3. NUMAC (Northumbrian Universities Multiple Access Computer, UK)

posted Oct 30, 2010, 6:09 AM by Jeff Ogden   [ updated Nov 25, 2012, 3:35 PM ]
The third site to install and run MTS, the Northumbrian Universities Multiple Access Computer (N.U.M.A.C.) was a collaboration between the universities of Durham (UD), Newcastle upon Tyne (UNE), and Newcastle Polytechnic that shared a S/360-67 at Newcastle and ran MTS starting in 1969. NUMAC installed a second MTS system at Durham in 1982.

The Reports of the Director of the University of Newcastle Computing Laboratory provide an interesting history of events at the Laboratory and N.U.M.A.C. from 1957 to 1989.

In 1991 Ewan Page gave a public lecture at Durham, "NUMAC : Courtship, Conception, Culmination and Continuation". We haven't been able to find a copy of that lecture on paper or in digital form, most likely because there wasn't one.

Tony Young from Durham University offered these memories with what he called the "obvious proviso that they are a 40+ year memory" in several e-mail messages sent to Gavin Eadie and Jeff Ogden in late October 2010:

My own memory of why we adopted MTS is simply that with our small 512K memory(!) 360/67, TSS paged itself to death with more than about 4 terminals; we tried CP/CMS and MTS - as far as I remember we eventually opted for MTS because it then had much superior integration of batch and terminal use.

I don't know the full story as I wasn't there at the very beginning of NUMAC - I only came to Durham in Summer 1967 and a lot of  things were already decided by then. And in any case the OS decisions were being made up the road at N'cle - at Durham we were concentrating on how to produce a working service via a 1130 batch station (initially over a 2400baud line) plus slightly later also via 9 individual 300 baud lines.

I understood that from the beginning the main aim of the NUMAC link-up was to provide a terminal based service (as its name implies)  to both Univs, and that IBM was selling us the 360/67 on the basis of being able to run TSS. My understanding was that the original config had only had 256K memory but that this had to be re-jigged when IBM found that  TSS wouldn't run in any system smaller than 512K;  this all happened before I came (or got actively involved) so I don't know what bits of configuration we had to lose to stay within the budget - I think we ended up with less disk space among perhaps other things.   I think when the 360/67started in early 1968 we ran experimental timetabled slots of TSS and otherwise MFT. We at the Durham end were particularly interested in the support for remote batch workstations, and I don't think TSS handled these well (or at all !). Indeed I'm not sure that standard MFT handled them well either - which is why I think  from the start we adopted the 'renegade' HASP system.

I'm afraid I don't remember a visit by either Bernie or Bruce - but this would almost certainly have been to N'cle. They had a very active Computer Science Dept as well as an active and well-connected professor (Ewan Page) and he would have known what was being developed elsewhere in the field of terminal systems.

I don't know if Ewan had a 'plan B' right from the outset,  or only when it became apparent that the version of TSS as delivered in 1968 was just not usable  - or at least not on the machine configuration that we had.

And the following appears in An Interview with Bernard A. Galler, OH 236, August 1991, Charles Babbage Institute, Center for the History of Information Processing, University of Minnesota, Minneapolis, Minnesota, USA:

[pages 26 and 27] And it was actually that concept [virtual memory] that led to my trip around England after our sabbatical, which I'll come to later, because I was lecturing on that topic. It was new for most people, and I was lecturing on that all over England, which led to my connection to Newcastle, because they were very interested in that, and eventually they used MTS at Newcastle because of those lectures and that time-sharing system.

[pages 37 and 38] At the end of that year, I was invited to make a trip through England to give a series of lectures at a number of universities. That was June of '66, and that's when I mentioned previously that I made the important contact with Newcastle, Ewen Page and Bryan Randell and Elizabeth Barraclough and Jim Eve, all of whom turned out to be very important for MTS and virtual storage on the one hand, and later on for our second sabbatical over there in 1973. It was a very happy and productive year, that sabbatical academic year 1965-66.

