Athena Bendix G-20 and
CIT G-21		 Univac 1108 IBM S/360-67 IBM 1401 Site Home
IBM 7040 Hybrid Lab G-15 Bogart DEC PDP-10 CMU Photos from Chris Hausler

CIT G-21

This Page last updated on .

This page continues the documentation of the computers at Carnegie-Mellon University  in the 1960's. You can click on the links above to learn about some of the other computers.

The CIT G-21 was a home grown computer which consisted of two Bendix G-20 CPU's connected together with a shared memory.  Most of us who worked on the Athena, cut our programming teeth on the G-21 (using ALGOL-20 and dozens of other obscure languages like THEM THINGS and SNOBOL).

Here are some old G-20 computer ads.  Wikipedia Page.

Here is a paper written by Jesse Quatse : Design of the G-21 Multi-Processor System. It was published in 1965. Jesse was involved in the memory architecture of the G-21. This paper was scaned by Pat Stakem who found it in his junk pile, er, archives....
   
Quatse, Jesse T. .Design of the G-21 multi-processing system. Pittsburgh, PA : Carnegie Mellon
University, Dept. of Computer Science, 1965, ENGR&SCI 510.7808 C28R 65-5, Pittsburgh, PA
Title :   DESIGN OF THE G-21 MULTI-PROCESSOR SYSTEM,
Corporate Author : CARNEGIE INST OF TECH PITTSBURGH PA
Personal Author(s) : Quatse, Jesse T.
Report Date : 26 FEB 1965

At the CS50 Symposiom at CMU in 2006, Jesse Quatse and Dick Shoup had some slides about the G-21. The videos used to be on-line but now seem to be gone. Dr. Quatse passed away in 2013 (findagrave).

John Yurkon sent me a scan of  A Visual Display System Suitable for Time Shared Use by Jesse T. Quatse published in 1965.


The two men in this ad might be Hal van Zoeren and Carl Lefkowitz.
 John Yurkon found this advertisement on eBay and sent it to me in Dec 2016. From the March 1962 issue of Scientific American.
Some reponses to the photo:
Interesting!  It's not a configuration I've ever seen.  There was some evidence that the first G-20 was installed for a while in GSIA as I saw a photo of a G-20 console being unloaded from a truck parked in front of GSIA and I know the "comp center" first started at GSIA with I believe an IBM 650 (which may have been the 650 which had been in HH53 the year before we came there).  It's possible this photo was taken in GSIA.  The photo shows what appears to be the LP-10 in use for output.  I wonder who the two guys in the photo are? 

I think I said this once before but I came across information that 48 G-20 processors were produced and I've always wondered where the others went.  There were actually three at CIT (one, called the "engineering prototype" was over in the CS lab in the basement of Porter, I once saw it there) and one of those CS50 talks referred to one in Mexico, but again, where did all the others go...???

73, Chris Hausler
The ad says 1962.  The G-20 was first produced in 1961 and the Internet says Scaife Hall was designed in 1961-1962.   A GSIA machine is certainly possible.

Roy Engehausen
I was a Freshman in the Fall of 1962, and the G-20 was on the fourth floor of Scaife Hall at that time.  I know this because I took S-205, Intro to Computing, that term.  We learned GATE on the G-20. 

Bob McFarland

There's evidence that Scaife was completed no earlier than sometime in 1962.  Given the lead time for magazines, particularly ads, I suspect the photo was taken no later than late 1961.  Further that round column shown behind the processor doesn't look like anything I remember from Scaife but as I spent little to no time in GSIA, I cannot say whether such structural members were located in GSIA.  Further such a large column might be found at a lower level in a building and apparently the comp center did get its start in the basement of GSIA according to that article in Carnegie Mellon today quoted on Mark's 1108 page.  But that article implies that the move happened in 1961 which doesn't match a 1962 completion date for Scaife, so who knows?  History is disappearing before out eyes...

All the computer items shown in that photo were at CIT, even the never apparently used when I was there LP-10.  It has been suggested that maybe the photo wasn't taken at CIT but one of the other ads for sale showed a G-20 installation said to be at Humble Oil in Houston and it was certainly a different configuration.  What's more, the "Now in Operation" at the bottom of both photos suggests that the CIT photo was of the installation at CIT.  That Humble Oil photo at least tells me where another G-20 went.  What with the three at CMU, the Humble Oil one and that CS50 talk which indicated there was one in Mexico, now we only have to find the other 43 missing G-20's...

73, Chris Hausler  
I can add one more G20.  I believe that there was a 4th cpu in Pittsburgh at Duquesne University maintained by the CMU folks.

David Chou
From:    "Matt Batt" <mbattsprynet.com>
Subject:    CIT G-21 Page
Date sent:    Mon, 26 Jun 2017 00:23:41 -0400

Hi Mark:

Just ran into your website while Googling old G20 photos.  Your site certainly brings back memories.  I'm one of the lucky people that was at Carnegie Tech computer center from 1960-1963.  Thought I'd send you a couple of comments on the CIT G-21 page.

The two folks in the picture on the Bendix G-20 Computing System Ad at Carnegie Tech are Hal van Zoeren and Carl Lefkowitz. 

Hal worked with Al Perlis and Art Evans when they were writing the early compilers - GAT for the IBM650 I think. (General Algebraic Translator). It was the precursor to GATE. (General Algebraic Translator Extended.)

Carl was the master of the G-20 Operating System.  I still have a listing. (THEM - Tech's Helpful Executive Monitor).  I met Carl in 1960-1961 when I worked as Tech's first operator in the basement of GSIA where the original G-20 was installed.  I remember the G-15 and the IBM650 that were also installed there.

I remember the move from GSIA to Scaife Hall. I was in charge of the G20 cabling layout under the raised floor at the new location in Scaife.  The problem was that they had specified a 5-6" raised floor but it was built at 1-2 inches.  The thick G20 cables didn't have enough clearance to cross each other under the floor, so it became an interesting layout.  It all worked out.  The G-20 was brought in through the 4th floor windows. 

Eventually, I maintained and extended the GATE compiler for a couple of years.  I got to meet and work with some incredible people at the Tech computer center in those early days.  I mostly programmed in G-20 assembler. After a stint at Control Data in 1963, I ended up at US Steel in Pittsburgh providing system support for the G-20 that they installed in 1963.

I'll quit now and see if this reaches you.

Martin (Matt) Batt
Matt,

Thanks for your email.

I've forwarded it to our small group of former students at CIT that spent way too much time at Scaife Hall. We would appreciate any and all comments about your time there.

Mark
Date sent:    Wed, 6 Sep 2017 17:03:34 -0400 (EDT)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    Re: Hal van Zoeren

Hi Mark,

I think Matt in his email (I just carefully reread it) is referring to the black and white G-20 advertisement you've recently added to the beginning of the G-20 page, not the color photo I submitted some years ago from that slide show for prospective students.  The man sitting in the color photo looks like the same guy standing to the right in the advertisement and I suppose that could be Carl (however, if you look at the rather dark photo of Pat, Russ and Carl I took in fall '76 at Carl's Penn Ave. establishment, Carl looks a lot thinner and otherwise somewhat different than the man in the other two photos.) 

So anyway I'm guessing that the man standing to the left in the advertisement is Hal and the man leaning over in the color photo is Jim.  I have no memory of Jim at all but do recall hearing Hal's name mentioned back when we were there but again never knew him.  Maybe he was gone by the time we arrived. 

Note that the man in the advertisement looks older than the man leaning over in the color photo.  My 2011 Alumni directory shows Hal getting an MS from MCS in 1955 (and lists his email address as vanzoerengmail.com in case you want to contact him :-) and living in Cupertino, CA.  The advertisement was likely from circa 1961 and the color photo a few years later.  The listing for Dr. Jim McElroy says he graduated from CIT in 1965 but his address is shown as "unknown" and so there is no further detail about him or his degree.  I checked my 1997 directory and likewise his address is unknown in that one too.  But the graduation date would at least indicate to me that he was a younger man.

73, Chris
I sent Hal an email.

Are you the Hal van Zoeren who was at Carnegie Tech back in the 1960's?

I and a few others have been collecting information and photos of that time and we have a Bendix G-20 advertisement from 1962 which might include a photo of you.

Could you take a look? We would be interested in hearing from you about your experiences at Carnegie Tech.

Mark DiVecchio
EE'70
From:    Hal Van Zoeren <vanzoerengmail.com>
Date sent:    Tue, 19 Sep 2017 22:21:19 -0400
Subject:    Re: CIT/CMU Computer Center

Hi, Mark.  Yes, I was at Carnegie Tech in the 60s.  In fact, I was part of the initial 3-person staff of the Comp Center when Carnegie's first computer (an IBM 650) arrived (along with Alan Perlis) in 1956, and I worked at the Comp Center from then until 1976, except for the years 1961-66, when I was at Stanford U.

I remember watching the G-20 being lowered into the GSIA building in 1960, and I even took a picture of it on the way down.  I haven't looked at that slide for many years, and unfortunately I can't look for it today.  It's in a closet in my home in Cupertino, but right now I'm in Massachusetts getting packed to return to California from a 3-month family road trip, and we won't get home until mid-October.

I hadn't thought much about the G-20 era for many years, so your note and your collected info brought back lots of buried memories.  I don't remember posing for that 1962 photo, but that certainly looks like me, and the other person looks like Carl Lefkowitz.  I'm not sure how many other relevant photos or documents that I have, but I'll check when I get back home and let you know what I find.

Hal




Some emails:

From: Mark DiVecchio <markdsilogic.com>
Sent: Mar 25, 2008 11:28 AM
Subject: RACE Unit

http://archive.computerhistory.org/resources/text/RCA/RCA.3301Realcom.1964.102641286.pdf

Pg 14 RACE unit??
Model 3488
Date:            Wed, 26 Mar 2008 10:31:47 -0500 (GMT-05:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         Re: RACE Unit

Hi Mark (and All :-),

It certainly looks like "Son of RACE" however it is not the units which were at CIT/CMU attached to the G-20's.  For one thing the packaging is sexier, but from a distance, the idea is the same.  This unit has removable card cartridges.  I do not recall that the "bins" on the units at CIT were removable, at least not easily removable, and if they were, we never removed them.  And if they were removable I believe it was only for maintenance purposes, not as an interchangable data cartridge.  This is likely a later designed unit and may have been competing with the IBM Datacell which did have removable cartridges (and I have a couple :-) 

CMU did have a Datacell unit attached to the 360/67 when it was still in Scaife but I don't know if it moved to Wean.  It had not been there very long, maybe arriving in 1970 or so.  Of course the Datacell was a disaster, making the RACE for all its problems look like good solid engineering :-)

As to the RACE.  The link refers to a 3301 computer but mentions the 301 and the 501.  A question for you all (and maybe for Dave Rodgers specifically) is the CS talks mention that the machine at CIT was a 301, however, I remember it being called a 501.  Which was it? 

Another story (or two).  I was in the room with the RACE relatively early after I first arrived at CIT, long before I was employed by the Comp Center, watching a demonstration of the RCA 301/501/whatever playing music (and my memory tells me that Dave was "master of ceremonies" :-) This was the typical "playing music on a computer" of the time, you'd tune an AM radio and listen to the RF energy being given off by the computer.  This was not unusual, I still have some paper tapes I used on the Hybrid Lab's PDP-9 which would do that.  What was unique about the 301/501/whatever demonstration was that in addition to the music coming from the radio, the computer operated its console typewriter to provide percussion :-)  I'd never before and never since seen that done. 

Second story.  It was rumored that the two RACE drives that I "cleaned" in spring 68 as a Comp Center operator were the second set of drives at the site.  They had replaced an earlier set of drives.  Now, just like the photo in the link, the drum around which the cards were wrapped was at one end of the machine covered by a plexiglass cover.  To clean the drives one opened this cover to gain access (it was hinged at the back and would flip up).  Even with the cover opened on the units I worked with, they would operate correctly.  However, correct procedure before opening the cover was to go over to the 301/501/whatever and stop its execution as there was no interlock on the covers.  To clean the drum it was necessary to remove the head assembly.  The story was that on the earlier drives, there was a deflector plate on the plexiglass cover that as the card came shooting out of the box toward the drum that this deflector plate directed it into the drum.  The later units I worked with had this as part of the head assembly.  With the cover open on the earlier units, the story was that the card would miss the drum and just fly out into the room.  Now, the cards were a stiff plastic.  The rumor was that with the earlier drives, someone at CIT forgot to close the cover when restarting the computer and it selected a card and shot it out into the room, the card remaining aerodynamic long enough to hit the side of the 301/501/whatever and leave a small dent in it.  True or false?  The story is that these units were replaced because this was in fact a safety issue, a flying card could seriously injure someone.  (As I recall the process of unbolting the head assembly in the newer units did trip an interlock and so it wouldn't shoot a card out into the room, but the cover was still not interlocked.)

Now, while I'm here, a story about the Athena, or actually the B250's, which I have not seen on the web site.  If you recall we had quite a number of these machines.  The story was that Williams had gotten them from Bank of America and that they had been used on an experimental system for magnetically marking and reading checks, some process that had preceded the now universal MICR markings still on checks today.  One problem with the B250 is that we had no way of saving programs written on it so one had to laboriously re-enter them.  However, after all, they used core memory.  We pulled the core planes from a number of the machines and tested them in the one machine we were using.  I specifically remember that we found four planes which would work without any adjustment in our 250.  Now I played a lot with the 250, wrote as much code for it if not more than for the Athena.  It was a character machine (6 bit words) and instructions were variable length.  I recall that an "add" was four words, the op code, the first source, the second source and the destination.  This was the first character machine I had ever played with and when I got my first "microprocessor" in 1977, a Motorola 6800 D2 kit, one of those hex keypad engineer's evaluation kits, one of my first thoughts was that programming it was just like the B250 ;-)  Anyway, for a month or two, I had my "own" B250 core plane.  Before I would start playing I would remove the one in the machine and install "mine" and then when done put the original back.  Thus I kept a copy of my programming.  There were a number of people involved in this testing of the core planes but whether anyone else used them as much as I did, I don't know (and I only did this for a month or two).

Finally, back to the G-20, if you haven't listened to Richard Shoup's talk I suggest you do.  He includes several pictures of the G-20's in his talk.  And Gorden Bell is, as usual, fun to listen to as well...

Chris Hausler, JH37
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$

RACE cards.
Photo by Chris Hausler, Jan 2009
Date:            Wed, 26 Mar 2008 09:42:17 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: RACE Unit

Hi guys..

I remember the RACE.  The computer as I recall was an RCA 301.  I also remember the legend of the card dent.

I think the  IBM 2321 Datacell showed up in 68.  Could have been late 67.  It didn't stay very long because it never worked well.  IBM would pronounce the box as working and I would run a job I had to stress it and it would break.   About a decade later I was transferred by IBM to San Jose.  The 3850 MSS was just coming on line (mag tape cartridges) and I mentioned the Datacell.  I was informed that the memories were painful and were best forgotten.

The Univac 1108 had its Fastrand drum.  Not quite as Rube Goldberg as the RACE and datacell

Roy
I think I was RE01
From: "Mark DiVecchio <markdsilogic.com>"
Sent: Mar 26, 2008 2:15 PM
To: Roy <r.engehausengmail.com>
Subject: Re: RACE Unit

Roy,
   I think you were HE01.
Mark
Date:            Thu, 27 Mar 2008 15:22:36 -0500 (GMT-05:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         Re: RACE Unit

Hi All,

Yes, Roy and I are both "first name challenged".  Until I was about 10 years old, I knew I was Chris Hausler, but I didn't know about this other guy named Jon C. Hausler who was shadowing me (and I've been trying to lose this dude, unsuccessfully, ever since ;-)  Of course, we can't forget H. Guyford Stever, so we don't suffer alone...

By the way, Roy, I have a brief video of you taken our freshman year during Spring Carnival 1967 as we were tearing down the WRCT field set up after the sorority relay races (as I haven't been to Spring Carnival since I graduated, I wonder if they still do sorority relay races :-)  There was also the faculty (read graduate student) egg toss done at the same event...

I managed to miss filming the buggy races that year (I have them for the later years, particularly 68) as I had spent both the Thursday night and the Friday night up all night playing with the IBM 7040, not getting to bed until about 6 AM, and although I set my alarm to wake me up in time for the races, I never heard it either day...

Regards,
J. Chris Hausler (see that dude sneaks in there most every time...)
From:    David Rodgers <dave_rodgersmsn.com>
Subject:    RE: RACE Unit
Date:    Wed, 26 Mar 2008 11:36:29 -0700

The RCA machine in Scaife Hall was a 301.  The two Race units were back to back.  Although the magazines were not easily removable, you could instruct the machine to unload a card for replacement since there was a lot of wear on the surface from all the pinch rollers.  Depending on factors unknown, sometimes the card would fly out of the front Race and shoot the person sitting at the 301 console in the back.  It hurt but was not fatal ;-)
From "Mark DiVecchio" <markdsilogic.com>
Date: 03/27/08 8:54 PM
Subject:    RACE Unit

http://www.feb-patrimoine.com/projet/gamma30/rca_301.htm   RCA

BullRAC

http://www.feb-patrimoine.com/english/gamma_30.htm    Gamma 30
http://febcm.club.fr/english/corporate/computer_innovation.htm    Bull and computer innovation

http://febcm.club.fr/english/ge400.htm    GE

http://www.dvorak.org/blog/?page_id=8215    IBM and the Seven Dwarfs - Dwarf Six: RCA
Date:            Thu, 27 Mar 2008 23:53:03 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: Other RACE links

Mark DiVecchio wrote:
>
> http://www.dvorak.org/blog/?page_id=8215
> » IBM and the Seven Dwarfs - Dwarf Six: RCA Dvorak Uncensored
>
Lots of inaccuracies in this one.  We all know RACE was Random Access Card Equipment.  The Univac Fastrand was simply very large drum memory.  One was installed on the 1108.  No mention of the IBM datacell.
Date:            Fri, 28 Mar 2008 00:06:21 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: Other RACE links

I also remember RCA having a bunch of IBM 360 Clones.  In the early 70s everyone was making them.  My brief Army career was spent studying Soviet computers and even they built 360 clones.  They even ran OS/360.  They stole the software and hardware but bought the books.  The Soviet embassy in Washington was the largest customer of IBM manuals in those days.

http://shashwatdc.blogspot.com/2007/07/feature-cias-take-on-computer-it.html
Date:            Fri, 28 Mar 2008 13:28:11 -0400
From:            "Patrick Stakem" <pstakemloyola.edu>
Subject:         Re: Other RACE links

and, did I mention I have a copy of the Bogart Programmer's manual?
..p

Mark's note: I had found that manual on-line. Click here.
Date sent:    Wed, 14 Mar 2018 14:37:08 -0400 (GMT-04:00)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    Early RACE...

Hi Pat and Mark,
 
While looking for something else, I came across the following video describing a mass storage device developed by Potter Instrument Co (the same folks who made the G-15 and G-20 tape drives) with replaceable magnetic cartridges with each "card" a bunch of strips of half inch magnetic tape.  This was apparently in the late 50's.  Anyway, interesting...

https://www.youtube.com/watch?v=664p6K6F59g
 
73, Chris
Date:    Mon, 17 Mar 2008 13:11:20 -0500 (GMT-05:00)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    Re: Jesse Quatse

Hi All,

Once again, what a trip!  I can highly recommend viewing this video.  It brought back lots of memories and confirmed a memory of mine that the scopes had three screens.  I hadn't been sure whether two or three.  And speaking of the scopes, Pat recently sent me a photo of me in front of the scopes, likely taken my sophomore year (67/68).  I've included it below. 

Somewhere I have some slides taken of the G-20's that day in 1976 I previously mentioned when Pat, Russ Moore and I found them in an old bank building being maintained by Carl Lefkowitz.  As I recall Carl was not too happy that I was taking pictures but I did get a few.  One of these days I'll have to locate them and get digital copies made.  I don't have any photos of the G-20's made while they were at CMU. 

The talk refers to 2 CPU's.  There was actually a third which was kept in Porter Hall in the CS engineering lab and referred to as the "engineering prototype".  I believe this was acquired after the other two when Bendix / Control Data abandoned the G-20 line.  While at a CMU homecoming some years ago I saw a couple of photos in an exhibit in the Library showing the machine being off loaded from a truck in front of GSIA.  This has lead me to believe that the first one might have been installed in GSIA but I'm not sure.

Regards,
Chris Hausler

Chris Hausler in the Scopes Room
about 1967-68

Scan from Philco Integrated Memory Display System Operators Manual
Sent to me by Tom Engelsiepen
From:    Charles Putney <chputneygmail.com>
Date sent:    Wed, 10 Mar 2021 22:49:04 +0000
Subject:    Philco Scopes

Mark,

I came across this 1969 writeup by Alan Bond (An Interactive Graphical Display Monitor in a Batch-Processing Environment with Remote Entry).  He was in charge of the Scopes room and allowed me and others to use them.  At the time, I remember Dave Vavra, Fred Schuman, John Yurkon were all using the scopes.  Fred Schuman wrote a 'space war' program based on what we saw running on a PDP5 at PIT. At the time, the program ran on two scopes until one player shot the other.  Then you had to wait 15 minutes to resubmit the program.  We had the complete listing of the code so we quickly found the time check and removed the limitation so that we could run the program all the time.  Of course, nothing else was running so Roy Wile came over to have a chat with us.
--
*Charlie*
Date:            Sat, 22 May 2010 11:30:33 -0400 (GMT-04:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         G-20 Scopes Art

Hi All,

Some of you who worked/played with the G-20 scopes might remember one individual who was different from the rest of us in that he was a fine arts student using the scopes to explore the idea of computer generated dynamic art.  I recall at least a couple times siting in the scopes room in the dark watching over his shoulder as his ephemeral art appeared and changed on the screen.  I seem to recall rotating splines as one of the "works".  Beyond this fuzzy memory I had otherwise almost forgotten him and this.

Well, last night at a local CMU event (the Rochester Clan's annual banquet) I crossed paths with this man for the first time since those times in the dark there in the scopes room.  Although I have been an active member of our local Clan for decades this to my knowledge is the first time he ever attended.  His name is Duane Palyka (I had remembered the Duane part and that his last initial was P).  It turns out he is a recently retired professor from the "School of Film and Animation" here at Rochester Institute of Technology.  He continues to experiment with computerized support for the arts, see <http://people.rit.edu/dpalyka/ScanFix.html>  In his interest in dynamic graphics arts it would seem that his work with the G-20 scopes lives on.

 Anyway we had a nice chat.  Small world...

Regards,
Chris Hausler
I remember Duane's programs running (painfully) on the G21 printers.  He generated many abstract "pictures" on the printers.  Often they would print for many minutes (hours?), wearing out the printer ribbons.

Dave
Thanks for the writeup.  I had forgotten Fred’s name so thanks for the refresh.

John

Hi Charley,

Neat article, thanks for sending it. One interesting thing I note is that the article appeared in the November 1969 issue of the ACM magazine. I recall that the G-20's along with the scopes were removed from service at Tech and sold to Carl Lefkowitz sometime summer 1968. I had started working as a part time computer operator at the beginning of 1968 and ran the G-20 on and off that spring (I have a G-20 break page showing me as the operator :-) but when I returned for fall semester 1968 only the 1108 and 360 remained for me to operate.  When Pat, Russ and I visited Carl in fall 1976 he still had the machines and the scopes.  I took some photos that day and you can see two of the scopes consoles in the back of the fourth and last photo on Mark's page:  http://www.silogic.com/Athena/Carl%20Lefkowitz.html . I don't think he ever did anything with them.

At CIT/CMU we had a very advanced for the time computing facility available to us in the late 60's.  A decade later I was in a position to hire just out of college engineers and math majors as computer programmers and I was somewhat shocked at the primitive computing environments most of them had available to them in college when compared to what we had a decade before.