On Oct 31, 2010, at 2:26 PM, Denis Russell wrote:

Jeff (and Gav and Tony),
        . . .
        I guess you have found have your answer wrt Bernie visiting. I wouldn't have known about the earlier visit since I was then a PhD research student in Physics plus also a "night operator" of the KDF9 computer. I joined the staff of the Computing Laboratory for a year approximately September 67 to August 68 as the 360/67 arrived and began operation. I then left for two years back in Physics at the University of Alberta approximately August 1968 to July 1970, when I came back to Newcastle for the long haul in computing. It is entirely a coincidence that MTS seems to have stalked me on that journey, and I can take no blame... :-)

        One other part of context that helps to understand events was that at Newcastle there was a "Computing Laboratory" in which the science and service sides worked quite closely, especially in the early years.

        WRT the second question about why MTS was adopted on NUMAC - well the 360/67 was bought on the premise of introducing the new Time Sharing technology into the UK. There was then a very strong government policy of supporting domestic industry via Government procurement, and this was government money. The argument for buying IBM was broadly that it was the only practical way of getting time sharing. To do this meant breaking a very strong procurement policy. The domestic product (an ICL knock-off of the 360 architecture) was late, and the one that Edinburgh got (after our 67) was eventually programmed by an internal team lead by Harry Whitfield some years before he came to Newcastle). The fact that when we got it TSS was a dog was mightily embarrassing, especially politically. In order to get some work out of this very expensive pile of gear, we ran timetabled MFT for batch and sessions of TSS. I have a vague feeling that we dabbled with CP/CMS but don't recall any detail.

        I'm not sure about any hardware improvements that were made to memory at the last minute, but I do remember that we got 2314 disks at the last moment instead of 2311s (30 meg disks instead of 10 megs if I remember).

        The practical failure of TSS meant that Ewan (and possibly others, I don't remember) called up their contacts - I think there was an emergency trip to the States, almost certainly including UM, and I recall a confidential meeting when the decision to adopt MTS was announced by Ewan and explained. I don't remember how much CP/CMS was considered, but it must have been in the frame. I believe that it was run a bit at Newcastle, but I don't recall any details. I don't recall the detailed dates either, but I had left (Aug 1968) for the joys of Alberta and the Rockies before the move to using MTS for service was made.

        Interestingly, and entirely by coincidence, there was a 360/67 in Alberta when I arrived. As I recall, that ran timetabled MFT (I think), and CP/CMS. They were hot on APL which was one of the drivers for CP. I was only loosely connected to the computing service there (my wife, Marion, became program librarian and Newsletter editor in the computing service during our time there), but the decision was made by UQV to move to MTS while we were there. I vaguely recall a visit by one Jeff Berryman (I think) and a.n.other from UBC as part of the decision making process. I also recall a visit by Keith Barnett from Newcastle (the chief systems programmer). Again, I left UQV before the changeover was made, to return to Newcastle and a service experienced in running MTS.

        I hope this helps. I think if I tried to put any more detail in, it would be somewhat shaky. The person who can probably add a lot more is Elizabeth. I've added her to the recipient list of this email. I'm sure she can put me right (she spent lots of years doing just that. :-) ). The other interesting possibility is Ewan page himself. I don't have contact with Ewan (my email address for him is years out of date). If we could get his comments on these events, they would surely be worth hearing. I don't know whether Elizabeth is still in contact.

        Best wishes to everyone,

        Denis

From: Tad Pinkerton
Date: December 21, 2010 12:02:22 PM EST
To: Jeff Ogden
Subject: Re: MTS at Grenoble?

Hi Jeff,

Yes, I know a fair bit.

When I was finishing my PhD in late 1967 or early 1968 I was trying to decide where to go for an overseas year of teaching before starting a career here in the States.  Bernie [Galler]  arranged for me to be interviewed by Professor Louis Bolliet from Grenoble and Professor Sidney Michaelson from Edinburgh.  The overall institution was called the Universite de Grenoble and the sponsoring group at the time was the Institute of Informatics and Applied Mathematics (IMAG), which I gather later became INP.