73, Chris Hausler

Hi Dave,

I started working as a part time operator at the beginning of 1968 and recall that my salary was $1.61 an hour.  I recall that when I quit as an operator to go to work for the CS Engineering lab the end of spring semester 1970 my operator salary was around $2.25 an hour.  I'm trying to understand Lauston's memory that he was using the scopes in fall 68.  I know that I never operated the G-20's after spring 68 but it is possible that they were still there although I don't know doing what.  I know that all normal student course use had been moved to the 1108 by that time but maybe there were still some research projects finishing up on the 20's.  But I have no idea who was running them as the operator staff was no longer doing it.  Lauston, do you recall the last time you used the scopes?  In fact even in spring 68 the 20's were little used.  I think there was one introductory course (S-something) still using them but also recall hearing that spring 68 was the last time this would be done.  In fact I recall when working on the 20's during my shifts spring 68 that I would frequently play with them, screwing around with my tapes, as they were otherwise idle.

73, Chris
CMU was blessed in the 60’s with Alan Perlis as founding leader of computer science department and polymath Herbert Simon who drove the creation of exceptional for the time facilities.  As president, economist Richard Cyert made those early investments in CS curriculum and GSIA.

Dave Rogers
Chris,

I arrived in the fall of '68 and played the green sun program on the scopes. Whoever owned them at the time, they were still on campus.

Lauston Stephens
I remember spending two (possibly three Christmases) as the only operator on the 4th floor of Scaife.  Doris Crabb (DC05– I was DC08) would assign me since I could run all three systems.  Of course, it was dead quiet and I ran no more than a handful of jobs.  As usual, the 7040 ran by itself before it was replaced by the 1108.  At least one year, I even assembled a Heathkit during the shift to keep me busy.  Imagine soldering in a computer room today – it would trigger all kinds of fire alarms.  And I was paid my regular student wages at $1.60/hour – no holiday pay!

Dave
At this point, there is reason to accept everyone's memory with a grain of salt. John Yurkon was my room mate for the '68 fall semester. He enlisted in the navy after that, so it was that semester because he was working on the program butI beat him at playing it. As I recall, he didn't appreciate that.

Lauston
Actually it was the Air Force.

John
John,

Sorry. You make my point about memories.

Lauston
There’s a lot that I’ve forgotten or misremembered.  I think I find it easier to remember id’s like DV03 and CP05, if I’ve gotten those right.

There was a small tube based computer that I don’t think was working properly while I was there, the G-15.  It resembled a refrigerator.  Not sure which building that was in.

John
Ahh, I remember it.  It was a tube based 4k drum memory machine made by Bendix before it released the G20 (https://en.wikipedia.org/wiki/Bendix_G-15).  It was located in one of the subbasements in Hamerschlag Hall and students could play with it at will.  It was configured to load from paper tape with a Friden as a console and used a one-write tape drive which required a bulk erase when it was filled.  I ran a few jobs on it but gave up because it was so temperamental, and it took forever to do a compile.  It was cooled using a exhaust duct and worked fine in the winter, but overheated in the summer.

I saw on of these in an electronic surplus store in St. Paul (Crazy Louis) in 1978-79, but because it was so big (I had a small apartment) and in bad shape, I couldn’t convince myself to take it home.  I instead chose to make a copy of the Altair using surplus parts from the same store.  I already had an ASR30 to use as the I/O device.

Dave

Hi All,

The G-15 was located in HH53 along with the PDP-8. HH53 also had an ASR35 in it so one could dial into the G-20. I played a number of games of Chess from that 35 :-)  HH53 was Gordon Bell's anti-office and there was a room within HH53 with a door to Bell's private office. 

I've attached a photo showing yours truly at the G-15 console.  A similar photo (taken the same day) is on Mark's site.  I only ran Algo on it and only used it during my freshman year 66-67 as I recall it was becoming marginal by the end of that school year.  I remember hearing that some student had been maintaining the machine but he had graduated.  I recall hearing that some other folks were running Intercom on it, an interpreted "machine". 

The G-15 was apparently the last commercial machine designed based on some of the particular (peculiar :-) ideas about machine design put forward by Alan Turing and was very difficult to program in its "native tongue".  Some data I have on numbers of machines manufactured says that as to vacuum tube computers, the IBM 650 is the clear winner.  There had been a 650 in HH53 the year before I arrived (visited campus with parents for fall 65 homecoming and recall seeing it) but all that was left of it was the side panels when I got there.  I think it was removed to make room for the then new PDP-8   More or less tied for runner-up in number of vacuum tube machines made were the G-15 and the LGP-30 both with somewhere in the range of 400-500 of each manufactured.

About 15 years ago there was a project in Australia to restore a G-15 to operation but apparently it fizzled out. 

Somewhere I read that there was a G-15 given to CIT as an appeasement for late delivery of the first G-20.  I'm assuming that "our" G-15 was the same one.  About 20 years ago when attending a CMU fall homecoming there was a display of old photos in the Hunt Library and one of them showed a G-20 console being offloaded from a truck in front of GSIA where, as Scaife (RIP) was not yet complete, it was apparently first installed.  I assume the G-15 had been in GSIA as well as after all Herb Simon was associated with GSIA.  Although G-15's apparently originally sold for around $50K by the early 60's they could be had for about $10K from Bendix.  I sometimes wonder if these were returned leased machines and Bendix was just trying to unload them, maybe why they just gave it to CIT.

The G-15 was removed a year to two later.  I recall hearing that it was given to some high school.

After that year we also lost access to the PDP-8 in the room as a graduate student project took it over.  They added a single platter disk to it and I think this led to the project described here:  http://bitsavers.trailing-edge.com/pdf/cmu/van_de_Goor_A_Special_Purpose_Time_Sharing_System_for_the_PDP8_Computer_Aug67.pdf  .

As to the computer science department not only Perlis and Simon but Newell was also involved.  Newell was very friendly to me and even gave me access to his private library (his secretary was a stern librarian :-).  One of these guys apparently said that the computer science department was handing out PhD's before the school even knew the department existed :-)

73, Chris

Date:            Sat, 26 Dec 2009 11:16:06 -0800
From:            Roy <r.engehausengmail.com>

What was the name of the library system that ran on the G21?  You could store your source code in it.  The files were kept on the G21's disk or in the Race.  I remember that it was ported to the 360 but  never was a success there.

Roy
Are you thinking of "AND" -- Alpha Numeric Directory?

The G-20 was the first computer I ever used.  The Fortran like compiler was called "gate" -- Generalized Algebraic Translator, Extended.  The new version of gate was called "pile" for Prison In London England (Newgate). The Assembler was called "that" --To Help Assemble Translators.  There was a version of that called "this" -- To Help Instruct Students, which could check student's answers to problems.  And of course, eventually, there was an ALGOL compiler.  And also, if I recall correctly, IPL-V, Information Processing Language Five, which was a list processing language.
Bob
I thought it was AND but wasn't trusting my memory.  I don't remember "this" or "pile"

IPL-V was a machine language that ran on a simulated computer. Everything was for list processing.  There was also "Formula ALGOL" whcih actually stored equations rather than evaluating them.  There was  SNOBOL of course.  I don't remember if LISP ran on the G21.  There was also an analog computer simulator.

Anyone remember the analog computer in Hamerschlag around 69?  It was on the floor at the same level you entered from the parking lot facing Scaife.

Roy
You can see an IPL-V Programmer's Reference Manual in PDF at
http://www.rand.org/pubs/research_memoranda/2008/RM3739.pdf

Bob
Hi All,

I think IPL-V was sort of an assembly type machine language for an imaginary machine implemented in software by or based on the ideas of Newell and Simon.  It thus didn't actually compile into anything and was directly interpreted.  I looked at it briefly and don't have much further memory.  I think it was ported to the 360 but I believe the AI world was hooked on LISP (which I recall was based on the IBM 7090 hardware or at least one of those 7000 series machines).  I once had the IPL-V manual but I don't believe I do anymore.  I do still have the SNOBOL (4?) manual.

73,
Chris Hausler


From:    <jpollock1sbcglobal.net>
Subject:    Dialing Card
Date sent:    Sat, 1 Jun 2024 11:26:47 -0700

While cleaning up a drawer, found this dialing card from back in the day.

Jim
Great picture!
I remember those!
PS08
Jim,

The card says 360/20 which would seem to say it was from a 360 model 20 dialup. But maybe it means something else.
There is a 683 exchange in the 412 area code which seems to be in the Oakland section of Pittsburgh.

Mark
Hi All,

Looking at the card there are two 7  digit phone numbers, one immediately after the other, on it.  You will note that the "stop" hole is punched adjacent to the first digit of each number and I have this fuzzy memory that one could put one of these cards in the dialing unit and cause it to pass one of the stop holes and proceed to the next without dialing the first number.  So maybe one number was for the 360/67 and the other for the G20's.  But I have no records showing what the phone numbers for the systems were.

Just MHO...

Ah yes, the memories...

73, Chris
I vaguely remember the 7701 as the series for the G21 modems.

Dave Chou


Computer History


Date:            Thu, 24 Dec 2009 13:47:42 -0500 (EST)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         Computer History

Hi All,

About a month ago I went crazy on Amazon adding a number of computer history books to my already substantial collection.  The first new ones I'd bought in quite some years.  A couple of note.  The first, "A Few Good Men From Univac" by David E. Lundstrom, published by MIT Press in 1986 discusses early Univac history, early Control Data history. Control Data more or less came out of Univac and Cray Research history as Cray more or less came out of CDC.  Unfortunately the Athena gets just one sentence in the book and although it is mentioned that CDC bought a number of other computer operations in the early 60's they are not named and there is no mention of the G-20.  However, its still a good read.

The second book came out of a four part BBC radio program (or is that programme :-) and is titled, "Electronic Brains / Stories from the Dawn of the Computer Age" by Mike Hally, a BBC radio producer.  It tells a number of stories, some, obviously British, LEO being an example, but one section talks about a man named Loring Crosman. 

In 1943 he came up with an idea for an all electronic business computer, before any real research into such a possibility.  He took this idea to James Rand and although he then worked for a rival company to Remington Rand, Rand liked his idea and hired him.  After several experimental models the machine was produced commercially and called the RAND 409.  It was a decimal machine, not binary and could handle ten digits, but it did it with bi-quinary math, four counters and a carry flag like an abacus.  It did not have any stored program capability but was programed with plug boards. 

With Remington Rand's purchase of Eckert-Mauchly and the follow on wild media success of the UNIVAC I, the 409 design was updated into two models, branded the Univac 60 and Univac 120.  These sold well, with about 1500 total of all models sold by 1960 but the vacuum tube technology was then out of date.  Because of the good sales, it was decided to re-host Crosman's design once again in a transistorized machine.  The result was first available commercially in 1963 and was branded as the ... wait for it ... Univac 1004.

Three years later a follow on design, the Univac 1005 was released, the first in the line with some stored program capabilities but this was to be the last of the machines designed based on Crosman's ideas (he had passed away back in the 1950's) however by the end of its run about 10,000 Univac 1004/5's had been sold making it one of the most successful machines of its type, it was advertised as a "punch card electronic computer".  It was very rugged and reliable, fully capable of payroll and tax calculations by itself.  Popular with the military as the first "transportable computer", the very first 409's had in fact gone to the IRS for checking tax returns.

As Paul Harvey used to say, "Now you know the rest of the story"...

Regards,
Chris Hausler
As I remember, Carnegie had at least two Univac 1004s.  The first was used to print or reproduce card decks.  It was located in a room between the RCA 301 and the 7040.  A second one acted as the card reader for the 1108.

There was a special "feature" called the 1004 card fountain.  Failing to properly place the weight where the cards exiting the reader were stacked allowed them to bounce out of the tray and arch gracefully though the air.

Roy Engehausen

>As I remember, Carnegie had at least two Univac 1004s.  The first was used to print or reproduce card decks.  It was located in a room between
>the RCA 301 and the 7040.  A second one acted as the card reader for the 1108.

Yes, they did, but the first one, the one you mention being used for the 301 and 7040 is the one first used with the 1108 and which had a separate card punch added but only on the arrival of the 1108.  I don't think it had a punch on it before this time. It also had a Uniservo VI tape drive.  Originally instead of using the 7040's IBM printer one could write a tape on the 7040 and then list the tape using this 1004.  I did this one night in early '67.  With the arrival of the 1108, the 1004 had two new programming plug boards wired for it, one which made it a reader/printer/punch on the 1108 and another which connected it to the Calcomp plotter so we could mount "plot tapes" on its Uniservo VI tape drive for plotting.  Later, a second 1004 was purchased and placed on the 3rd floor in the I/O area so I/O clerks could read in cards too.  This unit did not have a punch but with its arrival along with the high speed printer which was also installed about that time there were now two 1004's making a total of two card readers and three printers on the 1108.

>There was a special "feature" called the 1004 card fountain.  Failing to properly place the weight where the cards exiting the reader were
>stacked allowed them to bounce out of the tray and arch gracefully though the air.

Yes, sadly, in my job as part time computer operator I "exercised" this option a few times.  The problem is that the receiving card bin had a slight slope to it and the weight would roll up this slope as cards exited the reader, the building deck pushing it up the slight slope.  Unfortunately, if you removed the cards from the receiving bin and forgot to manually push the weight back down the slope, the weight would not always roll by itself back down the slope, leaving a gap.  If you then started the reader without the weight all the way down, cards, which the drive fired column 1 down into the bin would spring back out thus launching them into the air.  I question "gracefully", it was more like rats abandoning a sinking ship. OOPS! Sorry...

Regards,
Chris Hausler
As an 1108 operator, I always dreaded loading a large deck of cards into the 1004.  If there were any cards which were "bent, spindled or otherwise mutilated" there was no possibility that the deck would remain intact and I hated writing a note to users of why his cards were shuffled.  Another feature of the 1004 was that any deck that went through the reader more than a half dozen times would almost certainly be shredded.

I remember one time when someone from GSIA submitted a deck which consisted of more than 30-40 boxes.  The lead programmer (I can't remember his name) rescued these decks from the 1004.  He wrote and ran a JCL card-to-tape utility on the IBM and loaded the tape on the Univac.  I think it took more than an hour on the IBM reader.

I attended the University of Minnesota some 10 years later.  Upon taking a programming class, I walked into the user facility and discovered (to my chagrin) that all the I/O was user operated on an ancient soupped up 1004.  I rapidly decided to store all my programs on-line and only use a small deck to print outputs that might be needed.

Dave Chou
Hi Dave,

The real problem was that the first die the card had to pass through covered about a quarter of the lead (9 row) edge of the card right at the center.  This was the same spot that rubber bands around the decks would ding the cards.  Further with a number of short user decks (typical on the 1108), once you had (hopefully remembered to) removed the rubber (or in Pittsburghese, "gum") band, you would have several dinged cards, the end of one deck followed by the beginning of the other, right in a row.  Rubber bands and paper clips on cards were always a problem.  I used to find separate rubber bands in decks between cards.  That said, I much preferred cleaning up jams on the 1004 than with the IBM reader (1402?).  The 1004 came apart easily and usually the card if inside the machine had jammed going under the read head which was just held in by a couple jack screws which were easy to remove, the rest of the card path being quite open.  Further the 1004 usually only chewed up two or three cards before it stopped. I recall 1402 jams with half a dozen cards shredded and although my memory is fuzzy, I seem to recall the 1402 being a pain to disassemble enough to easily clear the jam.

Regards,
Chris
I agree with the "gum band" problem on the deck.  Unfortunately, I found the IBM card readers not difficult to disassemble, probably because I learned on the IBM reader used on the G21.  In any case, one reason for the better handling was that the stacker on the IBM released cards into the reader in a controlled fashion into a secondary hopper and shuffled the deck on the way into the hopper so that the cards did not stick.  It was also infrequent that I had to take the reader apart.

Dave
I would have thought the 2540 would have been better.  The card path was straight through (the famous face down 9 edge leading) and then the card fell in the hopper.  The 1004 was a path where the card went in lengthwise, through the reader sideways by column then bent up and
finally down into the hopper.  It was a 270 degree change.

I was listing a deck for a user one night and the 1004 was printing away.  Then I noticed the cards weren't coming out.  They were all jammed under one of the rollers.

1004 pictured here
http://archive.computerhistory.org/resources/text/Remington_Rand/Univac.1004.1964.102646252.pdf

2540 against the back wall in this picture
http://www-03.ibm.com/ibm/history/exhibits/mainframe/mainframe_PP2040.html

Roy Engehausen
Hi Dave,

Yes, the 1402 "fluffed" the cards for you and kept an even and light weight of cards on the load mechanism.  As I only ran the G-20's a handful of times in spring '68, I don't ever recall a jam on it but I do recall a couple bad jams on the 360's reader.  I also spent two summers running 360's at Kodak and had some reader "fun" there too.  But as I read many times more cards into the 1004 at the 1108 than all other card readers ever, I did have more jams with it than the others.  The 1004 operators needed to manually fan or fluff the cards before putting them in the input hopper if they didn't want problems.  I got into a regular procedure for fluffing and fanning (and examining) card decks as I read them into the 1004.  Sometimes I would notice that the gum band had caused a small tear in the card.  These decks I would lay aside until a quiet moment would allow me to go over to a keypunch and make a new card.  Sometimes that quiet moment was several hours later and delayed the running of that job and sometimes the pile of damaged jobs would grow.  I also always wondered what the user thought when he found a different type or color card in his deck although frequently it was the job card or the end card with the damage and all job cards looked the same.  It was really bad when some project was due and all the last minute programmers were trying to finish the last night.  I don't think they had any idea what the card had to go through to work.  In fact a funny story...

One night it was really bad, the hall on the third floor of Scaife was jammed with sleep deprived people.  Now it was not uncommon for most of the operator staff to go out to lunch together between 2 and 4 AM, leaving one person to keep the machines (usually at this time the 1108) running.  But this night we were all disgusted.  I think we were short one person and two of us (I was one) were spelling each other on the 1108 (this was before the second 1004 reader had been installed on the third floor) the other one covering the 360.  I'm guessing this was the 68/69 year.  Anyway the entire operator staff and all but one I/O clerk left to go to lunch at the same time (there were a couple of popular restaurants, one on the south east side of the East Liberty loop and another on B? Blvd (next road north of Center) just west of Margaret Morrison).  Anyway we all had a good lunch, not being in any hurry to get back.  When we did return about 4 AM we were all pleased to see the population of the hallway, although still busy as compared to a normal night, quite reduced ;-)  Sometimes fast turn time is not a good thing...

Another funny story, particularly when you were reading in a lot of short student jobs (compile, load, crash).  As you would get a tray of cards you would start loading them into the input hopper, placing the gum bands on your wrist.  When done, you would of course pull the gum bands off of your wrist one at a time to re-band the decks.  Sometimes, when "done" you'd have an extra band on your wrist or sometimes you'd be a band short.  What had happened!?  I remember searching through the tray to find out where I had screwed up.  Sometimes I found the problem, sometimes I didn't...

Another dangerous thing to do with the 1004 was to start the reader without the backup weight on the input hopper.  You could do this if there were enough cards in the hopper.  This allowed you to start the reader before having loaded all the cards, particularly if you had several trays of cards to load and thus you would get done quicker.  One had to be careful however to gently place the additional decks on the stack because if you dropped them or pressed them too hard (or when you were done dropped the weight too hard) the reader would jam. The same was sort of true with the output hopper.  You could pull decks while the unit was still reading if there were enough cards in the hopper but as I said in the earlier post, you had to be careful with its weight to make sure it was rolled down against the remaining cards carefully.  If it wasn't or it hit too hard, the "fountain" which Roy talked about would occur.  Considering the hundreds of thousands of cards I read into that thing over the five semesters I worked there, I could still probably do it in my sleep.  Fun and games...

Regards,
Chris Hausler
Such customer service, repunching a card for who knows who and at how much  per hour? Where can we find such care today? On behalf of all "compile, load,  crash" students, "Thank you," a bit late.

And accounting for gum bands when you could have invoked the "sock in the  dryer" theory. For years, I thought I was losing an occasional sock until I realized that either they were mating in the dryer and the "extra" sock was the  next generation or the lint in the dryer filter was the earthly remains of one  entire sock.

These acts of devout customer service far outweigh taking a two-hour lunch break.

Where are the unsung heroes of today? Somewhere on Christmas day, unknown technicians are monitoring banks of web servers. What opportunities are there  for _heroic deeds_ (http://heroicdata.com/) ? Like the boy with his finger in the dike, is some underpaid technician using his own body as a conduit  to keep the web up and running?

Merry Christmas
Lauston Stephens
Hi Lauston,

As I recall it was a $1.76 an hour...

As to heroic deeds, one does occasionally get the chance.  I once early this decade had a customer's corporate jet coming to get me when I was in Savannah, GA at another one of their sites because a computer system in Albany, NY had failed (my boss had made a change, failed to test it then went on vacation) but before it got there I had driven an hour to get back to Savannah (I had been up the Tybee Island light house when the page came in, it was a Sunday) managed to dial in from the hotel where I was staying and although the line was bad, it would only hold about three minutes and then I would have to dial in again, managed to figure out what he had done wrong and got it all running again.  And then there was the time in 95 in Salisbury, NC for another customer with a pair of ancient DG Nova 3's on a system design  I had never previously seen.  I was a hero for about 6 months due to that one, but that story can wait for another time...

73,
Chris Hausler
$1.76? I got $1.61

an remember that pesky $20 Pgh employment tax that came out of the first paycheck every year?

..ps08
I don't remember what I was paid (it was definitely less than $2), but I do remember that I did not get any extra pay for working Christmas!

Dave Chou
At 08:11 PM 12/24/2009, Roy wrote:
>The 360 card reader/punch was a 2540.  The 1401 and 1403 were printers.

Actually, the 1401 was a computer. It was not very intelligent, and was used mostly for copying data between card reader. printer and magnetic tape. I remember trying to program the 1401, for what I don't remember.  It  had instructions like "p" to print a line. The 1402 was a card reader, but I seem to remember a 1442 card reader on the G-20.

Bob McFarland
You are correct, I mistyped.  The actual printers were a 1403 (maybe 1403-3) and two 1403-N1.  The N1 were the newer models.  They were the one where the cover would raise and lower by itself.  Never leave your coke on the top of the printer.  If it ran out of paper, the cover would swing open automatically.  I saw it happen at least once.  The N1 also
had the power paper stacker.

The 1442 was also a card reader but I don't recognize it as the one for the G20.  I was never a G20 operator so didn't send much time there.

http://en.wikipedia.org/wiki/IBM_1442

1402 card reader info located here.  This appears similar to my feeble memory

http://www.bitsavers.org/pdf/ibm/140x/A24-3072-2_1402_rdrPunch.pdf

I spent the summer of 67 working for Brooklyn Union Gas.  They had four 1401s that did nothing but card-to-tape and tape-to-print.  They even had a 360/50 with the 1401 emulator.  These machines fed their 7074.  I was a programmer working on their new accounting system that would run directly on the 360/50.  The code was all assembler and we used decimal arithmetic for most of it.  I got real handy on all the decimal instructions.

Roy Engehausen
I remember the 1403-N1 on the 360.  I called it the :Lotus Blossom because of how it opened when out of paper.

>The 1442 was also a card reader but I don't recognize it as the one
>for the G20.  I was never a G20 operator so didn't send much time there.

You are right.  The 1442 looks more like the card reader on the 360.

Bob McFarland
The G20 definitely had the IBM 1402 card-reader punch.  The picture of it is unforgetable.  Late on the midnight shift, I would start a long job, turn off the reader, stick my head in the reader output hopper and go to sleep. It was the only warm place in a room that was religiously kept at 68 degrees to keep the germanium transistors happy.  When more cards were needed, the console would err out, beep, and I would restart the reader.  Unlike the IBM 360, the reader was not programmed to start and stop on demand, a good thing when your head was resting in the reader.

Dave Chou


More of Pat Stakem's Photos

Two photos from Pat:

Pat sent me these photos in March of 2010. They had faded to blue just like my old slides. I'm no photo expert, but I tried to correct the color as well as I could.


Warner Hall
Pat told me that it was a light leak onto the film that caused the images to have the blueish tint. This frame was only half affected.

That's yours truly, Mark DiVecchio
Pat wrote:
I assume you still look the way you did in the picture of you taken in Scaife?
Of course!
To the right in the photo are those wooden mail boxes which were in the back hall of the third floor Scaife. When I started working there in early 68,  I had been assigned one of those mail boxes.  Note the location of the clock hanging from the ceiling by your shoulder as well.  Pat was apparently sitting at Linda Hammond's (comp center secretary) desk when he took this photo.  This also puts this photo into 1968 or 1969 as Linda's office was in a larger area to the left and forward of this shot the first year (66/67) we were there but later used for a couple other purposes over time, one being part of the I/O area (and I forget the order of the changes or even all of the details).

Chris Hausler

G-20 Manuals

Manuals at bitsavers.org:   http://www.bitsavers.org/pdf/bendix/g-20/

Machine Language Programmer's Manual

David Chou (DC08) another CIT alumnus, sent me scans of the G-20 Machine Language Programmer's Manual.

Dave wrote:

From:    "David Chou" <dchouu.washington.edu>
To:    "'Mark DiVecchio'" <markd@silogic.com>
Subject:    RE: CIT/CMU Comp Center
Date:    Sat, 29 Mar 2008 00:38:51 -0700

Mark,

I have included half of the G20 machine manual and sending it since it is a little big.  I am scanning the other half now.   I hope it makes it to you. I will break it down into fewer pages if needed.  You will notice that my user code is on the manual (DC08).

I also have the WHAT, THAT, and ALGOL20 manuals.  These are a little bigger and may be more difficult to scan.  Let me know which you would like me to do first.

Do you plan to post these manuals to your web page?  I wonder if anyone else is interested in them.  I enjoyed reading about Carl Lefkowitz from one of the others.  It is interesting that all our rememberances are rather similar about the nature of these devices.

Dave
Here is Dave's first email to me:
--------------------------------
From:    "David Chou" <dchouu.washington.edu>
Subject:    hello
Date:    Wed, 12 Mar 2008 22:19:22 -0700

Mark,

I somehow tripped over your web site looking for articles on the CMU robotic car.  Its been many years, like you and Roy Engehausen, I worked at the CMU Comp Center and graduated in 1970 (Chemistry).  I am on the faculty of the University of Washington (in Seattle) and supporting computer systems in the hospital/medical school environment.  I drifted back to my comfort zone with computers after spending time in medical school.  Things really haven't changed much in what I do.  But unlike the tech world, the academic environment requires that I still work.  Your pictures of the Atlas launch computer brings back memories.  Unfortunately, I have tried to get a few pictures of the infamous Bendix G21 computer, but have not had any luck.

Again it was nice to see your old pictures and the memories it brings back.

David Chou
dchouu.washington.edu
dblchoucomcast.net




It is scanned in two parts PART 1 and PART 2.

G-20 Service Manual

Pat Stakem scanned this in March of 2010. He wrote: "Look what I found in the attic. Unfortunately, not the whole thing."




Dave Chou wrote:
"Gee - that means I can fix a G20 again???!!! 
Ironically, I found a G15 in a junk/surplus electronics dealer in St. Paul, MN in the late 1970s (Crazy Eddies on University Ave).  I was tempted to buy it, but lived in a small apartment, and it looked like it was badly abused."

THAT


From:    "David Chou" <dchouu.washington.edu>
To:    "'Mark DiVecchio'" <markd@silogic.com>
Subject:    RE: CIT/CMU Comp Center
Date:    Sat, 29 Mar 2008 12:12:57 -0700

Mark,

Here is the THAT manual.  I haven't looked at this one for a while.  The G20 wasn't really as sophisticated as the 8080.

Dave



From:            "David Chou" <dchouu.washington.edu>
Subject:         RE: CIT/CMU Comp Center
Date:            Fri, 28 Mar 2008 16:30:03 -0700

Mark,

I also worked in the I/O desk, so that it probably where we met.  I do not know the people in the picture for the Athena.  I actually never worked on the computer even though I saw it many times.  I worked on the G15 for a while.

The G21 was purchased for very few dollars by, I believe, Carl Lefkowitz, who was a director of the Comp Center at one point.  CMU tried to sell the machine, but there were no takers (surprise).  This is consistent with the comments made by one of the others in your blog on the Athena.  There were several others who programmed the G21 who joined him in supporting it and getting it going.  I heard that they were trying to sell computer time, but personally I doubt that they made any money and I don't know who would buy time.

An old picture of me is on the UW web site  (http://myprofile.cos.com/choud2).

Dave

WHAT





From:    "David Chou" <dchouu.washington.edu>
Subject:    WHAT manual
Date:    Fri, 11 Apr 2008 20:13:19 -0700

I don't know if this will make it at 21 MB.  WHAT was the embedded assembler in ALGOL-20.
 
Dave
Date:            Sat, 12 Apr 2008 12:16:41 -0400 (EDT)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         Re: WHAT Manual

Hi Mark,

Thanks for posting those manuals from Dave Chou. Looking at the THAT manual, it doesn't seem (at least in the first few pages at which I looked) to explain that THAT is an acronym for "To Help Assemble Translators".  The computation center was fond of acronyms, but I don't know for what WHAT stood unless just a play on THAT.  The file system AND stood for either "Alpha Numeric Directory" or "Alpha Numeric Director System".  They also had THEM THINGS which stood for "Tech's Helpful Executive Monitor" and "To Help Implement New Generalized Systems".  Anyway, you can see where I got the idea and encouragement for SHE and HER THINGS ;-)

Since Dave is looking for pictures of the G-20's don't forget to tell him to view Richard Shoup's talk as it contains a few pictures of it.  I also believe I sent out that photo Pat took of me in front of the scopes if he's interested in those as well.  I still haven't dug deep enough to find those photos I took in 1976 of the G-20 at Carl's place and somewhere I believe I have an early photo of the G-20's at Tech which was once part of a slide show provided in the early 1970's by the CMU Alumni office to show to prospective students (haven't yet relocated that slide tray either and that photo may be one of the ones shown in Shoup's talk). 

Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Date:            Sat, 12 Apr 2008 12:40:30 -0400
From:            "David Vavra" <davavravverizon.net>
Subject:         RE: WHAT Manual

IIRC, "WHAT" stood for "Which Helps ALGOL Translate" or some such. Might even have been recursive: "WHAT Helps ALGOL Translate." It's undoubtedly meant to be a sly reference to THAT.

WHAT was the embedded(?) assembler in ALGOL-20.

These really bring back some old memories.

DAV
Date:            Sat, 12 Apr 2008 09:53:37 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: WHAT Manual

WHAT = Which Harms ALGOL Terribly

When I was working on ALGOL-67 (the 360 version of ALGOL-20), we had two great tools:  TREWQ and QWERT.  I never heard what they stood for except the characters on the upper left of the keyboard.

Some other old languages:  Formula ALGOL, COMIT, SCADS(?), and IPLV

Roy

CIT User Manual - Computation Center

I had forgotten that I even had this. Dated 10 Feb 1966. Introduction signed by David Nickerson, Assistant Director of Computation Center.

Do you remeber the Dump Count?  The DOLDUMP command? Man Numbers? How about the Input-Output counter in Scaife Hall SH 421 (before it moved to the third floor) or the User Consultant in SH 316?

You could always call Ext 580 for a recorded System Status Report. I must have recorded hundreds of these when I worked at the I/O desk but I had completely forgotten about them. How about the courier service that left the I/O counter every hour on the hour and went to Hamerschalg Hall, Baker Hall, Doherty Hall, GSIA, and Porter Hall?

The manual also warned you not to wait until the last two weeks of the semester lest you suffer the "last minute rush when turnaround is exceptionally high." "Week-ends, in particular Sunday mornings, are often good times to run programs since relatively few users wish to employ the machines at these times" -- but, of course, we knew that and we were one of those few.

Chris Hausler wrote: "As to the dump count, as the red users manual says, you could always override it at your PERIL ;-)"

This manual sold for $1.10 in the bookstore.



CIT User Manual - Table of Contents and Introduction
CIT User Manual - Hardware
CIT User Manual - AND Ssytem
CIT User Manual -  ALGOL-20 (will scan on request)
CIT User Manual -  WHAT (will scan on request)
CIT User Manual -  GATE (will scan on request)
CIT User Manual -  Backus Naur Form (will scan on request)
CIT User Manual -  FORTRAN (will scan on request)
CIT User Manual -  IPLV (will scan on request)
CIT User Manual -  LIPLV (will scan on request)
CIT User Manual -  TIPLV (will scan on request)
CIT User Manual -  COMIT (will scan on request)
CIT User Manual -  LIST (will scan on request)
CIT User Manual -  THAT (will scan on request)
CIT User Manual -  Double Monitor (will scan on request)
CIT User Manual -  THEM THINGS (will scan on request)

ALGOL-20

Written by Janet Fierst, David Blocher, Robert Braden, Arthur Evans Jr and Richard Grove.  Preface signed by Arthur Evans and dated January 1965. I had forgotten that ALGOL-20 did not have recursion as described in ALGOL-60. That made me remember first writing a program with recursion using Univac 1108 ALGOL which contained that feature.

Sold for $1.25 in the bookstore.



Chapter 0 - Table of Contents and ALGOL Ready Reference
Chapter 1 - Introduction
Chapter 2 - Notes on ALGOL at Carnegie Tech
Chapter 3 - ALGOL-20 Input/Output
Chapter 4 - System Statements
Chapter 5 - The ALGOL Library
Chapter 6 - Miscellaneous
Chapter 7 - ALGOL-60

Emails

Date:            Sun, 16 May 2010 21:46:36 -0700
From:            Paul McJones <paulmcjones.org>
Subject:         ALGOL-20 for Bendix G-21

Dear Mark,

I am working on a project to collect and preserve as much information as possible about the ALGOL programming language and its historic implementations; I'm working in cooperation with the Computer History Museum in Mountain View, California and I'm maintaining a web site at http://www.softwarepreservation.org/projects/ALGOL/ .

I'd run across an old paper by Arthur Evans, Jr. with the title "An Algol 60 Compiler". It said he was at Carnegie, but it didn't mention a machine. This evening a friend said he'd run Algol on the Bendix G-21 at Carnegie around 1964, and with that clue I came across your web page discussion of ALGOL-20, and the scan of parts of the manual.

I've added this information to my web site, but it would be great to have a scan of the entire ALGOL-20 manual. I wasn't sure if it was you or Dave Chou who scanned the manual, but I thought I'd ask you first. I appreciate your help in this matter.

Paul McJones
Paul,

Thanks for your email.

I'm happy that someone out there in web-land ran across our web site. I'm pretty sure it was me who scanned the G-21 Algol-20 manual. I will be happy to scan the rest of it for you. You might have to give me a few weeks to fit in the time. As a high school junior in 1965, I had the pleasure of learning Algol from the editor of that manual, Jan Fierst.

I also have a CIT User Manual which has a section on Algol-20. That section may have more information.

Did you also see the Univac 1108 Algol-60 manual? Its at: http://www.silogic.com/Athena/Univac%201108.html. I already scanned that complete manual.

I'm happy to support your work. I've forwarded your email to the dozen or so other ex-Carnegie students who are mentioned on my web site. They may have more to add.

Mark
I did not scan the manual, but I still have a copy of the original.  I too was taught by Jan Fierst in the NSF program, but it was a year earlier (1964).  That is how I got the lifetime addiction into computers.

Dave
I sent Paul a note about Algol-67 and Forumla Algol (which ran on the G21)

Roy
Mark and Dave,

It's nice to hear from you (as another person who acquired a lifetime addiction to computers in high school, but in California). Perhaps you met my friend Dennis Austin -- he was another high school student in the NSF-funded program. (He went on to work at Burroughs for a number of years, and later worked at a tiny company called Forefront, where he implemented the original PowerPoint.)

Re the Univac Algol manual, Al Kossow also has a scan at bitsavers:
     http://bitsavers.org/pdf/univac/1100/algol/UP-7544_1108_ALGOL_Jul68.pdf
plus an "extended" version:
     http://bitsavers.org/pdf/univac/1100/algol/UME-7636_1108_Extended_ALGOL_1968.pdf

Do you folks know anything about the origins of these compilers?

By the way, re scanning the manuals, I find that it's usually best to scan in B&W (one bit/pixel) mode rather than grayscale, unless there is a continuous-tone image (e.g., photograph) on a particular page, in which case I switch to grayscale or color. This generally results in smaller files with better contrast.

Paul
Hi all,
   I was a student at CMU from 68-72. During that time I took and "Intro to Algol" course taught by A.J. Perlis one of the group that defined ALGOL 60 in January 1960 in Paris.  I still have my book from the class. Its entitled  "Introduction to Algol Programming" by Torgil EkMan  and Carl-Erik Froberg; Stduentlittleratur, Lund Sweeden; Oxfor University Press, London, Copyright 1967.
   It has a short history ~5 pages , of ther language as chaptor 1.

Glenn Sembroski

ps: Apparently I bought the book used for $3.00
Hi All,

I too still have that Algol textbook as well as the 1108 Algol manual in my collection of "junque".  Several years ago due to a chance posting I made on alt.folklore.computers, I got in email contact with Art Evans who was apparently in charge of the Algol project for the G-20.  He told me in one of the emails that Jan did most all of the "heavy lifting" when it came to developing the G-20 Algol compiler and that at least at that time (early this decade) she and her husband were still living in Pittsburgh and that he had been in contact with her.  Possibly she would be a source of further information on Algol-20.

73, Chris Hausler
Chris,

I would be interested in talking with both Art and Jan -- about G-20 Algol, and I'm also interested in the PAL programming language Art was involved with at MIT.

I just Googled around in alt.folklore.computers and came up with the email address " EvansSL21 at earthlink dot com (Fix the address and change 'com' to 'net'.)" -- is that the one you had for him?

Paul

Hi Paul,

Unfortunately, due to failed computers and lack of proper backup procedures, I no longer have record of any of those conversations with Art and thus cannot verify the email address you provided. Although I've visited Pittsburgh and the CMU campus frequently in the last 15 years, I have never seen Jan there nor have I tried to contact her. It is unlikely she would even remember me after 40 plus years. Although I remember talking with her at several Computer Center parties we never worked together on any project and our association was at best brief and fleeting. I think she had moved on after my Sophomore or Junior year. As I recall from the conversations, Art was already gone before I arrived on campus in 1966. It was a thread on the aforementioned interest group early this last decade on the G-20's at CMU which first brought us together. My suggestion would be to just send Art an email at that address and see what happens.

Regards,
Chris Hausler
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Chris,

I'm sorry to hear about your disk drive problems; I've had my share of  them too! I'll go ahead and try to get in touch with Art, and keep you  posted on what I learn. (I just discovered an additional way to reach Art: through the multicians.org Mail facility.)

Paul
Date:            Sun, 30 May 2010 17:54:31 -0700
From:            Paul McJones <paulmcjones.org>
Subject:         Re: ALGOL-20 for Bendix G-21

More about Carnegie in the G-21 era:

     Report on a Visit to the USA
     R F Churchhouse, F R A Hopgood, E B Fossey
     November 1965
     17th May to 12th June, 1965
     Atlas Computer Laboratory
     http://www.chilton-computing.org.uk/acl/literature/reports/p011.htm

contains a section "Visit to Carnegie Institute of Technology (Pittsburgh) 31st May/1st June 1965" that mentions the G-21, S-205 and other courses, and what various people were working on at the time; it ends with a description of a modem with an acoustic coupler, although neither of those terms is used.

Paul

P.S. I came across this while adding Formula Algol to my web site.
Sent: Tuesday, September 14, 2010 9:05 PM
To: J. Chris Hausler; Patrick Stakem; Mark DiVecchio; David Chou; Dale Dewey; Dave Rodgers; Dave Vavra; Glenn Sembroski; Lauston Stephens; Roy Engehausen; John Yurkon; Charlie Putney; Bob McFarland; James Pollock; Tom Engelsiepen
Subject: ALGOL-20 Manual Scanned
 
I've scanned the ALGOL-20 manual that I have. If anyone wants a copy, its on the G-21 web page.

Mark
From:            "John E. Yurkon" <yurkonmsu.edu>
Subject:         RE: ALGOL-20 Manual Scanned
Date:            Tue, 14 Sep 2010 21:11:34 -0400

Thanks Mark!  I really liked Algol.  Wasn't there an Algol-60 on the G21?
 
John
The Algol-20 implementation of Algol-60 had a couple of major variations
that I remember:

1) I/O was not defined in Algol-60.  Algol-20 used a local formating language which (in my opinion) was one of the better printer I/O formatters that I have ever used.  It had better functionality than Fortran, and I have never seen anything similar.  Of course, given the variable pitch fonts used today, this formating is arcane and not very applicable.

2) Recursion was deliberately removed due to memory limitations. I still love this dialect of Algol, but that's probably another function of an aging memory.

Dave

COMIT

COMIT was a text processing language. I didn't scan the manuals but will on request.








SCOPE User Manual

By Alan H. Bond

From the Preface:

This manual describes how to use the Scopes. It is concerned mainly with software, as the hardware is treated definitively in the Carnegie Tech document: "A Visual Display System Suitable for Timeshared Use" by Quatse, Jesse T., latest version December 1966, obtained from the computation center documentation staff. In order to use the Scopes, it is sufficient to read the Quatse document and the Scope User Manual. The Quatse manual should be read first. A brief description of hardware concepts and terms is given in section 3 of this manual.

Click on the cover to download.




Input-Output Envelope

Didn't we all have dozens of these?

Univac 1108 Manuals

These have all been moved to a new 1108 web page.


Date:            Sat, 12 Apr 2008 10:07:10 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Mark's web page

A great web page including a picture of the G-21 tape drives :  http://www.ece.cmu.edu/visitor/history.html

More about 4th floor of Scaife Hall


From:            "David Chou" <dchouu.washington.edu>
Subject:         RE: Mark's web page
Date:            Sun, 13 Apr 2008 00:54:27 -0700

Roy,

I looked at the picture in the ECE web site.  The tape drives are exactly as I remember.  It appears to be an earlier picture of the 4th floor of Scaife for two reasons: (1) The drapes on the windows are missing or open.  The G21 was so temp sensitive that we closed the curtains to keep the CPU from high-temping.  Being a germanium transistor computer, any temperature over about 68 seemed bad.  (2) I remember the two G21 consoles faced the computer and tape drives rather than being perpendicular to them.

Dave
Date:            Sun, 13 Apr 2008 12:07:17 -0400 (GMT-04:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         RE: Mark's web page

Hi Dave and Roy,

I did work as a part time comp center operator starting spring semester 1968 through spring semester 1970 when I went to work for the CS dept engineering lab as a technician and did work a number of shifts that first spring as the G-20 operator (it is my memory that the G-20's were removed during summer vacation in 1968 but it could have been 69).  Although I rarely worked as a 360 operator at CMU, Roy, I did spend the summers of 68 and 69 working as a 360 operator at Eastman Kodak on 360/44's, 360/50's and 360/65's running OS/MVT.  The 360/65's in addition to their development role were acting as back ups for the older generation (IBM 7080's IIRC) production systems and so had walls of tape drives :-)

Anyway, Dave, you are right, the configuration shown on the ECE web site is an early one.  The first time I saw the G-20s was when I was a high school senior at Homecoming 1965 when I came down with my parents (my father was a CIT grad) to see the campus.  At that time, the G-20's were in their "final configuration" that I could tell, not the one shown in that photo.  Speaking of that photo, as I mentioned in a previous post I had an earlier photo of the G-20's which had been part of a 60 slide slide show put out by the Alumni office which our local (Rochester, NY) Carnegie Clan (I'm still a member of the board) used to show the campus to prospective students in the 1970's.  Well, I have located that slide tray, and the photo, a Kodachrome color slide, is the same one shown on the web site and also one of the ones shown in Shoup's talk.  The slide set is marked September 1968.  Shoup does show another picture of the G-20's in his talk which does show the final configuration as I remember it so you might want to check that out (and its a great talk too).  I have yet to locate (get to) the slides I took (Ektachrome so I don't know what the color will be like) in 1976 of the G-20's at Carl's.  I still have some serious "mining" to do in my utility room to get over to where my collection of slides I took is located :-)

Speaking of the G-20 tapes (and I still have my three spools :-) does anyone else remember that they were "blocked" just like DECTAPE.  They had directories and you could use them just like a (rather stringy) disk.  Of course, floppies wiped out DECTAPE.  I've always been curious if this blocking was an inherent capability of the G-20 tapes or just something that CMU developed.  Remember with "normal" half inch computer tape once you write to a spot on the tape anything after that point is effectively destroyed.  Both G-20 tapes and DECTAPES needed to be formatted just like a disk and I know DECTAPES had a timing track so I assume G-20 tapes did too. 

Remember that most of the G-20 software was developed at CIT and in many ways was way ahead of the rest of the world of computing.  In fact, when the G-20s were replaced by the 1108, the computing environment at CMU in some ways, IMHO, degraded.  In the late 1970's I was in a position of hiring quite a number of just out of college engineers as computer programmers.  I was shocked at that time how primitive their college computing environments had been as compared to what CMU/CIT had 10 to 15 years previous.  I then came to realize what a privileged environment we had at CMU/CIT when we were there.

Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Date:            Sun, 13 Apr 2008 04:35:54 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: Mark's web page

Hi,

I was never a G-21 operator so I didn't deal with the drapes.  I recollect one evening being frustrated at all the computers being down due to an A/C failure despite the sub freezing weather outside.  I also have a fond memory of smoke pouring out of all the computers due to a fire in the basement next to the A/C intake.

I do remember the G-21 card reader and console being just in front of the tape drives though.  The memory units were on the backside facing the railroad tracks.  You got to the G-21 area by going through the separate room that held the printers.

My days as a computer operator were all on the 360/67.

Roy
Date:            Sun, 13 Apr 2008 13:33:58 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: Mark's web page

Hi J :-)

I thought the PDP-8's tape drives were blocked.

After my Army stint, I started with IBM in Manassas, VA.  Our project was some of the software for the Trident submarine.  The computers that were to be used on the ship were the Univac AN/UYK-7.  The tape drives were blocked as I remember.   They held the normal 1/2" tapes.  They were ruggedized for shipboard use.

Roy
From:            "David Chou" <dchouu.washington.edu>
Subject:         RE: Mark's web page
Date:            Sun, 13 Apr 2008 13:46:12 -0700

Chris,

You bring back some memories.  I did some programming for the G20 tape drives and remember the blocking.  There was a timing track on the tape as well as some unusual capabilities to handle skewing, an inherent problem with 1" tapes.   I do not know, however, whether CIT wrote programs to take advantage of this or whether the hardware just supported addressable blocks.

Of interest is that the blocking design allowed for the mapping out of bad blocks.  You could continue to use some really bad tapes, a useful feature by the end-of-life for the G20 when the tapes were no longer being made (thus my job as a student programmer).  Reliability was a big problem since the MT10 tape drives used mechanial arms rather than vacuum columns to manage tape slack.  During heavy search (and destroy) motions, the tapes really got a workout.

I have an old HP 35mm slide reader which I have kept attached to an an old Dell W2k computer.  For the most part, I have had pretty good success in adjusting the color on Ektachome slides.  Unfortunately, the process is slow.

Dave
Hi Dave (and All :-)

The female operator's name was Mary Noe.  She married Joel Bloom who I recall was a physics student (S68) working on a graduate degree, but that's a fuzzy memory.  I have a brief video of Mary and Joel (and a few others) at a comp center picnic.  Someone had tied some gas filled balloons in her blond hair.  If you have any memory of Mary, you can almost imagine her expression.  You're right Dave, she was a "pistol".  She was a fun person to be around.  A 20 year old CMU grad directory has them living down near NYC and he working for Chase.  Another incident involving Mary was that she bent over the back of the 1004 printer (on the 1108) to get some output.  If you recall at some point they added a "paper puller" to it which used rotating brushes and it caught her hair.  They got it stopped but her hair was vary entangled in the brushes.  The unit was detached from the printer and she carried it around on her head until they could get a UNIVAC CE there to disassemble it.  She said it had quite frightened her, but once again I can still see her wry expression as she carried the thing around...