I received job offers for the 1968-69 academic year from both places and was really torn between them: Grenoble was the site of the 1968 Winter Olympics and the TV coverage was spectacular.  Both jobs were half-time in the department and half in the computing center.  But if I taught in Grenoble it had to be in French, which I had only mastered well enough to pass the reading exam for my degree.  So the family went to Edinburgh, but with an invitation to visit Grenoble while I was in Europe, especially to advise the computing center folks on MTS.

. . .

I went more often to Newcastle [than Grenoble] because it was only about 120 miles south of Edinburgh and an easy train ride.  The joke that year was that if someone from Europe called Michigan to ask about MTS they could be asked "have you called our Edinburgh office?"  A story from that institution:  after they received the MTS tapes and loaded up the system and booted up, they couldn't get past the time and date query.  They could not imagine that the format of the latter was so 'user friendly'!

Tad

On Dec 21, 2010, at 1:56 PM, Mike Alexander wrote:

This largely agrees with my recollections, although I didn't know so many details.

I think the story about the first IPL at Newcastle is apocryphal since I carried the tapes there myself and was present for the first IPL.  I remember it well since they insisted that I come pretty much directly from the plane to the Computer Lab and I more or less IPLed MTS in my sleep.  . . .

         Mike




At 14:01 -0500 9/12/10, Jeff Ogden wrote:
Am I correct in remembering that the NUMAC S/360-67 was the first Model 67 installed in the UK?
  -Jeff

On Dec 9, 2010, at 3:06 PM, Denis Russell wrote:

Jeff,
    That is almost certainly correct - certainly in a university. However, I couldn't be quite 100% certain for other communities outside universities. It might *just* have been installed in a government facility, or within IBM itself (they had a research institute near Southampton as I recall - though I never visited it). However, I'm very dubious about any other facility.

    I've copied this email to Brian Randell. Brian wasn't at Newcastle until a year or two later but he came to Newcastle from IBM research - in the USA, but he has a broader knowledge of the field then, and has a particular interest in computing history.

    Brian, can you confirm or deny my feeling about the /67 ?

    Denis [retired Assistant Director, University Computing Service]

PS Just occurred to me that Roger Broughton was also very much around then and may have more concrete knowledge than I.

On Dec 10, 2010, at 3:52 AM, Elizabeth Barraclough wrote:

Denis,

We certainly had the impression that it was an early trial of the system and the software but whether it was the first I don't know.

Elizabeth
[Director Emeritus, Newcastle University Computing Service]

On Dec 10, 2010, at 2:22 PM, Roger.Broughton wrote:

i do not know
but think it was the first in uk
ewan page the director at the time
had a hard fight with the university computer board
to get the 360/67 rather than the british icl mainframe
because he wanted vitual memory and interactive computing
good for ewan

[Roger was the long serving operations manager at Newcastle]



The quest continues with assistance from Denis Russell and Ewan Page:
 
To: "Ewan Page"
From: "Denis Russell"
Cc: "John Lloyd"; "Jeff Ogden"; "Gavin Eadie"
Sent: Monday, April 18, 2011 12:26 PM
Subject: Reference to a lecture you gave in 1991

Ewan,
It was good to see you at Brian's birthday bash, especially as you were looking so well. Sadly such convivial get-togethers of the old Computing Lab crew are all too infrequent.

I'm emailing you with a request for a reference to a public lecture you gave in 1991. First a little context: Jeff Ogden with the assistance of Gavin Eadie (to whom I've ccd this email) is compiling an archive of the history of MTS, see <http://archive.michigan-terminal-system.org/>. This is already a very impressive piece of work. One of the missing pieces is a reference to a public lecture you gave at Durham in 1991 (see the second paragraph of <http://archive.michigan-terminal-system.org/discussions/how-did-sites-learn-about-and-make-the-decision-to-use-mts/3numac>). If you could help them out with this I'm sure they will be very grateful. I have no doubt that any further additions, suggestions or recollections that you may have will also be most welcome.