Dave, are you sure it was the 360 which was dropped in through the roof?  It is my memory that it was the 1108.  I think I told in an earlier post about being one of the folks who disassembled the 1108 processor in summer 1971 for its transfer to Wean.  If you recall both the 1108 processor box and its memory box were long cabinets.  Very roughly 1/3 of this length was the power supplies and on the processor, this 1/3 was also the piece with the lighted pushbutton maintenance panel on it.  These had been delivered whole and thus the need to drop them through the roof, however, it was possible to separate them resulting in pieces which would fit in the elevator but this entailed some rather intricate unwiring of wiring harnesses which ran between the two sections.  It was my understanding that UNIVAC wanted more than the school was willing to pay to do this disassembly and so us CS dept engineering lab employee's were drafted.  It was my joy to be the one who took the processor box apart.  I always recall the older gentleman who was the head UNIVAC CE occasionally coming around behind the box where I was working to see how I was doing.  The worried expression was priceless ;-)

That same day as we were cleaning up and pulling cables from under the floor I found an old paper box under the floor with two reels of tape in it.  These had been the "on the machine" (not the master copies) of the IBSYS operating system and libraries for the tape operating system which ran the 7040.  These would wear out and occasionally get replaced, sometimes by shorting the tape beyond the used area and putting a new load mark on the tape and then copying from the master copies.  Someone had obviously put them there when the 7040 was removed in 1967 or early 68.  Thinking them "historic" I "acquired" them and still have them in my pile of junque to this day.  One is labelled S.SLB1 and dated 2/13/67 and the other S.SLB2 and dated 7/13/67.

And as to the Scaife elevator, it was quite the toy for certain bored adolescents like me and as I recall some others in this group.  We would get on top of it and take manual control of it.  But those stories are for another time...

J. Chris Hausler, JH37
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Hi,

I was a G21 operator prior to (and after) the arrival of the 360 and 1108. The fire event brings back a couple of events, many details I heard only second hand, so the details may be embellished.

Many of the 360 modules were too big to bring in through the elevator.  As a result, a hole was cut into the 4th floor ceiling and the computers were dropped with a crane.  In preparation for the 360, the 4th floor I/O desk and programming staff were moved to the third floor.  The rest of the 4th floor housed the 7040, located near where the 1108 was finally placed.  One day, I noticed that all of the CO2 fire extinguishers had disappeared. These were big units, about 4 feet high and very heavy.  One of the staff told me that the wood in contact with the cut steel plate caught fire after the workmen left.  The operator on duty was a thin and small 100 pound female (Mary N.) who I assume must have been really fired up with adrenalin. Fortunately, the room was mostly empty.

Some years later, I learned that IBM was upset over a roof leaking in one of the 360 CPU modules, resulting in some expensive repairs.

Incidentally, the Scaife Hall elevator was never very reliable.  I was stuck in it many times, often during campus tours for prospective students when the elevator was full. I should have surveyed to see if any of these attended Carnegie.

Dave
I remember that practically all the later machines were dropped in Scaife by crane.  The elevator was simply inadequate.  The 360 CPU unit was rather large and not easily broken down, especially since the 67 had a DAT unit attached as part of the console/CPU.

Dave
Hi Dave,

I've also got to argue about your memory of the 360 coming through the roof.  I know the 1108 did as I recall watching it.  When I arrived in fall 66 the I/O desk had already been moved to the third floor (directly underneath where I had seen it on the fourth floor during my visit in fall 65) and the walls in the area where the 360 was to be installed on the forth floor were already being torn down and false floor installed, but the machine was not there yet.  The machine was moved in, up in the elevator, later that fall.  When the 360 was moved to Wean in 1971, the same time as the 1108 and PDP-10's were, all went down the elevator.  Only UNIVAC had been "stupid" enough to build a machine which wouldn't fit in an elevator without some significant unwiring.  As I recall a little better use of plug connectors would have made this easy, but the 1108 had multipoint wiring harnesses running between the two sides of the machine, some breaking out into individual wires tied to individual screw points (says he who unscrewed them all :-).

But of course, my brain cells have been atrophying for a number of years and all the above may be false memories, or nightmares, as the case may be...

Regards,
Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Hi Chris,

The 360 was installed in late 66.   I used it enough that I was able to get a job during summer 67 as a 360 assembler language programmer.

The 1008 Fastrand drum was the box that had to come through the roof.  It may have been able to fit in the elevator but it weighed too much

Found a picture of the CIT job card : http://www.cs.uiowa.edu/~jones/cards/collection/i-logo.html

Roy
Hi All,

My memories of the Fastrand II are somewhat vague.  What I do know, however, is that it did not show up until quite some time after the rest of the 1108 was installed.  And it the processor and memory did come in through the roof.  The hole in the roof had been somewhere near where the processor ended up and I don't believe it wouldn't have been possible to open it up again without moving everything.  Remember the Fastrand was installed back in the area where the 7040 had been and the 7040 remained in service for a while after the 1108 was installed.  Eventually the wall between the two areas was taken down and the 1108 stuff expanded into that area (although it never filled it up).  The Fastrand came after this wall was removed IIRC, but I don't recall exactly how long after or when.  If they dropped it through the roof I certainly don't recall it.  Further everything was removed in 1971 by the elevator including the Fastrand.  The Fastrand was a long cabinet more or less like the processor and memory but I'm assuming it too came apart or it wasn't THAT long.  The elevator was (is) somewhat deep.  However I had nothing to do with either installing it in the first place or moving it in 1971 so my memories, once again, are vague and I could be wrong...

I also don't know what was wrong with and what happened to that PDP-8 RJE front end.  I recall rumors of hardware problems but that's about all and that memory too is vague. 

Another 1108 oddity was the DCT100 (I think that was the number).  One of these was installed over in the Hybrid lab and I recall submitting a job to the 1108 through it once or twice.  The printer part was sort of a one column drum printer and definitely somewhat Rube Goldberg.  So there was some kind of RJE on the 1108 other than the PDP-8.  There was also some leased line modems (202's ?) there which linked it to some company which apparently occasionally rented time on the 1108.  I recall hearing that there was a 1004 on the other end but don't recall the company name. 

If anybody knows anything more about this, please speak up.  As the ad says, Inquiring minds want to know...

Chris
I can't speak to 1971 or the remote job entry but the card reader/punch for the 1108 was the 1004.  There was one installed for years in the room next to the RACE and it was used for card listing and duplication.  Programmed by punch board.  When the 1108 was installed, the 1004 was used for it too.  I think there may have been other printers for the 1008 and the 1004 was only used for card operations.

Roy

PS.. I thought it was the 1040 but the Wikipedia says 1004
Hi Roy,

When first installed, the 1004 which combined a card reader and printer in one unit was used as the sole hard copy I/O device for the 1108.  That had been the unit that together with its UNISERVO VI tape drive had been an offline lister/reader for the IBM 7040 before the 1108 had arrived.  I never saw anyone else use it for such as the 7040 did have its own reader, punch and printer, but one night, just for the hell of it, I did :-) 

This 1004 ended up being the unit on the fourth floor with the 1108.  At the time the 1108 was acquired, a separate UNIVAC card punch unit was purchased and also connected to this 1004.  There were two plug boards for it, the one which configured it as a printer / card reader / punch on the 1108 and the one which connected it to the Calcomp plotter such that "plot tapes" mounted on its tape drive would plot to the plotter.  When the 1108 was first installed, the default object option was DECK so
every time you compiled a program, an object deck would be punched.  This was quickly changed to NODECK :-)

Exec II was a pretty lean OS, basically a batch system with spooling.  The 1108 rapidly would get way ahead of this one 1004.  Eventually a second 1004 and in addition a "high speed printer" was purchased and installed on the third floor and the 1108 could keep them all going at the same time, three print streams and two card streams.  As the I/O clerks were now reading in some of the cards, the operator's job was made a little easier (Although again, with the typical student job, "compile,
load, crash", the machine could frequently keep ahead even with both readers going full out and the machine itself would go "idle" but the printers would all rattle on for some time emptying the spool.) 

That high speed printer was a "trip".  It was so fast that paper jams were a mess to clean up.  It was a drum printer like the 1004's printer but vertical registration was very poor and sometimes a particular column would end up in the next row or go missing altogether.  This unit had its own small motor generator set to power it. 

Roy, as to the 1040, its a vague memory, but although there wasn't one at Tech, I believe such a unit did come into existence, a "new design" 1004.  I'm not sure about the number and never saw one but I do recall reading about such a follow-on product.

Chris

Misc Photos

From:    "David Chou" <dchouu.washington.edu>
Subject:    RE: Computer History
Date:    Thu, 24 Dec 2009 20:11:07 -0800

Three pictures that I took in the 60s of/in Scaife Hall and scanned in from my collection.  If they were processed in my usual fashion for Thistle pictures, the film is TriX developed with HC100 at ASA400 and taken with my Minolta SRT101, most likely with a 135mm lens:.

Dave Chou

1) a picture of Anne Rindt, one of the operators.

2) Scaife Hall during a Spring Carnival buggy race

3) Scaife Hall from Hamerschlag roof during one of the war protests on campus. I was chased down by the CMU authorities shortly after the picture was taken.
Hi Dave,

Great photos!  I'm still wracking my brain trying to remember the name of that other woman operator in the picture with Mark and Roy and made a list of all the female operators I could recall.  But I had completely forgotten Anne.  I remember she had a boyfriend who would come and pick her up, a tall guy.  I thought at one time he had worked for the school or even the comp center (a courier?) but that memory is fuzzier than all my others.  I think she was gone by fall 67 semester.

The buggy in the lead is, of course, PIKA's "Shark".

I don't know when you took the photo of the roof of Scaife but if it was late 67 or later, I'm going to suggest that the rectangular area nearest the corner of the roof to which you were nearest when you took the photo and covered with a something with a cross hatch pattern is the famous "hole".  Its about the right size to fit the 1108.  I know the hole was left in place and a cover fabricated for it but I don't recall it ever being used once the 1108 had been dropped in.  But you were not on Baker but Machinery Hall, Hammerschlag when we were there.

73,
Chris Hausler
Did they require the hole for the 1108 or was it for the FASTRAND drum for the 1108?  I always thought it was for the disk.  It definitely didn't go in the elevator.

http://en.wikipedia.org/wiki/FASTRAND

Roy
Hi Roy,

The processor and memory modules for the 1108 certainly came through the roof.  As delivered they were too big for the elevator.  The Fastrand did not come initially with the computer, it was added later and I have no memory of the roof being opened again.  Of course, my little gray cells have been rotted over time by too much wine, beer, other spirits and not to mention high living so just like I totally forgot Anne Rindt, I could have totally forgotten the arrival of the Fastrand.  In fact, beyond that it did not originally arrive with the machine I don't recall anything in detail about the arrival or departure of the Fastrand.  Someone else here will have to fill in any further information if they have it.

As I already wrote on the ASDG web site, although it didn't come in that way, the 1108 processor and its memory were taken down the elevator because when that happened I was a full time CS Engineering Lab employee and was drafted to help with this effort as I heard Univac wanted too much money to do it.  As I wrote, I was the one who disassembled the 1108 processor unit into its two parts, the processor rack itself and the power supply maintenance panel (flashy lights) rack.  They were interconnected by many cables not all plug connected, some breaking out into individual screw points and I unscrewed them all.  Someone else did this for the single memory module which was constructed the same way.  Once all the cables were undone and pulled into either one side or the other, I unbolted the rack supporting frames from each other and separated the two parts.  These were then individually moved down the elevator.  They had no built in dollies so we had to use those jack trucks, they look like a pair of hand trucks with hydraulic jacks to lift the blade relative to the wheels.  You'd put two of them on opposite ends of the assembly, strap them tightly together, then jack the unit.  The whole thing could then be wheeled around on the two jack truck's wheels.  I recall the longer two pieces (the processor / memory pieces were longer than their related power supply pieces) just and I mean just fit into the Scaife elevator as I remember clearly squeezing them in and then riding down with them.  Also as the elevator only came up to the base floor level and as most of the 4th floor of Scaife was at that time raised floor, there were sideways ramps leading down to the base floor level in front of the elevator and I remember a problem turning the longer pieces into the elevator, it was a tight fit.  I have no real idea why the 1108 and its memory were not originally delivered in separate pieces as well.  Putting the pieces together would of course be tedious and error prone and maybe Univac didn't trust their field service folks to do this back when the thing was first delivered but the cost of the crane and the hole in the roof had to be significant.  To my knowledge the entire 360, G20, 7040, PDP-10 and RACE came up and down on the elevator.

I don't remember moving the Fastrand.  In fact I don't recall whether the Fastrand was moved as I don't remember seeing it over at Wean but my employment with the school ended within about a month of this effort and I was not back on campus again until 1976, five years later.  The 1108 was still in Wean in 1976 and I visited it.  Although running that evening it was not being used.  I cannot remember whether I saw the Fastrand that evening or not and I've been told by the Unisys History folks that 1976 was rather late for 1108's to still be in service. 

If you look at the photo of the Fastrand II in the link you provided, it looks like there are three cabinets, two shorter ones on each end of the unit and possibly one or both are removable making the unit shorter just like with the processor and memory.  I don't think the total length of the Fastrand II was as long as the processor/memory boxes with their power supplies attached.  With that and no memory of the roof being opened up for its arrival, I've always just assumed that the Fastrand II came and went via the elevator.

Next time I'm on campus, possibly next April, I'll have to remember to take a tape measure and measure the depth of the Scaife elevator ;-)

That's the sum total of what I remember about the affair...

73,
Chris Hausler
I think the Fastrand fell into the same situation as the IBM 2321 Datacell. It was intended to hold source code files much like the RACE. The datacell was a great idea that never worked.  We would measure the MTBF in hours.  I had a job set up that used the datacell as the temporary storage for an assembler job.  I don't think it ever completed without the datacell breaking.  Years later I was transferred to IBM in San Jose and mentioned the datacell to my colleagues.  I was told that one did not bring it up in polite conversation :-)  For the 370 series machines, the IBM 3850 Mass Storage Device was much better!

CMU brought in the 2321 because the original 2311 disks were so small (7.25 MB).  The 2314 (29MB) arrived and the datacell disappeared.

I don't think CMU ever came up with a use for the Fastrand.  I think it was gone by 1970.  Disk drive capacity was making leaps and bounds and I think it just became obsolete too fast.

What was the name of the library system that ran on the G21?  You could store your source code in it.  The files were kept on the G21's disk or in the Race.  I remember that it was ported to the 360 but never was a success there.  I think they might have thought of using the Fastrand for it on the 1108.

Roy
FWIW, I also remember that the Fastran device was planned as a swap store for a 1108 time sharing system that was never installed, just like the Ampex LCS was intended as a swap device for the IBM.  I am not sure that the LCS was ever used as intended except as
tty buffers?

Dave Chou
Hi All,

Roy is correct about AND and mentions a number of other languages I recall but THIS and PILE are new to me.  I love the name PILE.  As I said in a post on the web site some time ago, the computation center was fond of acronyms and of course that's where I got the inspiration for SHE on the Athena.  I also recall a purchased FORTRAN II on the G-20 which I tried to use but it did not work very well.  Thus wanting to learn FORTRAN, I started playing with the 7040 very early in 67.

The 1108 for as long as I was there ran EXEC II which was just a batch system with spooling.  Univac did have EXEC 8 which supposedly did have some time sharing capabilities but it was in the same state at that time as was TSS/360.  It didn't work very well and CMU had the need to get work done, thus they stuck with EXEC II at least until I left in summer 71 right after the move to Wean.  I recall reading a long time ago that Univac was having the same problems with EXEC 8 as IBM with TSS/360 and that EXEC II was a quick fix written by some outside firm but that's something I read 20 years ago (although I still have the book, I don't know which one it is :-) 

Case Western did use EXEC 8 and I recall (Unisys History folks again) hearing it eventually was somewhat successful.  I recall being at Case's campus for a football game with Tech, so it must have been fall 68 as that was the last season I did radio remotes for WRCT and before that semester I wouldn't have known what an 1108 was, and remember walking past their computer center and its glass walls (stopping to put my nose against the glass :-).  They had two 1108 processors and four (64K 36 bit word) memory cabinets (CMU had one of each) and have read since that Case in fact did major work on getting EXEC 8 successful.

I don't know what the plan was, if any, for the Fastrand II relative to EXEC 8 but with EXEC II it was used as a primitive user file system much like how AND used the RACE and those large platter disks on the G-20, but I don't remember the details.  Without the Fastrand II, there was no magnetic user program or data storage on the 1108 (unless, like me, you had access to half inch tapes :-)  I remember I was taking some course and some of the course materials which each person had to access to do some exercise for the course were stored on the Fastrand.  I, however, had copied them to one of my tapes.  The Fastrand failed and was down for about a week.  They ended up using my tape.  The operators just left it mounted most of the time that week.  However, the Fastrand was still there and I think still in use at least until the move in early summer 71.  However I had given up my part time operator job early summer 70 when I went to work for the CS Engineering Lab, first part time then full time, so I'm quite fuzzy on the Fastrand's use that last year before the move.

The 1108 did have several very high speed "swapping" drums (FH432's ?) which I recall were head per track with flying heads but these were all used by EXEC II with no user access.  I recall hearing that their speed was one of the reasons for the 1108 being so fast when compared to the 360 running OS/MFT and whatever spooling it had at that time, HASP or ?? (I recall the running of performance benchmarks between them and the 1108 cleaned the 360 on most of them and was more or less even on the few it didn't).  Likely this was due to the fixed heads due to no seek time, just rotational latency and I believe they rotated a lot faster than the 360 disks as well but that's a very fuzzy memory.  These high speed drums, however were not very reliable.  I recall the Univac CE's having to swap out the actual drive units (a big aluminum box) with new ones on a regular basis.

73,
Chris Hausler
LCS was used for all sorts of things.  When first installed, the IBM 2361 was used.  Originally when the 67 was being used for batch jobs, LCS was used as buffers for the teletypes (which never worked well).   When in TSS mode, LCS was originally used for paging.  I think later experiments showed it wasn't worth paging and was used as main storage.

IBM LCS http://www-03.ibm.com/ibm/history/exhibits/space/space_2361.html

There was also a device called a Storage Channel where you could cause a page to go from the high speed memory to LCS or vice versa via the I/O subsystem rather than using the processor.  The backup system was simply sixteen MVS instructions.  One day the backup was activated and the system ran faster.  The storage channel was never used again :-)

I also remember some company (Ampex?) coming in and running tests on the 360/67 and LCS.  They were amazed to find out that the "underlap" was disabled.  Evidently underlap allowed memory accesses to the faster main memory while LCS was still cycling.  It created more problems and the feature had been disabled.  As I remember IBM's LCS took 8 usec for a memory cycle compared to 750 ns for main memory

I guess the Ampex LCS was installed because it was much faster than IBMs.

Found this reference but I can't get the article :-(

http://portal.acm.org/citation.cfm?id=1465693&dl=GUIDE&coll=GUIDE&CFID=69895914&CFTOKEN=93465281

Roy

Input/Output Counter in Scaife Hall


From: Mark DiVecchio
Sent: Tuesday, December 22, 2009 5:26 PM
Subject: 1967 I/O Desk at Scaife Hall

During the summer of 1965 Roy and I were enrolled in an NSF summer program at CIT. As an option, we got to take a programming class in ALGOL. Roy, do you recall if the instructor was Jan Fierst?

(Interesting ALGOL link: http://archive.computerhistory.org/resources/text/algol/algol_bulletin/ )

I recall the I/O desk on the 4th floor of Scaife Hall in 1965.

Chris remembered that the I/O Desk had aready been moved to the 3rd floor by the fall of 1966 (also Roy's and my freshman year). My memory is not clear on that move.

As part of my financial package to attend CIT, I got a part time job washing dishes in the cafeteria. That lasted about two weeks until I was able to land a job behind the I/O Desk.

I never worked as an operator (I worked as a programmer for a professor in the History department) but I remember Roy letting me follow him around many times on the 4th floor.

Mark

Jan 1967 - I/O desk on either 3rd or 4th floor of Scaife Hall.
On the right is Roy Engehausen. I'm (Mark DiVecchio) behind the desk.
I don't know who the girl is but she looks familiar.
Part of my tuition assistance at Tech was a part-time job as a dish washer in the cafeteria.
That lasted all of two weeks until I got a job behind the I/O desk.
Roy was my roommate during freshman year. He worked as a programmer.
Photo sent to me by Ray Carson in Dec 2008. Ray didn't say but I expect that he took the photo.
Photo stamped on the back "Jan 67".
From:            "David Chou" <dchouu.washington.edu>
Subject:         RE: 1967 I/O Desk at Scaife Hall
Date:            Tue, 22 Dec 2009 17:38:03 -0800

My memory agrees with your memory (Folie à deux or madness shared by two). I also worked the IO desk, initially on the 4th floor (1966) and later on the 3rd floor when room was made for the new computers, starting with the IBM 360 in 1967.  Later additional space was cleared out for the Univac 1108.  There was a carpenter's strike going on at the time, and somehow the walls fell down during a long weekend, allowing for the false flooring and other changes to be installed.

Dave
Date:            Tue, 22 Dec 2009 21:50:21 -0500
From:            Robert McFarland <rm08alumni.carnegiemellon.edu>
Subject:         Re: 1967 I/O Desk at Scaife Hall

She looks like Carol Shanahan, but at my age, I can't be sure. I barely recognized Mark.
Bob
Hi All,

Yes, the I/O desk was on the forth floor when I visited as a HS senior at Homecoming 65 and by the time I arrived as a freshman fall 66 it had been moved to the third floor directly below where it had been on the forth.  I think I described that already in one of the emails on the web site. 

I'm not sure, but I don't think the girl is Carol Shanahan.  This girl is too short.  Somewhere I've got some brief 8mm movies of Carol when she and John Godfrey visited my lake camp in the late 60's, but I haven't watched it in years and never did get that later reel copied to video tape, so I would have to find the reel and then fix the broken 8mm projector to watch it and compare :-(

What's more I remember a very slightly built girl operator working there that first year.  I think the girl in the photo is her.  A very nice friendly person, she was going to Pitt part time but I don't recall what she was studying.  She may have been gone by the next fall (67) when I arrived back on campus but I also have this very fuzzy memory of talking with her one morning in front of the 1108 which would put her there the next school year.  She certainly was gone, however, by the end of spring semester 68.  Her name is on the tip of my tongue but refuses to come out.  The guy sitting behind her looks familiar too but I have no idea what his name is.

73,
Chris Hausler
I am not so sure it was the fourth floor.  My memory (which may be bad) is that this was the I/O desk on the third floor.  It was basically across from the elevator.  By Jan 67, the 360/67 would have been installed (or very close to it).   I worked as an I/O clerk for a while and then programmer and operator.  I operated the 360/67 but remember getting classes on how to operate the 1108.  Anyone remember when the 1108 was installed?

Roy Engehausen
If this picture was taken in 1967, it was probably the third floor.  The 360 had taken over most of the areas where the original i/o desk was.  The 1108 came in about a year after the 360, probably in the time frame around 1968. I worked in the Comp Center for about three years and remember working Christmas day 1968 by myself and making the rounds as an operator on a very quiet day running all three computers.  By this time, the 360 was in production and doing something useful, other than blinking lights and crashing.

Dave Chou
My memory is a little foggy, but I'm pretty sure the 1108 was there when I started in the Fall of '67.  Certainly by early '68.  The Rand drum for it was installed but not generally available yet.

I need a reference point.  Was the Bendix G-21 on the 4th floor or was there one floor higher?

John Yurkon
Easy answer.  Scaife only had four floors.  All machines were on the top floor.  The G21 in the corner closest to the Schenley Drive bridge.  The 1108 was put at the end facing Hamerschlag.  The 360 was between them on the side towards Baker Hall.  The 7090 (or was it a 7094)  and the RCA 301 were also in the middle but on the side next to the railroad tracks.   The G21 was replaced by a PDP10. by 1970.

Roy Engehausen
I don't recall her name, but it's not Carol Shanahan. G21 was on fourth (top) floor.  I/O desk was on 3rd floor.

Dave Rodgers
Hi Roy (and all),

It was a 7040 or 7044, not a 7090 or 7094 and it had those disgusting 7330 tape drives with the short sideways vacuum columns.  It was located on the north west end of the 4th floor, north of the 301 which was as you say in the middle of the west side (the G-20's were in the southwest corner, the west side was the side overlooking Panther Hollow and the railroad tracks). 

And speaking of the 7040 and female operators (I still can't remember the name of that one in the photo), there was another one, Yvonne Gartland (or something close to that).  My freshman year someone (and it was either Dave Rodgers or Walt Sullivan as I recall :-) caused the 7040 to print out the the message "IBSYS DEMANDS THE SACRAFICE OF A VIRGIN" one night when Yvonne was working on it.  She came out wondering what she should do ;-)  This was also the girl we couldn't find one night when she was supposed to be working.  As you might recall, the stairwells in Scaife had these large circular lamps at the top.  Yvonne was eventually located at the top of one of the stairwells plastered against the wall.  She was afraid of spiders and there was a big one on the lamp....

73,
Chris Hausler
The G21 was on the 4th floor.  Until Comp Center moved out of Scaife, all the big boxes were on the 4th floor.  The "5th floor" was the roof and the big boxes were dropped in by cranes.

Dave Chou
The 360 was run for a year or two in two modes. The machine  was scheduled for different modes at different times.  The modes were mutually exclusive.

The batch mode ran it as a 360/65 (no virtual memory). OS/MFT with HASP was used at first and then replaced by OS/MFT II with JES2.  This was a productive mode from the start and basically ran card decks.

What CMU really wanted was time-sharing.  It hitched its wagon to the TSS/360 operating system but stability was horrible in the first few years.  There were attempts made to use other time sharing systems.  I remember running PTSS (Pittsburgh Time-Sharing System) from Pitt.  As TSS became more stable, it took over more and more of the schedule.  By the 68/69 school year, the machine ran TSS fulltime.

My programmer job with CMU was first to work on ALGOL/67.  The first versions ran in the batch mode and then later on TSS.  They switched me to porting APL/360 to TSS in the 68/69 school year.

Roy Engehausen
was she the one that we played the "cover the modems so we can blow the dust of the datalines trick?

..ps08
I was thinking the same thing. Her name was Yvonne.

DAV
Hi Pat (and All),

Yes, one of the two AT&T phone guys, Dick (or I think we called him "Dicky") did this.  Yvonne started putting the desk phones on the computer consoles in bags as Dicky had suggested but when she went into the G-20 room she saw all the 103A modems on shelves there in the back and came down to the third floor looking for Dicky to find out what he wanted her to do about those...  Yes, she was a sweet girl, but gullible...

The other phone guy was George and he was the one who kept the various leased teletypes repaired.  Several years later there was a big strike against AT&T.  The union had printed up bumper stickers which said, "Ma Bell is a Cheap Mother".  In fact by that time the school had purchased a number of 33's for internal direct connection to the PDP-10 (I set many of them up) and we hired George during the strike to keep them maintained as well as teach a guy hired full time to do that as well.  Anyway, George gave me a couple of the bumper stickers.  I put one of them on the back of my car.  Now, of course, my car had NY plates.  I was driving one day and realized I was being followed by a white "Bell of Pennsylvania" car.  So, not being in a hurry, I spent about an hour and a half driving around Pittsburgh eventually leading this guy back to where he had first started following me at which point he must have figured out that he had been made and broke off.  I assume he thought I was an out of state union organizer or something like that.  He did however obviously record my license plate because for several years following at my parents house here in Rochester I occasionally got what I would call harassing phone calls from AT&T even though the local phone company was an independent.  Probably a small price to pay for the things I did to them ;-)  BTW I still have the second bumper sticker (see attached photo which also includes G-20 PT-10 manual, the G-20 paper tape reader/punch unit which Dave Vavra used to punch Athena tapes from his cross assembler)...
>
>"Never trust a computer you didn't build; never trust a computer you can't lift"
>
I thought that was supposed to be, "never trust a computer you can lift" :-)

73,
Chris
Hi all,

I remember the I/O desk on the third floor but I don't remember much from the summer of 65.

I worked as an I/O clerk beginning the school year 1966-1967.  I then got a job as a programmer working on the ALGOL-67 compiler (the one for the 360).   We used TREWQ and QWERT to build/run the compiler.  QWERT was the compiler compiler and TREWQ was the runtime libraries.  The criteria for ALGOL-67 was to pass the man-boy problem : http://en.wikipedia.org/wiki/Man_or_boy_test

After ALGOL, I was switched to APL for TSS on the 67.  I spent a number of weeks bracketed around ROTC summer camp in 1969 installing APL on the IBM computers in Yorktown Heights, NY.

I also got to work as an operator on the 360/67.  I had Midnight to 8AM on Saturday and Sundays for a few years.

Roy
Subject:         RE: Your Arm
Date:            Sat, 27 Dec 2008 20:29:54 -0500
From:            "Ron Herold" <ronsolverrr.org>

Mark,

RACE is correctly detailed as Random Access Card Equipment.  They were built by RCA and driven by a 301 or 501 computer.  There were only 5 501s built - and CMU had a 301.  The only 501 I ever saw was at RCA Labs where I worked after graduation and that one was actually doing billing for NJ Ma Bell.

The site is full of pix and memories.  Glad you have them down and visualized - as our memories fade.  College is so far in the past.
Best wishes to you and Sally for the holidays and coming year. Thanks for keeping in touch and find time to come visit.

Ron


Date:    Thu, 14 Jan 2010 12:32:07 -0500 (EST)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    G-20 photo

Hi All,

Taking advantage of after Xmas sales, I just bought myself an inexpensive slide and negative scanner.  As they say, you get what you pay for, but it doesn't do a real bad job.  As I have mentioned, back in the mid 80's I was given a tray of 60 slides which had been put out by the CIT Alumni office in the late 60's to the various Alumni groups (clans) to be used to show prospective students the campus.  I've been a member of the board (or is that bored) of our local clan since the early 80's.  Anyway, it was either me or the dumpster and as you know I'm a pack rat so I took it.  The slides say they are Kodachrome duplicates and all have a date of September 68 stamped on them.  One of those slides, no. 24, is that photo taken in the G-20 room which was seen on some CMU web site as well as shown in Richard Shoup's talk, but I recall both were B&W.  Here's the color version attached.  This configuration of the tape drives precedes my visit as a high school senior in fall 1965. 

Anyway, enjoy!

73,
Chris Hausler

Richard Shoup identified these two men in a talk that he gave at CS50 in 2006.
The man leaning over is Jim McElroy, one of the OS programmers and
the other man was identified as Carl Lefkowitz.
Chris,

I was a Freshman in the Fall of 1962, and at least one G-20 was in Scaife Hall. I don't remember my first trip to the G-20 machine room, but it was much later than 1962. I took "Intro to Computing" (S-205?) in the Fall of 1962.  We used  Gate on the G-20. Dr. Perlis did the lectures for that class.  He is one of the few  professors from that year that I actually remember.  I also took G-20 Assembler (S-206?) in the Spring of 1963. After that, I started working as a User Consultant, and did some work for the IBM 360/67 before it arrived.  We had a disk with a Model 67 simulator that we ran at night on the Model 50 at Gulf Reasearch.  That must have been the Summer of 1966.

I don't know how the G-20 compared with other computers of its era, but I do remember that it had a 6 microsecond memory access time, with a minimum 6 microsecond instruction execution time.  Since the index registers were in core, any indexed  instruction (mode 1 or 3, if I remember correctly) added 6 microseconds to its execution time.

The original G-20 had 14 address bits, allowing it to access 16K 32 bit words.  It was modified so that one of the instruction bits (indirect addressing, I believe) was used as the 15th address bit, allowing access to 32K words.

Bob McFarland

Matt,

Thanks for your email.

I've forwarded it to our small group of former students at CIT that spent way too much time at Scaife Hall. We would appreciate any and all comments about your time there.

Mark


From:            "David Chou" <dchouu.washington.edu>
Subject:         RE: Mark's web page
Date:            Sun, 13 Apr 2008 00:54:27 -0700

Roy,

I looked at the picture in the ECE web site.  The tape drives are exactly as I remember.  It appears to be an earlier picture of the 4th floor of Scaife for two reasons: (1) The drapes on the windows are missing or open.  The G21 was so temp sensitive that we closed the curtains to keep the CPU from high-temping.  Being a germanium transistor computer, any temperature over about 68 seemed bad.  (2) I remember the two G21 consoles faced the computer and tape drives rather than being perpendicular to them.

Dave
Date:            Sun, 13 Apr 2008 12:07:17 -0400 (GMT-04:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         RE: Mark's web page

Hi Dave and Roy,

I did work as a part time comp center operator starting spring semester 1968 through spring semester 1970 when I went to work for the CS dept engineering lab as a technician and did work a number of shifts that first spring as the G-20 operator (it is my memory that the G-20's were removed during summer vacation in 1968 but it could have been 69).  Although I rarely worked as a 360 operator at CMU, Roy, I did spend the summers of 68 and 69 working as a 360 operator at Eastman Kodak on 360/44's, 360/50's and 360/65's running OS/MVT.  The 360/65's in addition to their development role were acting as back ups for the older generation (IBM 7080's IIRC) production systems and so had walls of tape drives :-)

Anyway, Dave, you are right, the configuration shown on the ECE web site is an early one.  The first time I saw the G-20s was when I was a high school senior at Homecoming 1965 when I came down with my parents (my father was a CIT grad) to see the campus.  At that time, the G-20's were in their "final configuration" that I could tell, not the one shown in that photo.  Speaking of that photo, as I mentioned in a previous post I had an earlier photo of the G-20's which had been part of a 60 slide slide show put out by the Alumni office which our local (Rochester, NY) Carnegie Clan (I'm still a member of the board) used to show the campus to prospective students in the 1970's.  Well, I have located that slide tray, and the photo, a Kodachrome color slide, is the same one shown on the web site and also one of the ones shown in Shoup's talk.  The slide set is marked September 1968.  Shoup does show another picture of the G-20's in his talk which does show the final configuration as I remember it so you might want to check that out (and its a great talk too).  I have yet to locate (get to) the slides I took (Ektachrome so I don't know what the color will be like) in 1976 of the G-20's at Carl's.  I still have some serious "mining" to do in my utility room to get over to where my collection of slides I took is located :-)

Speaking of the G-20 tapes (and I still have my three spools :-) does anyone else remember that they were "blocked" just like DECTAPE.  They had directories and you could use them just like a (rather stringy) disk.  Of course, floppies wiped out DECTAPE.  I've always been curious if this blocking was an inherent capability of the G-20 tapes or just something that CMU developed.  Remember with "normal" half inch computer tape once you write to a spot on the tape anything after that point is effectively destroyed.  Both G-20 tapes and DECTAPES needed to be formatted just like a disk and I know DECTAPES had a timing track so I assume G-20 tapes did too. 

Remember that most of the G-20 software was developed at CIT and in many ways was way ahead of the rest of the world of computing.  In fact, when the G-20s were replaced by the 1108, the computing environment at CMU in some ways, IMHO, degraded.  In the late 1970's I was in a position of hiring quite a number of just out of college engineers as computer programmers.  I was shocked at that time how primitive their college computing environments had been as compared to what CMU/CIT had 10 to 15 years previous.  I then came to realize what a privileged environment we had at CMU/CIT when we were there.

Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Date:            Sun, 13 Apr 2008 04:35:54 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: Mark's web page

Hi,

I was never a G-21 operator so I didn't deal with the drapes.  I recollect one evening being frustrated at all the computers being down due to an A/C failure despite the sub freezing weather outside.  I also have a fond memory of smoke pouring out of all the computers due to a fire in the basement next to the A/C intake.

I do remember the G-21 card reader and console being just in front of the tape drives though.  The memory units were on the backside facing the railroad tracks.  You got to the G-21 area by going through the separate room that held the printers.

My days as a computer operator were all on the 360/67.

Roy
Date:            Sun, 13 Apr 2008 13:33:58 -0700
From:            Roy <r.engehausengmail.com>
Subject:         Re: Mark's web page

Hi J :-)

I thought the PDP-8's tape drives were blocked.

After my Army stint, I started with IBM in Manassas, VA.  Our project was some of the software for the Trident submarine.  The computers that were to be used on the ship were the Univac AN/UYK-7.  The tape drives were blocked as I remember.   They held the normal 1/2" tapes.  They were ruggedized for shipboard use.

Roy
From:            "David Chou" <dchouu.washington.edu>
Subject:         RE: Mark's web page
Date:            Sun, 13 Apr 2008 13:46:12 -0700

Chris,

You bring back some memories.  I did some programming for the G20 tape drives and remember the blocking.  There was a timing track on the tape as well as some unusual capabilities to handle skewing, an inherent problem with 1" tapes.   I do not know, however, whether CIT wrote programs to take advantage of this or whether the hardware just supported addressable blocks.

Of interest is that the blocking design allowed for the mapping out of bad blocks.  You could continue to use some really bad tapes, a useful feature by the end-of-life for the G20 when the tapes were no longer being made (thus my job as a student programmer).  Reliability was a big problem since the MT10 tape drives used mechanial arms rather than vacuum columns to manage tape slack.  During heavy search (and destroy) motions, the tapes really got a workout.

I have an old HP 35mm slide reader which I have kept attached to an an old Dell W2k computer.  For the most part, I have had pretty good success in adjusting the color on Ektachome slides.  Unfortunately, the process is slow.

Dave
Hi Dave (and All :-)

The female operator's name was Mary Noe.  She married Joel Bloom who I recall was a physics student (S68) working on a graduate degree, but that's a fuzzy memory.  I have a brief video of Mary and Joel (and a few others) at a comp center picnic.  Someone had tied some gas filled balloons in her blond hair.  If you have any memory of Mary, you can almost imagine her expression.  You're right Dave, she was a "pistol".  She was a fun person to be around.  A 20 year old CMU grad directory has them living down near NYC and he working for Chase.  Another incident involving Mary was that she bent over the back of the 1004 printer (on the 1108) to get some output.  If you recall at some point they added a "paper puller" to it which used rotating brushes and it caught her hair.  They got it stopped but her hair was vary entangled in the brushes.  The unit was detached from the printer and she carried it around on her head until they could get a UNIVAC CE there to disassemble it.  She said it had quite frightened her, but once again I can still see her wry expression as she carried the thing around...

Dave, are you sure it was the 360 which was dropped in through the roof?  It is my memory that it was the 1108.  I think I told in an earlier post about being one of the folks who disassembled the 1108 processor in summer 1971 for its transfer to Wean.  If you recall both the 1108 processor box and its memory box were long cabinets.  Very roughly 1/3 of this length was the power supplies and on the processor, this 1/3 was also the piece with the lighted pushbutton maintenance panel on it.  These had been delivered whole and thus the need to drop them through the roof, however, it was possible to separate them resulting in pieces which would fit in the elevator but this entailed some rather intricate unwiring of wiring harnesses which ran between the two sections.  It was my understanding that UNIVAC wanted more than the school was willing to pay to do this disassembly and so us CS dept engineering lab employee's were drafted.  It was my joy to be the one who took the processor box apart.  I always recall the older gentleman who was the head UNIVAC CE occasionally coming around behind the box where I was working to see how I was doing.  The worried expression was priceless ;-)

That same day as we were cleaning up and pulling cables from under the floor I found an old paper box under the floor with two reels of tape in it.  These had been the "on the machine" (not the master copies) of the IBSYS operating system and libraries for the tape operating system which ran the 7040.  These would wear out and occasionally get replaced, sometimes by shorting the tape beyond the used area and putting a new load mark on the tape and then copying from the master copies.  Someone had obviously put them there when the 7040 was removed in 1967 or early 68.  Thinking them "historic" I "acquired" them and still have them in my pile of junque to this day.  One is labelled S.SLB1 and dated 2/13/67 and the other S.SLB2 and dated 7/13/67.

And as to the Scaife elevator, it was quite the toy for certain bored adolescents like me and as I recall some others in this group.  We would get on top of it and take manual control of it.  But those stories are for another time...

J. Chris Hausler, JH37
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Hi,

I was a G21 operator prior to (and after) the arrival of the 360 and 1108. The fire event brings back a couple of events, many details I heard only second hand, so the details may be embellished.

Many of the 360 modules were too big to bring in through the elevator.  As a result, a hole was cut into the 4th floor ceiling and the computers were dropped with a crane.  In preparation for the 360, the 4th floor I/O desk and programming staff were moved to the third floor.  The rest of the 4th floor housed the 7040, located near where the 1108 was finally placed.  One day, I noticed that all of the CO2 fire extinguishers had disappeared. These were big units, about 4 feet high and very heavy.  One of the staff told me that the wood in contact with the cut steel plate caught fire after the workmen left.  The operator on duty was a thin and small 100 pound female (Mary N.) who I assume must have been really fired up with adrenalin. Fortunately, the room was mostly empty.

Some years later, I learned that IBM was upset over a roof leaking in one of the 360 CPU modules, resulting in some expensive repairs.

Incidentally, the Scaife Hall elevator was never very reliable.  I was stuck in it many times, often during campus tours for prospective students when the elevator was full. I should have surveyed to see if any of these attended Carnegie.

Dave
I remember that practically all the later machines were dropped in Scaife by crane.  The elevator was simply inadequate.  The 360 CPU unit was rather large and not easily broken down, especially since the 67 had a DAT unit attached as part of the console/CPU.

Dave

Email from Doug Jones


On Apr 22, 2008, at 11:16 PM, Mark DiVecchio <markd@silogic.com> wrote:

> Saw your web site. Here is a late 1960's 1108 Job Card from CMU. Sorry but its used (by me, MD05).
> I've also attached a CIT job card but the one you already have is probably a better example.
> I graduated from CIT/CMU in 1970.

I'm DJ02 from the same era -- I learned ALGOL-60 on the 1108 in Scaife Hall.  I may actually have all my old 1108 job cards. I seem to remember saving them.

And I remember taking a tour of the Athena.  Wonderful machine, with motor generators and pressurized nitrogen packaging for the logic.  I only visited the machine once.

My basement explorations were largely in Doherty hall, where my freshman roommate and I found the disassembled remains of Carnegie's old 11 inch refractor telescope.  There was a crane in the 4th subbasement high-bay where the parts were, so over a period of months, we assembled the telescope.  Several parts were over one ton, and there was no documentation at all, so we had a considerable amount of work to do to put it all together. Permission?  What permission.  We picked most of the locks in the building hunting for parts.  Some we found in attics and in crawl spaces under lecture halls.

Then, in spring 1969, someone managed to track us down and tell us to do two things:  Document the whole thing, then take it apart and pack it up for mothballing.  20 years later, it was donated to the Allegheny amateur astronomy club, which erected it in their observatory.

I graduated in 1973 with a BS in physics.

  Doug Jones
  jonescs.uiowa.edu
Amateur Astronomers' Association of Pittsburgh

John A. Brashear manufacturer of the 11-inch Brashear Refractor 
> Doug,
>
> I remember poking around in the basement of Doherty but I don't
> remember seeing a telescope.

It took us a while to recognize it as a telescope.  It was just a scattering of castings, sections of steel pipe, and whatnot. After assembly, it became something of a bit of folklore (a telescope in the 4th subbasement?  No way).  It was the same subbasement that held the (locked) entrance to the coal mine.

  Doug Jones
Date:            Wed, 23 Apr 2008 18:36:16 -0400 (GMT-04:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
Subject:         Re: Job Card Site

Hi Mark,

Yes, Doug is a great guy.  He and I had a discussion some years ago about the 1108 "sequence card", a blue card you would pick out of a card tray and put after your job card when you would submit a deck.  I have forgotten exactly for what they were used.  He has been the source of some and maybe all of the Smithsonian's card collection ;-)

Regards,
Chris

RJE and Teletypes

From: Mark DiVecchio <markdsilogic.com>
Sent: Apr 22, 2008 1:14 AM
Subject: RJE on the G-21

See attached scan from 1965.

IIRC there was a room on the 3rd floor of Scaife, middle in the back of the building, containing several different models of teletypes.

These must have been connected to one of the RJE boxes on the G-21.

Mark
Hi Mark,

There were two 35ASR's and two 35KSR's in that room, all with built in modems so to connect you would dial into the G-20.  You did this with a "dial card" (remember those :-) as the actual dials (remember those too ;-) were disabled to keep students from using them to "phone home".  Once the 360 was installed with its RJE system, dial cards were provided for it too.  For the G-20, however, the teletypes had to be in "full duplex" mode whereas for the 360, "half duplex" was required.  It seemed to me that IBM never did quite figure out how to properly do RJE terminal interfaces. 

There was usually a backlog of folks wanting to use one of these four TTY's and so you would put your name on the "blackboard" at the bottom of the list and as folks would finish (only one edit and run allowed per turn) your turn would eventually come up.  I use the word "edit" loosely as even though this was an RJE system it was batch.  The only interactive thing I recall available on the TTY's was "Desk Calc" (remember that :-)  You would enter batch commands to edit your AND file and then do a compile, load, and go and then had to wait for your job to come up in the job queue and run.  If you hung up the teletype, since the G-20 knew from which TTY you had dialed in, the "here is" drum was coded with its ID, i.e. "CIT Remote 20" and some "magic" characters, it would call it up and print out the results, usually a quick printout unless you asked it for a full one as you would still get line printer output.  Until your job ran, that TTY was unavailable for use by anyone else.

Adjacent (left) to that room was a larger "user room" with some tables and chairs for working and the user consultant's desk in the far corner and he had a 33ASR.  All of these teletypes and the others around the campus were leased from AT&T. 

On the other side (right) of the TTY room along the same outside wall of Scaife (facing Panther Hollow) were a couple of additional rooms with keypunches in them and a duplicator and interpreter.  A couple years later or so, the walls separating these areas were torn out and that became the enlarged I/O desk with both the UNIVAC high speed printer and one of the 1004's installed.  The user area was then moved (several configurations and I forget them all) to the other side of Scaife.  Up until some time in the late 90's, you could still see all the "damage" done by all these changes on the third floor even though Mech E had taken back the building after the comp center vacated it in 1971.  However, sometime in the last dozen years or so it looks like they finally rebuilt the interior, both the third and fourth floors, to its (more or less, I couldn't access all areas) original configuration.  So all "memories" of the comp center are now gone. 

Regards,
Chris

If you were editing an ALGOL source on the TTY's your last two lines typed in would be:
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
From: Mark DiVecchio <markdsilogic.com>
Sent: Jan 18, 2010 1:38 PM
Subject: Old RJE printout

So not like you guys don't have anything better to do, but....

I was looking at the old RJE printout that I have from the summer of 65 and it lists two of the G-21 operators:
ES06
JH09

Anyone remember who they were?

Mark
The name I recall that could be a match is Einar Stefferud.

Dave Rodgers
Hi Mark,

Neither of these ring a bell.  There was a John Hanyo but he didn't show up until much later, 68 at the earliest.  The only man numbers I have for operators are WZ02 for Bill Zippler and PG07 for Paul Gensler.  Both of these guys were lead ops in 66 so they may well have been there in 65.  I know this because I have print outs from the G-20's showing them as operators and I remember them very well, both good fellows.  (I also have one showing me as the G-20 operator from April 27, 1968 :-)  The girls I remember were Mary Noe, Yvonne Gartland, Nancy Barrett (briefly a girlfriend) and who could forget Doris Crabb who by 68 was the head of the operators.  Of course now I remember Ann Rindt but still don't recall the name of that girl in the photo even though I remember her reasonably well.  And anyway, none of these match your question.

73,
Chris Hausler
The name Jim McCrossin should ring a bell.

Roy Engehausen
Some users

MN01 - Mary Noe
DC05 - Doris Crabb
DC08 - me

Doris was one of my bosses.

Dave Chou
I thought Doris Crabb was DC09.

Bob  McFarland
I think you are correct.  Was DC05 David Cooper?

Dave Chou

CMU Chess Program

Joe Rubenfeld and Bob Walker wrote a chess program for the G-21. Here is a listing of the program dated 26 Oct 1968.

In December of 2022, I got these emails from Joe Rubenfeld.

From:    Joseph Rubenfeld <jrubenfeldacm.org>
Date sent:    Thu, 8 Dec 2022 22:15:34 -0500
Subject:    Hi Mark, still there?

I just came across your fascinating web site mentioning me and my chess program.

I remember your name, but that's all. Dave Rodgers was a classmate of mine in EE 1961-1964. I worked for Roy Weil at the computer center 1964-67. My chess program tied for first place in the world in 1980 and tied for first place in North America in 1984. Named CHAOS. If you respond and are interested, I will write a lot more history that all our fellow alumni may find interesting.

Thanks a lot and best wishes

Joe Rubenfeld ee '67
Joe,

You are famous!