Best wishes

Denis

cc John, Jeff, Gavin

From: "Ewan Page"
To: "Denis Russell"
Subject: Re: Reference to a lecture you gave in 1991
Date: Mon, 18 Apr 2011 15:45:48 +0100

Dear Denis ,

Until you gave me the reference to my lecture at Durham I had quite forgotten about it . I have no record of it; it wasn't published or intended to be. I'm sure that it only existed in note form and was otherwise delivered ex tempore . I remember that they put me up in Durham Castle in the Judges' bedroom.

The report you gave about MTS seems pretty accurate to me (you even spelt my name correctly - Ewan; not like some of my other friends ).

My lecture contained a fair amount about the politics of Newcastle and Durham getting together to operate a joint system; Durham was very much a junior partner in terms of cash and was nervous about being overwhelmed . I proposed equal representation on the governing board, the chairmanship alternating between the V-Cs of each university. I think, perhaps kidding myself, that agreement was helped by them knowing and trusting me - I had my first post, lecturer in Statistics there before being appointed to the Computing Laboratory. Of course they had been a minor user as of right of the Ferranti Pegasus that we had first, and had been well treated when we (Newcastle ) replaced it with the KDF9 .

Your comments about the IBM systems not performing were right and the enormous political cost (and possibly my job) if we couldn't get something decent going; I had made a great nuisance of myself to two ministries to escape from an ICL machine which wouldn't have worked for another 2-3 years. So MTS saved my bacon.

I didn't notice any reference to Newcastle Polytechnic in the site you referred me to. They joined us rather later when we ran a 370/168 as well as the 360/67.  At this time this was a cross-sector operation; the polys were run from the Ministry of Education (or whatever it was called then) and the U's by the University Grants Committee (or whatever it was called). We started quite a few things which caused head-aches for the civil servants as well as for us .

Of course, Elizabeth, you and the other colleagues had to deliver to keep the users happy; you did it pretty well with the happy result that you got rid of me, first within Newcastle, then further distant .

If any questions occur to you, do get in touch (by phone if you like , xxxxx xxxxxx) and I'll try and remember .

Best wishes

Ewan

From: Denis Russell
Date: April 19, 2011 3:41:40 PM EDT
To: Jeff Ogden
Cc: Gavin Eadie
Subject: Re: Reference to a lecture you gave in 1991 [Ewan]

At 17:01 -0400 18/4/11, Jeff Ogden wrote:
    ...
And you remembered a visit by Keith Barnett to UQV while you were there, but that had to have been after NUMAC was already running MTS I'd think.
    ...

The sequence was that I was in the Computing Lab from about Sept 1967 until about Aug 1968. During that time the 67 arrived and TSS was found wanting. MTS was tried. It wasn't in production when I left. I can't remember whether the decision to use MTS in anger had been taken before I left, but it had been tried and was under serious examination. At the latest, the decision to use MTS in production must have been taken soon after I left. I expect that MTS was running very soon after. There might be some information in the Computing Laboratory Annual Reports. I can have a look if you like (and if the relevant reports are still around).

When I arrived on Alberta they had a 67 on which I think they timetabled batch (of some sort) and APL on CP/CMS. I was in Alberta until about Aug 1970, and towards the end of that time they looked at MTS and had the visit from UBC as part of that. I was of course in Physics so my knowledge of the goings on in the Computing Center was not close. I think the decision had been made to run MTS some time during 1970, and as part of that Keith Barnett visited from Newcastle. He was then chief MTS systems programmer and came to give the CC the benefit of his experience of running MTS in production at Newcastle. The production operation of MTS started after I left Alberta in Aug 1970 I think. So, the sequence was definitely Newcastle/NUMAC and also UBC running MTS before Alberta.

I hope this makes this part of my recollection slightly clearer.