I am still here ( and in my 70's) and I have a mailing list for all of the people whose names you saw on my CMU web pages.

The list is mostly people who were associated in some way with the Athena in the basement of Hamerschlag Hall. Dave Rodgers is on that list as are many others. But as you probably saw, it has expanded to include all of the other computer at CMU during the late 1960's.

If its ok with you, I will forward your email to that list. Let me know.

I actually ported your chess program to FORTRAN as after CMU, I didn't have access to a computer running Algol. It worked fine after the port but I didn't have time in my life then to put in the effort to really understand how it workd (if I even could in the  first place!).

I would absolutely be interested in whatever history you can write up. I will add it to the G-21 web page. Photos of people or computers are especially appreciated.
From:    Joseph Rubenfeld
Date sent:    Sat, 10 Dec 2022 22:41:46 -0500
Subject:    Re: Hi Mark, still there?

Born in Pittsburgh, raised in New Castle Pennsylvania. In 1961, I started at Carnegie Tech majoring in electrical engineering, which today is ECE. In my sophomore year I overloaded a computer science class and my junior year I took more computer science courses and fell in love with programming. I realized that I did not want to be an electrical engineer. After that I took a two-year hiatus and went to work for the Carnegie Computer Center, as a programmer.

During that time I met Bob Walker, who was then a chess master. He had a degree in operations research and was working for GSIA which today is the Tepper School. He was responsible for running the Carnegie game. In Dr Feldman's compiler-compiler class I started writing a new language for writing chess programs. However I soon gave up on that and translated what I had written into a chess playing program. I wrote a little article for the Carnegie Technical magazine on computer chess.

I did all the programming, but Bob Walker gave me the chess expertise especially in designing the static evaluation function with a dozen factors and their weights. In those days chess was a very popular topic with a number of Herb Simon’s PhD students working on various aspects of programming chess.

My program filled a box of 2000 punch cards.  When Dick Shoup wrote the AND system, I put the program on the disk and made it available for whoever wanted to play. I think Roy Weil was the one who played it the most.

Then in about 1966 Ira Rubin, who was getting his master's degree in electrical engineering and going to work for RCA, as a number of our friends did in those days, asked me if it was okay to take my chess program with him.  I agreed and he did take the program to RCA. He put together a small team and translated the program from ALGOL into Fortran. They started entering the program into computer chess tournaments and were doing pretty well. But Ira Rubin forgot about me.

After my 2 year hiatus I went back and finished off my senior year, graduated, and started a master's program in computer science at Pitt. Meanwhile chess programming was becoming more popular. Rubin and his team renamed my program CHAOS. They also got more chess knowledge, added book openings, and another half a dozen factors in the static evaluation function. They did pretty well. They were now working for Univac which had bought out the RCA computer division. They were coming in second place in a number of tournaments. Eventually in 1980 Chaos tied for first place in the world among chess programs. Then in 1984 Chaos came in a four-way tie for first place in the North American computer chess tournament.

During this time I had lost touch with Ira Rubin and the team that was working on the program. I was working in New York as a programmer. Bob Walker, by the way, got married and had a couple children living in New Jersey and then left his wife and kids and moved to Las Vegas Nevada to become a professional gambler.

As artificial intelligence was becoming more popular in general especially in the form of expert systems, in 1984 I was recruited to become the founding president of a company specializing in expert systems development (Applied Intelligent Systems, Inc.) by a new venture Capital firm called Computer Science Capital. We built what was considered the leading expert system for life insurance underwriting for John Hancock Mutual Life Insurance Company in Boston. Based in part on our expert system that we built for them, they were then acquired by Manulife the largest insurance company in Canada. I went on to work for the City of New York where I rose to the level of deputy commissioner for advanced technology and telecommunications for the city government.  I introduced the concepts of artificial intelligence, expert systems, and Knowledge Management. My team at the City built a system called TowIT, that ran on newly developed handheld computers that allowed the City Marshals and Deputy Sheriffs to enter the license plate of a car as they drove down the street and it would immediately tell them that license plate indicated that the owner was delinquent in paying parking fines and allowed them to tow the car. That system brought in about $50 million a year while significantly reducing the billion dollars in parking fine debt that the City had on their books.

At some point I realized that the Chaos team had no recollection of me having written the original program. I contacted Ira Rubin who had totally forgotten about me.  We were never exactly friends, only acquaintances, at Carnegie Mellon; but after a little discussion he agreed to send me the latest documentation on the Chaos program.  After studying it, it was clear to me that the bulk of the chess program was what I had written and the improvements were in the chess knowledge. They had added, for example, six new factors to Bob Walker's evaluation function.

I thought of chess programming many times over the years, thinking that if I ever get to retire, I would write an even better program. But unfortunately even though I did get retired and moved to Florida (the promised Land), I got involved with other interesting and important projects and never had the free time to get back to chess. In the meantime as you probably know this guy David Silver developed what I believe is a major, major breakthrough in machine learning that they call RL or reinforcement learning. Using that approach they were able to beat the world champion Go player as well as any human chess players. Most recently I believe they have now been able to use RL to even invent new algorithms for solving many different types of problems.  I believe this will have significant advances in many different areas and really make artificial intelligence a significant part of the industrial and business worlds.

All the best. Thanks for your efforts in keeping our memories alive.

Joe Rubenfeld
JR04

PS Watch for my new book in the near future on Amazon entitled Choice or Chance, How Science is changing the narrative on Evolution, Materialism, and Life
From:    David Rodgers <dave_rodgersmsn.com>
Date sent:    Sun, 11 Dec 2022 18:07:54 +0000

Mark,
How great to hear from Joe Rubenfeld and his contributions to AI.
Dave Rodgers

G-15, G-21 and PT-10 Paper Tape Punch


Date:    Tue, 18 Sep 2007 18:23:44 -0400
From:    David Vavra <davavraverizon.net>
Subject:    RE: Details on Website

I wrote the assembler [for the Athena]. It ran on one of the Comp Center machines but I can't remember which one. Probably the G-21. Whatever it was, it had a high speed paper tape punch. The tape had to be hand wound onto a reel that the Athena used.
John Yurkon replied:

Hi Dave,

It's been a long time.  I heard from Charlie a number of years ago, to my surprise.  I don't know how he tracked me down.  Search engines were just starting to mature.

When I left CMU the transports on the paper tape drives were working but they were still trying to get them to read properly.

I thought that the Bendix G-21 was being decommissioned back then.  They were having a difficult time finding replacement germanium transistors and the disk storage looked kind of scary.  The platters were bouncing up and down about 1/8" if memory serves.

I don't know of another machine that had a paper tape punch though.  There was a water cooled monster that I never new the name of and there was an earlier Bendix G series that had vacuum tubes and was about the size and look of a refrigerator.

The Bendix G-21 and the Philco display units were what tempted me to get into trouble academically.  I survived after a few years in the Air Force and ended up in the Honors Tutorial Program in Physics at Ohio U.  Their computer facilities at the time were primitive enough to keep me out of trouble.

I know this group is about the Athena, but if anyone has a timeline for the history of the Univac 1108, the IBM 360 (mod 60 I think) and the Bendix G-21 at CMU I'd be very interested in hearing it.

Come to think of it, weren't there some PDP-8's that had paper tape drives? Could that be what you used?

John
Dave replied:

Hi, John,
Long time indeed!
 
The refrigerator sized machine was the Bendix G-15. Don't remember the water cooled one. The Bogart, maybe? It was in the same room as the G-15 IIRC. That used 6146W tubes in its flip-flops. I had a Drake 180W transmitter that used 6146A's just to give you an idea of the possible power draw. I get the impression from some of the web sites that the Bogart and Athena were somehow cousins.
 
Got me in trouble too. I worked at a market research firm in Pittsburgh run by an ex-GSIA dean for a while and got to do some nifty things unrelated to market research like stadium scoreboards and the passenger signs for BART in San Francisco. Almost became Hausler's competition by working at Union Switch and Signal but took a small sabbatical and visited Southeast Asia instead. Came back in 73 to finish things up. Pat Stakem got me fixed up with a small NASA contractor here so now I'm designing and building flight software systems for spacecraft and coming up on 33 years.
 
Funny how things work out.
DAV
Date:            Sat, 22 Sep 2007 11:57:44 -0400
From:            jchausler <jchauslerearthlink.net>
Subject:         Re: Details on Website

Hi All,

"John E. Yurkon" wrote:
> I thought that the Bendix G-21 was being decommissioned back then. They
> were having a difficult time finding replacement germanium transistors and
> the disk storage looked kind of scary.  The platters were bouncing up and
> down about 1/8" if memory serves.

I'm not 100% sure but I believe the G-20's were removed in summer 68.  I started working as a part time operator spring semester 68 and spent several shifts that spring operating them.

Those large platter disks were impressive.  Remember also the magnetic card memory system, the RCA RACE (anyone recall what RACE stood for?)  I remember when working the G-20, stopping the RACE units (there were two "readers" and the controller was an RCA 501) to clean the heads and drum around which the cards were wrapped to read or write them.  In 1970 or 71 while working for the CS department engineering lab, I had cause to be pulling floor boards in the room in which the RACE had been installed.  In a corner, I found a brand new RACE card, still sealed in its envelope. I'm assuming someone had dropped it and it had gone "flying" and they didn't dig for it.  It was in a far corner. The machine recognized which card it was by a pattern of "fingers" cut into the long sides of the card.  I believe looking at this one (I opened the envelope :-) that there were 128 individual patterns.  I believe this is how the unit would know which card to select out of a specific bin, I'm assuming the "bin address" was coded magnetically on the card.  And yes, I still have this card and its envelope too :-)  A favorite story of mine about the cards is that one of the Philco CE's told me he would take the used cards and use them as hinges on his "dog door", as they were made of rather stern stuff, necessary for the rough handling the card received. (Remember the debacle IBM had with the "Datacell", the cards for it were too flimsy.  And yes I have a couple of Datacells in my stash I picked up in the mid 70's :-)

Speaking of the G-20's, in the early 90's there was a 25th anniversary of the CS department.  I could not attend but was later told by someone who attended that he heard that that the G-20's still existed at that time at least, in private hands. Originally they had been sold to an individual, Carl Lefkowitz (and I may be mangling his name).  In fall 1976, Pat Stakem, Russ Moore and I first came back for homecoming.  We learned where Carl had the machines (in an old Bank building) and were staring through the front door at the machine when Carl showed up and invited us in.  The big disks were there but he was trying to interface the machine to a pack drive which looked much like an IBM 2311 type.  I think he was intending to try and sell time on the machine.  For the last dozen years or so I have maybe attended about 1/3 of the homecomings and on a couple of these occasions in the late 90's tried to locate Carl or the machine with no luck. Anybody know anything?

Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$

Date:    Tue, 19 Nov 2013 19:25:52 -0500 (GMT-05:00)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    PT10 manual

Hi Mark,

Lacking common sense, I just scanned the PT-10 manual. I also scanned three images of the loose leaf binder in which it is held. Those three images are attached to this email. I will send the "guts" in a following email as it is quite close to my max allowed email size. Everything was scanned at 100dpi to keep it small. I scanned some of the images at 75, 100 and 300 dpi grayscale and although I could see differences between the 75 dpi and the 300 dpi, the differences were not much noticeable between the 100 and 300 dpi images so I scanned the document at 100 dpi grayscale to keep it as small as possible. The binder was scanned at 100 dpi color (obviously :-). I don't know where you would put this as it fits under both the G-20 and the Athena. I thing I told the story before but...

Dave Vavra wrote an Athena cross-assembler which ran on the G-20's. To transfer the result to the Athena he used the G-20's paper tape punch, the PT-10 to punch the tapes. Likely he used this manual to figure out how to do this and as no one else was using the PT-10 at that late date, he got the manual. Well, a year or two later, after the G-20's had been removed, I walked into the Athena room one day and saw the manual in the trash. Someone was obviously cleaning up no longer needed items. Well, dumpster diver that I am I couldn't bear to see such an historic item lost to the future and retrieved it out of the trash. And here it is...

73, Chris Hausler

Click on the image to download the 18MB manual.
Chris,

The manual talks about an IPL mode.

Was the G-21 IPL'ed from paper tape or had they worked out a more "modern" technique?

Mark
Hi Mark,

To my knowledge, they IPL'ed the G-20's from the disks.  But like most computers, one could likely IPL from just about any device directly connected (I IPL'ed that PDP-8 in Bell's anti-office from the paper tape reader on its console ASR33 a number of times before they got the high speed reader later that fall or early spring 67, not to mention using an ASR33 reader to load my Motorola 6800 D1 kit a decade later, and I still have the paper tapes for that :-).  That said, I would suspect that if they ever IPL'ed from the PT-10 is was rarely.  I have no idea how system software was distributed back then.  Further, I think most "system software" for the G-20's was actually developed at CIT.  Certainly paper tape was popular with minicomputers into the mid to late 1970's and at least early in the 70's I would boot the DG minis at GRS and customer sites from paper tape although using a high speed reader, not the teletype reader.  I have a Data General high speed paper tape reader from the late 1970's which the BNSF Railway folks gave me in the late 1990's which had come off of a system installed in the late 1970's but it shows little use.  Likely it was customer specified as they commonly were for minicomputer systems back then but not really needed.  Most first and second generation mainframes, including IBM's, had high speed paper tape reader/punch units available for them but they were rarely provided with systems ordered by the early to mid 60's unless the application specifically required their employment.

The PT-10 manual copy you have has a couple of issues.  Sort of halfway through my effort to scan it I got a phone call and dropped a page (page 93) as by that time I was already numb with the effort.  There are also several pages which had fold-outs and so required multiple scans to get all the data.  Follow-on review of the scan shows that a couple of these didn't go real well.  I don't have the stomach to scan the whole thing again and the software I have doesn't allow editing of .pdf's.  However, one of my telegraph friends who lives in eastern PA has something called Adobe Pro 9 which apparently can do this and is willing to give it a go so if and when I get a correct copy I will email it to you...

73, Chris Hausler

Forever being the computer sleuth, Chris Hausler sent me this in Feb of 2023:

From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    G-20 disk drive
Date sent:    Thu, 23 Feb 2023 18:36:32 +0000

Hi All,

I've always wondered who manufactured the large disk drive that was on the G-20's  The cabinets don't match the G-20 cabinets and some years ago I communicated with a man who was involved of disk drive development for CDC in the mid 60's and he had no memory of anything resembling the G-20 drive.  Well, on my way to something else I came across a photo of a drive manufactured by GE in that time frame which looks a lot like the G-20 dirve.  Here's a link to the photo:  http://www.ed-thelen.org/GE-MassRandomAccessDataStorage-.jpg . 

If you look at the Carl Lefkowitz page on the Athena web site (  http://www.silogic.com/Athena/Carl%20Lefkowitz.html ), the first photo on that page shows the drive on the left and the second a better shot showing the disk controller rack on the left of the string.  It certainly well matches the below image.  Mystery solved?  You be the judge...

73, Chris
From:    David Chou <dchouuw.edu>
Subject:    RE: G-20 disk drive
Date sent:    Thu, 23 Feb 2023 21:55:20 +0000


I somehow remember that the disks were made by Dataproducts.  The Wiki describes disk drives that were OEMed to GE which explains why they looked identical to the ones I remember.  The printers on the G21 were also made by Dataproducts so clearly Bendix had a working relationship with them.

Dave

https://en.wikipedia.org/wiki/Dataproducts
From:    Charles Putney <chputneygmail.com>
Date sent:    Thu, 23 Feb 2023 22:25:02 +0000
Subject:    Re: G-20 disk drive

David,

I worked for Dataproducts from 1980 to 2008 but by then we were part of Ricoh. They were based in Los Angeles, but had manufacturing in Dublin, Ireland. When I joined, we manufactured OEM band printers and drum printers. At that time there were still a few core memories being repaired.   After daisy wheel printers, dot matrix printers, we finally we moved to small laser printers and then large laser printers (92ppm).

Charlie
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    RE: G-20 disk drive
Date sent:    Fri, 24 Feb 2023 02:09:17 +0000


Hi Dave,

I agree with you.  I came across a long interview with Erwin Tomash one of the founders of Data Products.  As the Wiki entry mentions they acquired the disk drive business from Telex.  Here's a short section of the interview about the disk business:


"We had to quickly assess the problem with the disk drives in St. Paul. It became evident that they'd made a mess of the electronics and that the logical system was just a jumble. They hadn't really had a decent logic designer at all. So Irv Weiselman, a proven logic designer from Ampex, was hired because we needed someone to sort that out. We also felt the need to get somebody to handle GE in Phoenix. We needed to hold their hand, and keep them in the family until we could get the product fixed. What was doubly worrying was that they could cancel the contract and retain full manufacturing rights. They didn't really want to cancel. They didn't really want to manufacture. But these disk units were essential to their program. If they lost faith in us and felt we couldn't fill their needs, we would drive them into the disk business, disk drive them into the business. What we needed to do was establish and maintain their confidence. We picked another Ampex/TMI person by the name of Jack Ogg and asked him to join us. I think he was our seventh or eighth employee. He went to Phoenix to keep things calm. At the time, we didn't know if we could fix the problems because we didn't know what was wrong yet.

That was the start of Dataproducts. We ultimately succeeded in keeping GE-Phoenix as a customer. We built a few hundred big disk files for them. They used big, huge platters for disks. They were not as large as this 4-foot table, but I think they were 32 inches across. It was like the early IBM RAMAC. We ultimately built some even larger units. I saw one of these a couple of months ago in the Science Museum in London. We shipped a huge unit to Cambridge for their Atlas machine, and it is in working order at the Science Museum today. It's on display there."

It mentions a disk diameter of 32 inches. I came across another reference that said it was 31 inches.  My recollection is that they were certainly larger than two feet in diameter.  The article does mention a few other folks to whom they sold the drives but not Bendix or CDC.  I wonder if CIT bought them directly.  Remember that the G-20 system at CIT was sort of a "Rube Goldberg" assembly of stuff quite advanced for its time.  In addition to that disk drive there was the RCA 301/RACE with the interface, I believe I heard was built at the school, and the teletype interface which I heard was built by Carl, not to mention the whole G-21 64 bit combination operation.  And most of the G-20 system software was also written at the school.

73, Chris
From:    David Chou <dchouuw.edu>
Subject:    RE: G-20 disk drive
Date sent:    Fri, 24 Feb 2023 08:10:06 +0000

Well, I don’t remember the G21 disk drives failing much, if at all, during my years at Carnegie.  I assume that the problems had pretty much been ironed out by then (or maybe it was because so many other parts of the G21 broke more often).  I can’t remember the exact numbers, but the available spinning storage was minuscule (in the megabytesᾠrange) and most of our AND files ended up on tapes or on RACE.  Another thing that struck me was how big they were compared with the IBM 3330™s or the 2714™s.

I wouldn’t be surprised if the disk drives were purchased by CIT and added on; almost everything in the G21 was kludged/customized.   This practice continued onto the IBM where the IBM ECS memories were replaced by AMPEX.  At least one bank of the ECS (both IBM or AMPEX) memories seemed to be down at any time.

Dave
Date sent:    Fri, 24 Feb 2023 08:22:27 -0800
Subject:    Re: G-20 disk drive
From:    Roy <r.engehausengmail.com>

I did a lot of programming and operating on the 360/67 with MFT, MVT, and TSS.  Even ran PTSS for a bit.  I don't remember LCS failing very much.  The IBM 2361 was a standard product.  I do remember that the "underlap" feature was turned off so maybe that helped.  The IBM "Storage Channel" feature was also found to be useless.  The IBM unit only had 2MB while the replacement ones were 4MB.

The IBM 2321 data cell at CMU was the biggest disaster.  When I was working at IBM in San Jose where the datacell was engineered, it was never mentioned in mixed company :-)

Roy
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    RE: G-20 disk drive
Date sent:    Fri, 24 Feb 2023 17:37:11 +0000

Hi Dave,

I did not see this and don't know when it happened but apparently although the heads were kept above the platters by air pressure, they were not "flying heads" as were the IBM pack drives like the 2311's and 2314's (and later ones by other vendors) I saw.  I was told that there was an air compressor supposedly down in the basement of Scaife which provided the pressurized air which kept the heads from crashing.  Well apparently, one day it failed and I recall hearing that the windows which allowed one to see the disks immediately turned brown with the scraped off magnetic coating.  That said, that is the only failure I ever even heard about.  As I recall the drive stored somewhere around 38 MB.

This compares to between 300 and 400 MB held by the RACE drive.  Although the individual RACE cards apparently held slightly less than the individual Datacell cards (or "strips"), each RACE bin held 256 cards compared to the 200 cards held by each Datacell.  However the Datacell drive held 10 cells compared to the 8 bins in the RACE drive.  So 2048 cards in a RACE drive and 2000 cards in a Datacell drive, so both held about the same amount of data per drive.  This was of course the reason for these units, they held roughly an order of magnitude more data then contemporary disk drives but of course disk technology advanced so quickly and mechanical complexity of the magnetic card units and their slower access speeds (much faster than tape of course :-) doomed them to be a footnote in history.

73, Chris
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    Re: G-20 disk drive
Date sent:    Fri, 24 Feb 2023 18:01:47 +0000

Hi Roy,

When I visited the campus for the first time after graduating for "homecoming" fall 1976 with Pat and Russ I vaguely recall the 360 had three Ampex LCS units and no IBM LCS units but I am unaware of their storage capability or access speed when compared to the IBM units.  I recall hearing that the IBM LCS units were slower than main 360 core.  An of course this was the same visit where I found two complete Datacells in a trash bin and saved them.  Despite all the stories about their inability to refile the cards in the cells, it appears this was never a problem with these two cells as they appear pristine.  That said although apparently a Datacell controller could handle 8 drives there was only one drive I ever saw at CMU and it didn't last long either.

I also don't remember ever hearing that we ran MVT at the school, only MFT (and of course TSS as well as TOS, BOS and DOS early on) but I could easily be wrong as I rarely ran the 360 while working as an operator at the school.  Those two summers working as an operator here in Rochester on 360's at Kodak we ran only MVT.  I remember returning to school one fall from one of these summers and demonstrating to "Skip" Grandy how one could kill off operation from a user punch card command (so much for a "friendly user" environment :-) and since this was something that worked in MVT at least maybe we were running it (the command was "//<space>p>space>init" something he quickly disabled :-)

73, Chris



The Hay Balers ???

Emails from Russ Herman.

Date:            Thu, 2 Dec 2010 14:40:17 -0800 (PST)
From:            Russ Herman <russ_hermanyahoo.com>
Subject:         CIT Comp Center 1960-66 (part 1)

Stumbled across this site during a bout of insomnia. I predate the other contributors (left the Comp Center in '66), and my cred is my identity as RH01 (SM03, DV03, and CR03). I should be able to clarify some of the conjecture about The Way Things Were. I was involved with the Comp Center from 1961 as student and staff, and was present for the move to Scaife Hall, the acquisition of the first G-20, then the second G-20, and their combination into the G-21. I was also the developer for the RCA301 and the RACE units (also known as the hay balers),

Acronyms. Yes, the naming could get convoluted. My favorite is the evolution of the algebraic compiler that was used before ALGOL-20 became dominant. It was derived originally from something called Generalized Algebraic Translator (GAT, a Fortran-like language in many respects) some time and some place else. The version in use when I started using the G-20 to model a chemical reaction for a presentation in Chem Seminar my sophomore year was known as GATE, from GAT Extended. The final branch of this tree was referred to as PILE. PILE you ask? Not for the reason you'd assume today. Actually, it was originally going to be called NewGATE, until someone realized that Newgate was a Prison In London England. Punning was a sport in that era. Other origins include AND (Alpha Numeric Directory) and THAT (To Help Assemble Translators).

Dave Fisher bootstrapped me into computing with that 1st program for my seminar. That is, he wrote (literally) it down, handed to to me bugs syntax errors and all, and said "Now it's yours." I think it was originally for the G-15, which was the 1st machine I pregrammed. (There was an IBM650 in the basement of GSIA that predates me, but this wasn't generally available. Suits have been like that forever.)

Student employment at the Comp Center was meritocratic. I started programming little useful utilities, which got me to peoples' attentions, and then applied for a job as a computer operator. Only problem was that I had never taken any Comp Sci courses, normally a prequisite. However, I offered to take CS-208, the 2nd semester project course, since I'd already mastered the contents of 1st semester (GATE or ALGOL, whichever was in play), and had no interest in IPLV (Newell, Simon, and Shaw's list processing language used for AI projects that got shouldered aside later by LISP). S207 and S208 were projects courses, so I dove into being part of a project team attempting to develop a COMIT implementation. We never got off the ground which in retrospect was no surprise, considering we were attempting to implement a string-processing language using a framework (Jerry Feldman's PhD project) intended to develop algebraic languages.

After paying dues as a computer operator carrying boxes of punch cards and separating outputs and returning them in folders to the wooden shelving some else mentioned in passing, the next step on the ladder was User Consultant, aka User Confuser. The good-size Keypunch Room contained a number of tables and chairs where users could examine their output and make corrections before resubmitting their jobs. A staff member was scheduled to be present to provide assistance to the baffled. This was a prestigious role. As an undergrad, you were helping PhD candidates complete their theses.

The next rung was summer employment as a developer. Staff and profs were writing software for operations (as opposed to their research projects), and positions were available. When you got this far, you were pretty well hooked on the field of computing, no matter what your undergrad major was.

After summer employment following my degree (barely squeaking through due to all those hours spent at the Comp Center), I attempted to continue part-time while enrolled in an MS Pharmaceutical Chemistry program at Duquesne. But my boss told me it was time for me a make a choice between chemistry and computers. After six weeks of grad school, I dropped out and went back to Carnegie as a staff programmer.

Now that I've nattered at length about my personal connection, further postings will explore hardware and software of that era, and name-drop various professors and staff.

RH01

PS The link for the whatever happened to Carl Lefkowitz is bad (Mark's note - fixed). That was actually how I got to the site. He crossed my mind so I Googled him.
Date:            Fri, 3 Dec 2010 12:57:53 -0500 (GMT-05:00)
From:            "J. Chris Hausler" <jchauslerearthlink.net>
To:              Russ Herman <russ_hermanyahoo.com>
Subject:         Re: CIT Comp Center 1960-66 (part 1)

Hi Russ,

Good stuff!  Keep it coming.  I really want to hear about the "Hay Balers".  I don't recall ever hearing them referred to as that.  I started my "operator career" when a sophomore, spring 68, which was the last semester of the G-20's and am the one with the RACE cards in my collection of junque.  There was a story that the same guy who "invented" the RACE also either before or after "invented" NCR's "CRAM" (card random access memory) which I saw at their pavilion at the 1964 NYC World's Fair.  It was rumored that this same guy designed yet a third magnetic card memory for yet another company (not the IBM Datacell) but I've never been able to verify any of this.  Any ideas?

73, Chris Hausler
Date:            Mon, 6 Dec 2010 00:08:25 -0800 (PST)
From:            Russ Herman <russ_hermanyahoo.com>
Subject:         Re: CIT Comp Center 1960-66 (part 1)

Hi Mark.

Surprised and saddened to hear of Carl's demise. I remember him being a bit of  an adventure junkie. There was a small group of people who went caving, and I  think he was part of that. I also remember him trying skydiving.

He was half of the team that did most of the work developing the G20 monitor. The other half was Jim McIlroy, who I recruited to RCA Computer Division (along with a few other Comp Center staffers shortly after I went there). That's probably why Jim's name hasn't popped up yet in the reminisces. I'll expand a bit later.

I'm trying to contact another friend and roommate from that era to let him know about the site and invite him to contact you. He predates me by 1 year and remained there awhile after the multiple departures to RCA.

Russ


More emails:

From:    J. Chris Hausler <jchauslerearthlink.net>
Subject:    Re: Happy New Year Athenians
Date:    Fri, Jan 2, 2009 at 12:53 PM

Hi Mark,

Happy New Year to you too!

I'm attaching a Christmas photo I recently sent out to some of my railroad and telegraph friends.  It shows my Christmas tree and the model railroad I put around it each year (called the BF&R, "ballast free and roundabout", carpet makes a terrible road bed).  (And you'll probably recognize the big Santa dude in the back :-)  However, you might note some interesting things in the photo.  Hanging from the tree more or less dead center down low is a G-15 flip flop module.  There's also one of those B250 cordwood modules hanging from the tree but obscured by branches.  To the right of the tree on top of some books is my G-20 tape 88.  Surrounding it is some CMU homecoming junk I've collected over the last 15 or so years.  Also hanging on the tree is an ornament made for me back in the late 60's by the comp center secretary, Linda Hammond and a souvenir ornament of the Duquesne incline acquired in the mid 1990's.  Those are the only things I believe to be CMU or Pittsburgh in photo.

Anyway I could take photos of the modules and the tape if you want and email to you.  I still haven't located those G-20 photos (slides) I mentioned some time ago and even if I did I have no way of making copies of them.  I've been meaning to investigate how to get this done locally but with work and dealing with the cancers the last two years (One of my New Year's hopes is that I won't have to renew my membership in the cancer of the year club this coming year :-) have just not allowed me the time.  I also believe I have an 8x10 B&W photo of me sitting in front of the G-15 in HH53.  As I recall it as taken by Pat with my camera.  I've seen the 8x10 sometime in the last 10 years (and I likely have the original negative somewhere too but haven't seen that in decades ;-) which once located I could probably get a copy of and email it to you.  Let me know if you are interested.

You can see a lot of other computer (and telegraph) junque in the photo.  My pack rat inclinations come naturally and are probably genetic.  Although my father died in 2000 and my mother in 2004, my sister and I are still going through their stuff (we both doubt we will live long enough to complete the task).  I bring this up because just a couple of weeks ago while we were looking through some of my father's stuff I came across another RACE card.  Now I had that one, the story of which I already told, but have no memory of this one.  Since I don't know why or where my father would have gotten it all I can assume is that I must have purloined it earlier and then forgotten it.  So anyway, I could take a photo of the two RACE cards and send it to you as well.

If you really want it I could send you the Comp Center "Red book" users manual from the mid 1960's if you want to scan and copy it.  I would eventually want it back.  I think I have a scopes manual around here somewhere too which I could mail to you under the same arrangement.  At the moment I don't know of any other G-20 stuff I have.  Now I do have other 1108/360/7040/PDP8/PDP9/PDP10 stuff, mostly manuals from back then.  There's also some DECTAPES and those last two system tapes from the 7040 in my stash.

It goes on...

73,
Chris Hausler

Date:    Sat, 31 Dec 2011 11:33:14 -0500 (GMT-05:00)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    Keypunch...

Hi Pat, Mark and Russ,
I just came across the following site and I thought you all might enjoy it. All it's missing is that pleasing crunchy sound every time you hit a key.
http://www.kloth.net/services/cardpunch.php
73, Chris Hausler


Mary Shaw and Roy Weil


From: Mark DiVecchio <markdsilogic.com>
Sent: Feb 19, 2010 7:05 PM
Subject: Mary Shaw and Roy Weil

In going through my old junk, I found that I had a copy of "Freewheeling Easy in and around Western Pennsylvania".

I bought it at a thrift store here in California. I never paid attention to the authors before but they are CMU computer department alumni - Mary Shaw and Roy Weil.

http://shaw-weil.com/fwe/
http://www.post-gazette.com/lifestyle/20030517rocs0517jp4.asp

Mark

RCA 3488 RACE


Date:    Mon, 2 Apr 2012 13:51:01 +0000 (UTC)
From:    foxbwcomcast.net
Subject:    RCA 3488 RACE

HI Mark,

I ran across your web site while googling the 3488 RACE machine.  In a past life I was a CSR (Computer Service Rep) for RCA Computer Systems, 1968 - 1972, working mainly on the 3301 Realcom system.  I had the rather dubious privilege of maintaining two of those devices.  It was a sort of love / hate relationship; quite a marvelous hydraulic / mechanic / electronic masterpiece of engineering and a tempermental and frustrating beast at times.

I had to chuckle over the part of your site where the card being ejected from the R/W head assy went sailing across the room to dent a cabinet.  One could do the same thing in a much safer manner by loosening the tension screw on the single card extractor chute at the rear of the machine.  The correct tension would stop the card right at the exit slot, but if you loosend the screw enough, it would sail right through that slot at warp speed, pretty cool.

I kept my maintenance manual and one card when I left RCA (actually UNIVAC as they had bought the firm) and still have it.  It is quite thick, 2/3 inches and weighs a few pounds.  I would dearly love to find some one who is trying to get one running again, and either help or donate the book.  I doubt if any exist.

BTW, they ran on RCA 301s, 3301s and also the Spectra 70 line, and RACE stood for Random Access Computer Equipment.

Regards,
Barry Fox
Sterling, MA
Barry,

Thanks for your email. I forwarded it to our small group of  1960's techno nerds...

We may complain a lot about the RACE unit but back then it was a very good way to keep offline storage that was still pretty fast to access. Much better than magnetic tape. The RACE system was the backbone piece of hardware that let us replace reading in boxes of punched cards with program source on them.

May I use your email on my web page?

Mark
Mark,

Yes, indeed, in its day it was the fastest way to retreive one record out of a zillion without searching 2400 feet of mag tape. We kept the entire state of Mass unemployment records on one and the other as a backup.  All day long people would go through the Division of Employment Security remote sites to get their benefits.  At the end of the day, the tape that was used to record these transactions was used to update the individual records on the RACE machine, using a program called "DAMU", daily maintenance update".  That ran the RACE machine at full speed for 3 / 4 hours straight, usually resulting in some spectacular card jams.

Sure, you can put my ramblings on your web page.

Barry

Hi Barry,

Mark forwarded your email to the Athena Group. I enjoyed reading your experiences with it. I never had a bad experience with the extractor chute, the card would always just stick out about an inch or so. When this would happen I would examine the card and if one of the fingers wasn't completely broken off (just folded over) I would "fix" it and stick it back in another slot which ran the card down the front of the drive and wrapped it around the read drum, then shot it back down the back of the drive and filed it back in it's proper place (usually ;-). Of course this was just delaying the inevitable :-) Again I believe the RACE drives I "operated" in spring 68 were not the first at the site as there was a deflector arm on the read head assembly which snagged the card as it came out of the box and directed it into the drum. As I recall the head assembly was held in by a single (or maybe it was two) jackscrew and loosening that tripped an interlock which prevented the drive from selecting a card. The earlier drives had this deflector as part of the plexiglass cover and opening that did not trigger the interlock (as it didn't with the units I serviced as well). I had to remove the head assembly once a shift in my job as part time computer operator to clean the drum with alcohol. Of course I also always remembered to first go over to the 301 and halt it :-)

I also don't believe that the units I used were the last design for the drives as some advertising I think Mark's site points to shows that the bins containing the cards were user removable. They were not on the units I serviced. This might have been an advancement to compete with the IBM Datacell which did have removable bins (I've got a couple :-) One funny story you might appreciate. The service engineers who maintained our RACE and the G-20 processors worked for Philco. One of them once told me that he took failed RACE cards and used them as hinges for his dog door as they were made of relatively stern plastic. They are a lot more rugged than the Datacell cards, the plastic is thicker. I assume you saw the photo of the two RACE cards I have on Mark's site. If you have any further information about the RACE, in particular details about the various versions of the drives, I, at least, would certainly like to hear about them. Despite it's problems, it successfully provided the major on-line storage for CIT's main computer system for a number of years. It was certainly more reliable than the Datacell.

73, Chris Hausler
From:            "Dave Chou" <dchouuw.edu>
Subject:         RE: (Fwd) RCA 3488 RACE
Date:            Mon, 2 Apr 2012 08:55:22 -0700

Wow - this is fantastic.

As a G21 operator, I somehow earned the dubious privilege of doing the monthly backups on the RACE units on third shift.  Most of the time, I could take a snooze, but when the monthly backups occurred, it was a constant marching between the RACE unit, the 301 console, and the G21.  I am not sure I would want to put the card into service - too many cards ended up as an accordion during the backup.

Dave

From: foxbwcomcast.net
Sent: Apr 8, 2012 2:56 PM
To: Dave Chou
Cc: Mark DiVecchio , Bob McFarland , Charlie Putney , Dale Dewey , Dave Rodgers , Dave Vavra , David Chou , Glenn Sembroski , "J. Chris Hausler" , James Pollock , John Yurkon , Lauston Stephens , Patrick Stakem , Roy Engehausen , Russ Moore , Tom Engelsiepen
Subject: Re: (Fwd) RCA 3488 RACE

Hi Guys,

Wow, had no idea there were so many RACE machine people out there.

I dug through the manual and found some specs for those who are interested:

7 bits per character
650 characters per block
166,400 characters per card
256 cards per magazine
42,598,400 characters per magazine
8 magazines per 3488-1
340,787,200 characters per 3488-1 (minimum system)
16 magazines per 3488-1 / 3488-2 (-2 is expansion unit)
681,574,400 characters per 3488-1 / 3488-2
5,482,595,200 characters per 8 3488-1 / 3488-2 (maximum system)

system architecture:

301 or 3301 processor connected to:
380 or 3388 RACE control module
one control module for from 1 to 4 RACE machines
so max system is 3301 / 2X3388 / 8X3488 each 3488-1 with an expansion unit (-2)

All of our machines had the removable magazines and its related magazine loading guide

We ran these things hard all day on the day shift and spent at least 2 hours each during 4-12 shift doing PM, replacing bearings (lots of those) pinch rollers (ditto) and running diagnostics. We cleaned the R/W drum with liquid Freon, had a 5 gallon can of the stuff in the shop. Neat stuff, would boil in your hand, OSHA would have a fit today.

If I haven't bored you too much and have any detailed questions, fire away.

Cheers,

Barry

Hi Barry,

Your data shows 256 cards per magazine. Looking at a card or the envelope which held the new cards (see photo on web site), the physical card address represented by the fingers on the sides of a card is only 7 bits, enough to address only 128 separate cards. It was my understanding that the unique fingers on each card was what allowed its selection out of a specific magazine (or bin as I recall them being called). This seems to imply that each magazine might have had two separate compartments. In fact a very fuzzy memory of mine is that the "bins" in the units at CIT/CMU might have had a dividing partition in each. Can you confirm or deny that the magazines on the units with which you worked had two separate compartments?

73, Chris Hausler
My recollection was that there was a single aluminum carrier which held the cards stacked side by side that was removable by an operator.  This carrier was about 4 inches in width, but I don’t know if a “set” consisted of one or two carriers.  I don’t think that 250 cards would have fit into a single carrier, so this impression is consistent with Chris’s comments that a magazine held 128 cards (or often less, because midnight operators didn’t replace cards that were damaged).

Dave
From: foxbwcomcast.net
Sent: Apr 8, 2012 11:03 PM
To: "J. Chris Hausler"
Subject: Re: (Fwd) RCA 3488 RACE

Chris, et al,

Roger that, each magazine had two halves, 128 cards each

A combination of push up bars, push down bars and the addressable teeth of the gripper assembly managed to select one of 128 cards, from one side of the mag or the other. Very complicated mechanical mechanism driven by simple discrete transistor logic circuits.

Each card has 4 sets of notches / tabs, labeled A, B, C and D.

The A set are on the bottom edge between the two labeled E and F, to be explained later
The B and C set run together on the right hand two thirds of the top of the card
the D set are on left hand third top of the card

Quotes from the theory of operation:

"Each magazine contains 256 cards, physically separated into two packs of 128 cards each. Each card is notched according to an address code determined by the 7-bit card address. It is this notch pattern that defines a card within the pack rather than its location with respect to the other cards.

Set A notches

The bottom edge set A notches are derived from the four most-significant bits of the 7-bit card address. Set A contains 16 notch positions, one for each of the 16 possible binary combinations of these four bits. The position corresponding to the card address is not notched, while the other 15 positions are notched.

Set B notches

The top edge set B notches are derived from the same four most-signifcant bits of the card address as the Set A notches. Set B contains 8 notch positions, a 1 and a 0 position for each of the four bits. The set B notch positions are notched in binary fashion, corresponding to the card address. The A and B sets of notches break down the 128 card pack into 16 8-card groups.

Set C notches

The top edge set C notches are derived from the least-significant bits of the card address. Set C contains 6 notch positions, and is notched in binary fashion, corresponding to the card address.

Set D notches

The top edge set D notches are derived from the same three least-significant bits of the card address as the set C notches. Set D contains 8 notch positions, one for each of the 8 possible binary combinations of the three bits. The entire set D area is cut away except for the notch position corresponding to the binary combinations. The uncut position is termed a tab. The C and D sets of notches break down an 8-card group into individual cards.

The bottom E notch (closest to the front ahead of the A set) is always located at this position on the card. It is used with a guide bar to properly position the card in the magazine.

The bottom F notch (closest to the rear behind the A set) is always located at this position on the card. In conjunction with the other bottom edge notches, it is used in the development of the block index signals by the photocells located at the capstan.

If you all are not bored to tears yet, I can describe the sequence step-by-step on how the mechanical system utilizes these tab / notches to grab just one card out of a magazine of 256 and send it to the R/W capstan. Any takers?

73 (guess there are some hams out there)

Barry - W1HFN
Hi Barry,

Good Stuff!! Keep it coming.

Yes, I'm a Ham but not an active one, call WB2TLL. My interest in the telegraph these days is land line American Morse telegraph. I collect the instruments (There is a photo on the site under "CMU photos from Chris Hausler" part way down which shows a bunch of telegraph sounders in resonators taken in my "family room", not at CMU :-) with well over a hundred in the collection at the moment. If you are interested there is a group, the Morse Telegraph Club, founded in 1942 of which I am an active member, see http://www.morsetelegraphclub.org/ and we also now have an Internet telegraph system (replacing all those pole line wires :-) free to download at: http://home.comcast.net/~morsekob/ . (Using this program there's a telegraph sounder clicking away next to my computer connected to wire 121 with today's news headlines at 20 WPM from a Yahoo RSS feed as I write this :-)

We also celebrate Morse's birthday each year on "Morse Day", the last Saturday of April (he was born the 27th) and I will traveling (taking Amtrak) down to VA and joining with MTC's "Washington-Baltimore" chapter at the Railway Mail Service Library in Boyce, VA for that event. I will also be driving out to Dayton, OH for the big annual "Hamvention" in May, to help man an MTC booth (right next to Vibroplex's booth :-) I don't know whether you attend this but if you do, please stop by and say hello! I believe Mark is a ham too.

73, Chris Hausler
From: foxbwcomcast.net
Sent: Apr 9, 2012 10:44 PM
To: jchauslerearthlink.net
Subject: 3488 card select

Hi Chris,

I think it would be better if I just sent this stuff to you and you can decide whether or not to forward it.

Very interesting your landline Morse interest. I could never figure out how clickey-clacks could be interpreted, while dits and dahs were easy for my mind to grasp. BTW, the local art museum has one of SFB Morse's paintings displayed prominantly. http://www.worcesterart.org/Collection/American/1907.35.html

I have been licensed over 50 years now, operate from 160 through 900MHz, actually have my own 900MHz repeater on a local hilltop. Main interests are RTTY, antenna experimentation and fox hunting.

3488 Card Select Sequence:

Preface: there were two kinds of card extraction grippers, the "clapper" and the "scissors" type. The only ones I am familiar with are the clappers, so the sequence will refer to that type. They are quite different in appearance and function and pretty difficult to describe in text. I do have a scanner, so if someone wants, I can scan the different images. BOTOH, I doubt if anyone has a RACE machine in their basement with which to compare the pics ;>)

The card selection and extraction process goes through 9 steps, controlled by 4 flip-flops, STA, STB, STC and STD. They are sequenced such that some are set while the others are reset, i.e.: 1=set, 0=reset.

STA STB STC STD

Step 1 1 0 0 0
Step 2 1 0 0 1
Step3 1 0 1 1
Step4 1 0 1 0
Step5 1 1 1 0
Step6 0 1 1 0
Step7 0 1 0 0
Step8 0 1 0 1
Step9 0 0 0 0

that may be more than anyone wants to know, but it is interesting, at least to me, how such a complicated sequential mechanism could be controlled with only 4 FFs

Step 1 - a. the selected bail assembly is activated b. the gripper knuckle is enerized c. the selected mag extracton pinch roller is activated for "pre-spin" to speed extracton d. the upper bar assembly pushes down all cards

Step2 - the lower bars to select the card are pushed up

Step3 - a. the pinch roller pre-spin is de-activated b. the lower gripper comb is activated c. the gripper assembly is lowered

BTW, all of these mechanical activations are verified by micro-switches which feed the logic matrix, thereby advancing the steps (or setting error bits if not true)

Step4 - a. deactivate holddown bars and lower bars b. lower the pushdown bars

Step5 - a. deactivate pushdown bars and bail assembly b. raise upper bars

Step6 - a. move gripper assembly forward (it has one card gripped in the teeth of its comb)

Step7 - a. open gripper jaws b. raise gripper assembly

Step8 - a.move gripper assembly rearward b. close pinch rollers to pull card into raceway and thence to the R/W capstan

Step9 - reset ABC&D FF to 0

Had enough? There's lots more

73,  Barry - W1HFN
Hi Barry,

Yes, please go on! Understand you may be the only remaining repository of this information. So getting it out there on the web on Mark's page may be the only real way of preserving it "for the ages". I have been recently reviewing the pages maintained by several members of an east coast "retro computing" group. One has an interesting history of magnetic media storage but completely misses the card media units of the 60's (the RACE, the NCR CRAM and the IBM Datacell, at least those are the ones of which I am aware). There is a lot of data out there already on the IBM Datacell and I've even found some data on the NCR CRAM but almost nothing on the RCA RACE. BTW there was a rumor that both the RACE and the CRAM were designed by the same principal engineer but I've never been able to verify this (in fact the rumor went on to say a third card memory system was designed by this man (but not the Datacell) but I've never found reference to any such thing. My only exposure to a CRAM was that it was the centerpiece of the NCR exhibit pavilion at the 1964-65 NY Worlds Fair. CIT (CMU) did also have a Datacell connected to their IBM 360/67 for a while in the late 60's but it didn't last long. I have couple of the cells from it (found in a dumpster when visiting the campus in 1976). Of course the big problem with all of these devices is that they traded complexity for storage capacity when compared to contemporary disk systems (from research it looks to me that they were capable of roughly an order of magnitude more storage than contemporary disk systems but disk technology and storage density advanced so quickly that the more mechanically complex card memory systems were quickly overtaken). Anyway, so photos, scans of manuals and all would be welcome!

As to understanding Morse Code on a sounder, it's just another sound. Each letter has a unique sound to it just like with tones and you learn the sound of the letters and sometimes whole words. If you try and follow the dots and dashes with either code, you're doomed to begin with ;-) I'm not all that good with either code. From what I've found with American Morse on sounders is you really have to start at a speed of at least 15 to 20 WPM and then add in a fair amount of Farnsworth between the individual letters. This helps you learn the sounds of each letter rather than paying attention to the individual clicks. That free to download MorseKOB program will allow you to set up some delay between letters when told to read from a local text file and can also mix tones with clicks to help you transition to clicks. Of course, it uses American Morse instead of International Morse but after all 15 of the letters are the same in both. Although connecting real telegraph instruments to you computer is a lot more fun, the program will generate the clicks quite well using the computer's sound system and the PC only version (2.5) allows you to select a range of different sounding clicks to boot. You might want to try it just for fun. You can read all about it on the program's web site.

There's also a couple Yahoo groups for land line telegraphy. The main one is "slowspeedwire" and there's one specifically for the MorseKOB program called "MorseKOB" of course :-) (Just in case you're interested ;-)

Thanks for the link to the Morse painting, I'm going to pass it along on slowspeedwire. He was quite a skilled painter (and writer as well) and his paintings are quite prized today.

My ham experience has been rather checkered. I was briefly active in the early 1960's but that was about it. The license I currently have I got in the mid 1970's and have never used it. I have been a member of the local ham radio group for decades. It was at the local hamfest flea market (once quite large, as big as Dayton is today) about 15 years ago that I found and purchased a telegraph sounder which started me down this "land line" path although at the time I was looking for old computer stuff rather than radio stuff. Maybe one of these days...