        Denis

From: Denis Russell
Date: April 19, 2011 4:05:06 PM EDT
To: Jeff Ogden
Cc: Gavin Eadie
Subject: Re: Reference to a lecture you gave in 1991 [Ewan]

At 17:01 -0400 18/4/11, Jeff Ogden wrote:
Thanks for contacting Ewan.

I've passed your questions on.
A couple more detailed comments of mine:

 . . .

Does he [Ewan] remember if the NUMAC 360/67 was the first 360/67 in the UK?  The first 360/67 at a UK university? You, Roger, and Elizabeth thought that it might have been, but weren't sure.

I'm as sure as can be that it was the first 360/67 at a UK University - not least because of the politics. Interestingly I do remember that there was a 360/65 somewhere (in London I think), and Cambridge got a 195 soon after - I think (*very* hazy on the dates here). However, we were definitely the only IBM machine with VM (i.e. the only 67) in a UK university for a considerable time. It might be possible to chase up the info about the other machines, but they are not really directly relevant to MTS.

        HTH

        Denis

From: Denis Russell
To: Ewan Page
Sent: Tuesday, April 19, 2011 8:55 PM
Subject: Some further queries

Ewan,
        Many thanks for your email the other day. Your comments are very useful.

Jeff Ogden has come back with a few more queries. I hope you can take the time to comment.
  • Does he remember how he learned about MTS or from whom?
    • I know Bernie Galler visited.
    • Mike Alexander came over to help install MTS.
    • And Tad Pinkerton visited when he was working for a year at Edinburgh, but I think that that was after MTS was running at NUMAC.
    • And I just finished reading Bruce Arden's memoir and learned that he visited Newcastle, sometime in the early 1970s I think, when Bernie Galler took the summer short courses that were taught in Ann Arbor on the road.
  • Does he remember if the NUMAC 360/67 was the first 360/67 in the UK?  The first 360/67 at a UK university? You, Roger, and Elizabeth thought that it might have been, but weren't sure.
  • What was the relationship with IBM with regard to NUMAC running a non-IBM OS such as MTS? Were they just happy NUMAC was allowed to buy a non-ICL system or did they push to have you run TSS or OS/360 or later MVS with TSO?  It is my impression that most of the fights that NUMAC had about continuing to run MTS were with the Computer Board, unlike the sites in the U.S. that had to fight with IBM. Is my impression more or less correct?
From: "Ewan Page"
To: "Denis Russell"
Subject: Re: Some further queries
Date: Tue, 19 Apr 2011 21:32:08 +0100

Dear Denis ,
 
I can't remember just where I first heard about MTS; if Elizabeth doesn't know for certain, no-one will.
 
Our 360/67 was not merely the first in the UK but the first in Europe - or it was going to be when we ordered it. I think it was the first installed too.
That was why I felt very exposed and said so to Eddie Nixon, the IBM head. I said that we would need their best people on the site and he promised to ensure that. I don't remember the names of the first two assigned to us but they were very good .
 
IBM would much have preferred us to run their systems but they soon accepted that we, and they, would get a bad name if we couldn't serve our users decently and they didn't press us any further - nor did they come up with the lots of free extra memory that would have been necessary to run TSS with a sensible number of terminals.
 
Rather later, when they had accepted CSS (? have I got that abbrev right ?) [probably CMS] as legitimate, they dangled that before us but my recollection (vague as it is) is that it didn't serve remote batch stations at that time, so it wasn't a possible alternative. Durham couldn't be abandoned .
 
The Computer Board weren't too opposed to MTS for quite a time - they weren't knowledgeable about operating systems, particularly IBM ones.
It was only later when Henry Chilton (?) became chairman that he leaned on us to give a slice of our 370/168 to serve Edinburgh that things became nasty. They wanted OS360, lots of batch power and in the daytime too; just not compatible with our people wanting terminal time. I think that Elizabeth acted like King Solomon and cut the 360/67 off and ran it like a 360/65 for several hours per day. Chilton just didn't understand  or didn't want to. He even complained about me to Derman Christopherson, VC of Durham, then Chairman of NUMAC. I couldn't have had a more staunch defender!
 
Ewan



Comments