I will forward all this to Mark as well

73, Chris Hausler

Hi Mark,

Barry has sent me more RACE stuff, including two photos of, if I understand correctly, two different designs used for selecting cards. If for some reason they don't go through, let me know and I will send them separately as I have saved the images. I will also copy you on my response to him thanking him (and asking for more :-).

73, Chris Hausler
(click on image for full size version)

Hi,

The two images of the clapper and scissors gripper are attached.  As stated earlier, the clapper gripper version is the
only one I ever worked on.






The 3488 saga continues:

The card has been extracted from its’ magazine and sits between the pinch rollers.   And we quote:

“Before the card can be extracted, it must be determined that the capstan does not have a card on it.   When the capstan is ready and the previous card had not left because of 32 unserviced revolutions, the activate pinch roller signal (APR) is generated to extract the card and place it on the capstan.

When the card enters the transport, its position is sensed by a photocell.   As a result, a card in transport signal (CIT) in generated which provides a signal to load the active register.   Thus the active register contains the address of the desired card.   This address is compared with the address of the extracted card as determined by photocells in the read / write station entry track.   When this comparison verifies the address of the extracted card, the recirculate gate is closed and the read or write operations can be performed.

Should an incorrect card be extracted, the address comparison will detect it, at 3he re76ciadjfl (cat just walked on the kbd, sorry) the recirculate gate will remain open, and an error indication will occur.   The incorrect card will pass around the read/write station and be returned to its magazine. “

Read / Write

“As the card revolves on the capstan under the recording heads, the card position is monitored by two sets of photo-transistors. These sensors detect the address code contained in the notches along the lower edge of the card, and determine which block on the card is under or approaching the read/write heads.   The block index signals enable the quadrant decoding matrix.   The matrix generates these control signals to the control module:   pre-begin read/write, block approach detected, begin read/ write, and next block pulse.   In the control module, these signals determine the time at which data transfer begins to the storage unit.

When the block address of a card on the capstan is changed, the head register is compared with the new address to determine if a head position is required.   If the head position requires changing, the head register check flip-flop will remain set allowing the head change enable flip-flop to be set.   The new head position data is loaded into the head register and the transfer of control signals from the quadrant decoding matrix is inhibited for two consecutive index pulses.   This allows the head positioner to move the heads to the new position and prevents the transfer of erroneous control data to the control module during the transient period.”

I wish the manual had a detailed explanation of the head position sequence, but it does not.  It was quite a kludge (as was quite a bit of the machine) using hydraulics to move the heads up and down (and the gripper assembly back and forth) and a very rudimentary binary sensor to determine where the head was.  The hydraulic unit was in a slide out drawer under the capstan assembly, and if memory serves, there was always a puddle of hydraulic fluid on the floor there.  Of course it was a British unit, made by Vickers, so that explains the leaks.  I can say that, having owned several British sport cars "back in the day".  

Thanks for all the info on the landline Morse.  I will take a look at the site you mentioned and see if I can get my ears around it.

73,

Barry
Hi Barry,

Once again thanks!! Keep it coming. I have forwarded the information and photos to Mark. One wonders where all the mechanical engineers who designed these things have gone. From my perspective, computers just aren't "fun" anymore, hydraulic oil and all ;-)

On the Morse Telegraph front, Saturday, April 28th, is the annual celebration of "Morse Day" by the Morse Telegraph Club (Morse was born on the 27th but we always celebrate it on the last Saturday of April). If you download the MorseKOB program or just run the Java version (3.0), there should be a bunch of folks on wire 11 for several hours likely mid day (eastern or central time, I haven't heard from any west coast chapters) that day. I'm going down to the Railway Mail Service Library in Boyce, VA and gathering with several folks and we will certainly be on the wire. I know at least three or four chapters from around the country will be having events and in addition, a number of individuals will be on too. Remember, you don't need any real instruments, the sounder clicks will be generated by the computer's sound system. The PC only version (2.5) has several different selectable sounder "sounds". With this you can listen in on the activity. Just FYI...

73, Chris Hausler

emails from Dick Shoup

Date:            Mon, 02 Apr 2012 10:00:51 -0700
From:            Richard Shoup <rgshouprgshoup.com>
Subject:         G-20 etc

Hi Mark --

Wow, I just discovered your Athena/G20/G21 pages.  What a great collection!  Brings back many memories.  Good to hear more about Carl L too.  I knew he had died, but the pictures etc are great to see.

Unfortunately, the link to "A Visual Display System Suitable for Time Shared Use by Jesse T. Quatse" is incorrect.  I might have an original of this document somewhere, but it would be great to have the scan as well.  (Mark's note: I fixed the link)

Best regards,

Dick Shoup
www.rgshoup.com
Dick,

Thanks for your email. I forwarded it on to the group of us that have been communicating over the past few years about the computers at Tech.
I also fixed the links that you mentioned were broken. You are first to point that out.
If there is anything you would like to share about the G-21 or Tech in general, I'd like to add it to the web page.

Mark

PS: I enjoyed the talk you gave at CS50.
Thanks.  I'm sure Jesse is also excited to see your collection.  I don't have anything profound to contribute right now, but will be cleaning out my garage soon and will perhaps find more stuff from this era.  Unfortunately, the site www.cs50.cs.cmu.edu has been decommissioned.  I've emailed them, and perhaps it will reappear.

Best,  -- Dick
I downloaded all the CS50 videos so I have copies here.
Ah great.  I have only Jesse's and mine.  If CMU won't put them up again, I may ask you for copies.  Best,  -- Dick
Regarding the RACE file:

I remember that it was supposed to be able to eject a card through a slot at the rear of the machine.  The slot contained a brake or clamp that would grab the card so that it would stop with part sticking out for easy manual removal.  Of course, the clamp seldom worked, and the result was a card flying at very high velocity out of the machine and across the room.

It was also known that those thin floppy cards at full speed could easily cut through a pencil on the way to the drum.  And the description of the machine's sound as that of a hay baler was common, and very appropriate, when all those internal bars moved up and down to select a card.  What a classic device.

Dick

Dick passed away in 2015



Richard G. Shoup

July 30, 1943 - July 18, 2015

Richard (Dick) Shoup, computer graphics pioneer, passed away peacefully at his San Jose home on July 18, after a long battle with lung cancer.

Born July 30, 1943, Dick was a native of western Pennsylvania. He earned a BSEE and a Ph.D. in Computer Science from Carnegie Mellon University. His Ph.D. thesis was the first to explore programmable logic and reconfigurable hardware, now widely used in computers and consumer electronics. Dick was one of the first employees at Xerox Palo Alto Research Center, where he developed the SuperPaint system, one of the first to use computer animation in television and movies. He received both an Emmy and an Academy Award for this foundational work. In 1979, he left Xerox to found Aurora Systems, a video graphics and animation company. Dick later worked for Interval Research Corporation in Palo Alto and founded the Boundary Institute to further explore a lifelong interest in physics and parapsychology.

An avid musician in his spare time, Dick played jazz trombone for many years in various Big Bands throughout the Bay Area.

Dick was most proud of his family. He leaves behind a wife, Nancy, of San Jose, and four children, son Randall of Pacifica, son Eric and wife Christina of San Mateo, son Alexander of San Jose; and daughter, Celanie, of Pittsburgh, PA. He is also survived by four wonderful grandchildren and a sister, Judith Sharp of Santa Cruz.

A Celebration of Life Service will be held at 1pm on Sunday, August 23 at the Computer History Museum, 1401 N Shoreline Blvd, Mountain View.

In lieu of flowers, please consider a memorial donation to the thoracic oncology research fund of Dr. Joel Neal, Stanford Cancer Institute. Please indicate that gifts are in the memory of Richard Shoup for Joel Neal Research-GHBYM. Make checks payable to Stanford University and mail to Stanford University Development Services. P.O. Box 20466, Stanford, CA, 94309-0466, or go online to https://makeagift.stanford.edu.
Published in San Francisco Chronicle on July 26, 2015

1968 Comp Center

From Pat Stakem, via Chris Hausler; here is an employee list from the Comp Center dated 29 Oct 1968.

On the list is my fraternity big brother, Bruce Black. Bruce died in 2007.  Also on the list is Carl Lefkowitz along with some of us who contributed to these web pages.

I received these emails and photos of the G-20:

Joe sent me photos of the IBM 7040 (which you can view on the IBM 7040 web page) and then he sent me these :

From:    joe rubenfeld <yosefsheagmail.com>
Date sent:    Wed, 28 Mar 2018 17:43:24 -0400
Subject:    Ron Elco's 1966 snapshots

Mark,

I met Ronald Elco at the CIT computer center during the 1965-66 school year and we became good friends.  He came from Donora, PA.  I also knew his older brother Richard (1936-2010), an EE professor at Tech.  Ron died in 1992, at age 47, due to chronic medical problems. 

Here are some of Ron Elco's Tri-X negatives I scanned last year.  They negatives were scratched and dusty.  Anxious to see the images and I didn't clean or spot them thoroughly.  If requested I could do some more dusting.  These images show the G-20 computer room circa October 1965.  While I recognize the hardware, having been a 2nd shift G-20 console operator around the time, I don't recognize the people.

Jon

Chris Hausler identifies this woman as Anne Rindt, see email below.

Chris Hausler identifies this man as Bill Zippler (WZ02) , see email below.
From:    David Rodgers <dave_rodgersmsn.com>
Subject:    Re: Ron Elco's 1966 snapshots
Date sent:    Sun, 1 Apr 2018 15:50:55 +0000

Great pictures of the 4th floor inhabitants.  From a time when people dressed for work.

Dave
Mark,

Double checking I saw that the date the G-20 pictures were developed was May 13, 1967, a bit later than I remembered.  By then the G-20 had been upgraded to G-21.  Also by early-1967 I had dropped out of CMU and was working on the Jersey Shore developing IBM's Hospital Information System.

In autumn 1962 the part of Scaife Hall top floor that later housed the 7040 was full of drafting tables.

Dr Alan Perlis taught the introductory computer course s205.  IMHO he was not the best choice for that class.  He often assigned GATE programming problems that required moderately advanced math or solving brain teasers.  I had taken some calculus in high school which sometimes (but not always) helped.

Perlis's mid-term was a two-part interception problem.  As I had recently come from NASA where I was programming graphics for Project Gemini orbital rendezvous, part 1 (target in a circle) was a snap.  An elegant solution to part 2 was so beyond my mathematical knowledge (target moved around a square) I solved only for the degenerate cases.

During the marathon s205 mid-term exam, with a late-night submission cutoff, the tables doubled as cots.  From time to time Charlie Pfeffercorn walked through the room selling snacks.

Fortunately for all concerned a glitch in the TEACH grading system failed and credit was awarded for just showing up!.

Jon
Date sent:    Mon, 2 Apr 2018 18:50:59 -0400 (GMT-04:00)
From:    "J. Chris Hausler" <jchauslerearthlink.net>
Subject:    Re: Ron Elco's 1966 snapshots

Hi Jonathan,
 
Thank you!!!  Great photos!!!  As to the people in them, the dark haired girl is a complete mystery to me, I have no memory of her.  However, the blond is Anne Rindt, there's already a photo of her on the G-20 page taken by Dave Chou.  The man in several of the photos I am almost sure is Bill Zippler (WZ02) a really great guy.  But again 50 years and rotting neurons might be fooling me so if anyone thinks differently, please let me know!  Bill was still there as a lead operator when I started working as a part time operator in 1968.
 
As to your other note, yes, sometimes a computer glitch can be a good thing :-)  I think I heard that there was a plan to port TEACH to the 1108 but it never happened.  When I was working as an operator on graveyard shift I remember a number of nights when some project was due the next day and the 3rd floor of Scaife was littered with sleep deprived people, some just sitting on the floor up against the walls snoozing, hoping to get just one more (hopefully finally working) run in before morn...

Chris 
Hi all,

These vintage computer center pictures came to light through a series of accidents. In 1995 I phoned Ron Elco to borrow his negatives from our December 1966 UK trip.  In 1967 he gave me a few B&W prints. I didn’t begin and serious photography until 1973. The phone call became an awkward conversation with his widow who told me Ron died three years earlier. I let the matter of the photos drop.

Sixteen years later, in 2011, I found myself passing through Pittsburgh more often, I called her again and we met. It was probably the first time since their wedding in the ‘60s. She was able to locate only one print. The next year we spent hours going through boxes of prints and negatives but came up with nothing more from the UK. While searching I came across the computer center photos and put them aside.

During a recent phone call an old friend (Joe Rubenfeld) brought up the subject of the CIT computer and some of the people who worked there. After the call I Googled a name from that era and found Mark’s site.

I started FORTRAN and FAP programming at NASA a few months before arriving at CIT (1962).   At Tech, I was science major who took s205-207.  I did no development work for the comp center.  At CIT I programmed and operated (1964-66) to pay my school bills.  After leaving in 1967 I continued earning money with (mostly mainframe) computers until my 2010 retirement.

Jon
Jon and Chris also talked about the G-15 which I put on the G-15 web page.

Hi Jonathan,

I never did any programming for the Comp Center, just worked part time during the school year starting early 68 (two shifts a week on average) and the income from that more or less covered my food expenses and text books.  Even then I did not see myself as a programmer but rather as an EE designing circuits.  In fact, I quit the Comp Center May/June 1970 and went to work for the Comp Sci Engineering Lab as a technician doing circuit design and repairs, first part time (about 25 hours a week as at that level I got free tuition for several courses) and then full time, leaving the end of July 1971.  I still have one each (bare) of the several printed circuit cards for which I designed the logic and did the actual taped layout that year (modem rack units).

My experience at Tech was pretty much my only contact with "mainframes" other than occasionally submitting a Fortran job to some 360 either at my employers or a customers at some job site in the early to mid 70's.  For the next 20 years as far as work went it was most all 16 bit minis, DG in the 70's and DEC PDP-11's in the 80's.  There were a few exceptions and I did play at home with 8 bit micros (Motorola 68XX) starting early 77 up until about 85 or 6 and actually spent about two months living in a hotel on the beach in Santa Cruz, CA for work in 81 designing the software for an early RFID reader which used a 6801.  In the early 90's I transitioned to PC's running a clone Unix (QNX) which continued until I retired in early 2010.  I could blather on further about my work experience but if you are really interested, here's a couple of links to pages hosting my "memoirs" on a web site maintained by a friend of mine for former US&S employees (runs on a 486 in his basement so don't mind the sometimes slow loads :-)  I never worked for US&S which is in Pittsburgh but did spend my entire career in the railroad control systems industry.  The downloadable articles also talk about some of my other weird hobbies :-)

http://has-been4.ddns.net:8080/wp2/index.php/?p=15045

http://has-been4.ddns.net:8080/wp2/index.php/?p=16226

As to old photos I got lucky in that quite a few of my CIT/CMU photos were totally destroyed in a flood at my parents home back in the early 70's (Hurricane Agnes).  The ones on Mark' site survived although one shows water damage.

You mention the 1401.  By the time I arrived, the 1401 was in the basement of Warner Hall.  There were rumors that previously it had been in Scaife in the northeast corner (where the 1108 was put, at the same end of the building as the 7040 (northwest corner).  Where was it when you were there?

As to the 7040 (on which I taught myself FORTRAN in spring 67), two different books by IBM folks say that it was designed starting in 1961 to provide competition with other companies making mid range mainframes, a hole in IBM's marketing.  It was announced in 61 when you could take delivery of a G-20 but was not available until 63, almost the dawn of the IBM 360 series.  The books mention several companies with which they thought they needed to compete, but only one machine, the G-20.  I have not been able to determine how many 7040/44's were built but there is some information I found indicating 48 G-20's were produced.  IBM was accused by their competitors of advertising vaporware hardware to blunt competition and to me the evidence is strong that the 7040 was such a product.  No one ever lost their job by buying IBM :-)  I actually came across this prejudice in customers twice, once in about 1989 and again a decade later.  The first time was when I sold a system based on a network of Micro PDP-11's to a mid sized railroad (MRL) over lunch with its president.  About a week later I got a call from the company's CTO calling me every name in the book for not providing a system based on IBM equipment and insisting that we change.  Of course we didn't.

I've never figured out why CDC bought Bendix Computer as to my knowledge they did not continue any of the products.  If you have any information and/or ideas on the subject I would appreciate knowing them.  By the time I arrived at CIT, the G-20's were being maintained by Philco CE's (of course the scopes system was Philco).

73, Chris Hausler
Chris,

I discovered the joys of systems programming in 1964 and found I was good at it.  I did a lot of consulting work in New York, New Jersey, San Francisco, Chicago, Milwaukee, and with the Feds in Washington, DC. For the last ten years of my so-called career I did security evaluations and support for CA (supporting CA-TSS, CA-CAF2).

In the fall of 1962 Scaife was new and had lots of unused space.  The G-20 installation seemed fairly recent.  I wasn't shown it during an April orientation visit.  I'm guessing it happened over the summer.  In '62 there was only one.  My best guess is that second G-20 arrived during the summer of 1964.

In autumn 1962 there was no evidence of an IBM 650.  One occasionally heard oblique references to it and GAT. I'm sure it was never in Scaife.

During my year away from Tech (1963-64) I worked for a Wall Street stock broker and learned pretty much everything there was to know about IBM 1401 operations, programming, Autocoder assembler, etc) When I worked with Tech's 1401, it was across the hall from the RCA 301 on Scaife's 4th floor.  It had a 1405 RAMAC (physically huge hard drive).  I did some outside programming work in it.  One was a analyzing commuter routes from a Port Authority rider survey. I also did a special project for the computer center that involved printing mag tapes from the 301.  The 1401/1403  default  print chain was the 48-character AN set.  Internally the 1401 supported 64 six-bit characters. The 301's 7-track tapes had 128 character combos.  The Comp Center has a special 1401 print chain with 96 (or 120?) characters on it.  It was a good trick making that work!

The budget 7040 came to the game late.  Much of its architecture, except for I/O, was based on the 704/709/7090 family.  Having learned 7090 FAP assembler, 7040's MAP was a snap. The I/O was far more modern than the 7090's allowing the use the same unit record gear as the 1401. The I/O channels look like they were adapted from the 1410.  I don't know how many were built but it was far more than 48.

I was a full time 7040 operator in summer 1965.  At the time Tech's 7040 supported an ALGOL compiler developed at the University of Grenoble in France. Its error messages were in French.  If memory serves it also ran an S/360 assembler and simulator.  This early S/360 support gave me huge advantage when those systems hit the street

I would guess CDC's purchase was either to reduce potential competition or acquire new developers. Perhaps they thought they could migrate customers to CDC machines.  With such a small installed base they must have bought Bendix Computer Division for a song.  In time CDC diversified with facilities management, Control Data Institute and S/360 compatible hardware. In 1974 I worked for them as they ran a couple of airline reservation systems on IBM and Univac hardware.  That's an entirely different series of surrealistic EDP stories.

Jon

Hi Jonathan, 

Yes, I never had any experience with a 1620 or even saw one in the "flesh", a loss I truly regret.  I believe that there was a 1620 over at Mellon Institute in the early to mid 60's however as some Pittsburgh based CMU students (Fred Brosi comes to mind) had some use of it while still in high school.  In my collection of computer modules I have one 1620-I SMS card, one of the modules I hang on my Christmas tree each year.  On Mark's G-20 page down near the bottom there is a photo taken Christmas 2008 showing me and my tree that year.  You can see the vacuum tube G-15 flip-flop hanging on the tree down low and center.  The 1620 module is just to the left of it, a little higher and behind it somewhat.

One of the principle reasons I became a computer operator at CMU was to be able to use their mainframes as personal computers :-)  I was always and still am a "hands on" kind of guy :-)  This was particularly true with the 1108 as it was so fast that it would usually go idle about 2AM in the morning leaving it for my personal use the rest of the graveyard shift, which is why I preferred working that shift.  I could do this as well when working on the evening shift but could only work on my own jobs between reading in trays of user jobs while the machine was busy with them.  I had done this the previous year before joining the staff but only in the middle of the night and with the cooperation of the operating staff (this is how I used the 7040 in particular) but as users weren't normally allowed on the fourth floor I risked being thrown off the fourth floor (and was occasionally :-).

As to the Potter drives, the most experience I had with them was the one inch G-20 ones and they tended to be rough on the tapes, crinkling the edges and even stretching them.  Possibly this is because by 1967 when I started doing this, they were old and possibly not being well maintained.  As I said previously I still have several of my G-20 tapes in my "junque" pile.  In that same Christmas tree photo you can see one just to the right of the tree in the background on top of some books.

I never did any 1401 programming despite providing the half inch tape for WRCT's record library program developed by Dave Rogers.  I know that the 360/67 printer had one print chain (or train, I never got straight the difference between "chain" and "train") which had lower case as well as upper case characters but as a result it was much slower as I think the character set was repeated only twice around the path rather than three or four times for the upper case only chain/trains.  I made use of this once to print lower case letters from a 9-track tape I had written on CS's PDP-10 using the Snobol-4 language, which although located in Scaife (where the G-20 had been) was not considered part of the Comp Center.  It even had its own separate operating staff.

Yes, I am aware that the 7040 as it was a late 7000 series design had the most advanced designs for the time including I understood the 7094's floating point processor design and better interfaces as you say.  Further I understood that tech's 7040 was completely "tricked out" short of having the faster memory of the 7044 (the only difference between the 7040 and 7044 basic systems).  And it was popular which is surprising in that IBM makes little reference to it in their histories.  Again, I think the competition was justified in complaining that it was promoted long before it was actually available to blunt the competition, probably the only reason for its existence as IBM was a fierce competitor.  It was probably most famous for being the original platform on which WATFOR was developed by the University of Waterloo (same place that Unix clone I used at work for 20 years, QNX, came out of). 

On the 7040 page on Mark's site I provided a link to a friend's web site which had 709/7090/7094 simulators and related.  Unfortunately the site is no longer up (the friend's wife is suffering from Alzheimer's and it is keeping him busy, a sad occurrence) but if you are interested I had already downloaded most of it and could send it to you.

Algol on the 7040!!!  That's a new one to me.  Also on Mark's 7040 page available for download are several CIT/CMU 7040 users documents which I scanned from my collection but none mentions Algol.  There is reference to a FASFOR compiler which I do not recall using.  Of course the big user of the 7040 was the nuclear physics folks.  The 1108 because it had a good Fortran compiler and a good Algol compiler more or less replaced both the 7040 for the physics folks and the G-20 for the Algol fans.  I know that there was a three step nuclear physics program (from the "scanning labs") which took hours on the 7040 and ran as a single step on the 1108 in 12 minutes.  The program as a whole and each of the three steps had names.  Two of the names were "Indian" and "Arrow" and the other two were related names.  Although sounding American Indian, it seemed to me that most of the users of it were India Indians :-)  Of course the 7040 was set up to handle 556 BPI half inch tapes, the best one could get with those 7330 drives, but I think this density was somewhat of a physics community standard as the 1108 even though it could do 800 BPI was always set up for 556 and while I was an operator we would get in tapes from other physics locations (Brookhaven was popular) and they were all 556 BPI.  I was also unaware of the 360 assembler simulator but by my first use of the 7040 in early 1967 the 360 at CMU was more or less operational.

As to CDC's purchase of Bendix, although they might have gotten some developers, they would have likely had to move them to their site in the Twin Cities and having seen such moves attempted in my own experience, the really good people tend to want to stay where they are and as they are good have no problem finding other remunerative employment in their area.  The folks who do move tend to not be, "the cream of the crop" :-)  I agree that CDC probably did get it for a song.  Further I never saw that CDC's prime focus was small and medium sized computers although they did have them.  What with Seymour Cray they seemed to focus on high end machines.  The only CDC machine I was ever, and then only very briefly, involved with was some 3000 series machine.  As to your airline reservation system, about 20 years ago there was a very interesting book published by MIT Press titled, "A Few Good Men from Univac" which was sort of a history of Univac/CDC/Cray as each kind of came out of the other.  The author was David E. Lundstrom.  The airline reservation system was discussed mostly in chapter 16.  I got a copy back then and it was a very good entertaining read.

73, Chris Hausler





G-20 Links

The Way It Was: Tales from a life in computing  -  My third computer: The Bendix G-20
This site prepared and maintained by Mark DiVecchio

email :  markd@silogic.com

DiVecchio HOME
Frazzini HOME
Site HOME