Talk:ICL VME

Jazzman0570 (talk) 09:03, 18 April 2011 (UTC)

MEEP
I think that B/MEEP deserves a mention here somewhere, as well as the VME/K equivalent (an Order Code board?). Is there any information out there? There is an article stub about DME but it looks very lonely! Perhaps a page combining the various 1900 emulation environments available on 2900s, S39s etc. is needed. TraceyR 13:58, 2 March 2007 (UTC)

'West Gorton' Link
presumably this is the same place as west Gorton? do you think anyone's ever going to write an article for West Gorton separately? it's not an enormous or in any way monumental place really... 148.79.162.143 (talk) 14:22, 21 November 2007 (UTC)

Burroughs Large Systems
Many innovative concepts from the earlier Burroughs large systems integrated software-hardware architecture and MCP operating system are evident in the design principles of the ICL 2900 Series and VME. For example: a system language based on Algol, direct system function calls, co-operative multiprocessors, hardware stacks, hardware semaphore instructions and memory words with control tags. Taken together, it seems probable that a designer in ICL knew of and preferred the intelligent Burroughs design to the prevailing brute force IBM approach. Who was it? --Vendeka (talk) 19:00, 9 September 2008 (UTC)


 * I get the impression that there was a lot of work surveying the state of the art at the time, and it's unlikely that you can identify one individual who brought this particular perspective. Certainly the Burroughs architecture is acknowledged as one of the influences (see for example Brian Procter's talk at http://www.computer50.org/mark1/mu5.html - but some of it might have come indirectly via the Manchester MU5.

It might be an idea to mention the further progression of VME into a service called VAaaS - Link here http://www.fujitsu.com/uk/services/application-services/application-development/vme/index.html


 * Mhkay (talk) 23:30, 9 September 2008 (UTC)


 * I don't have ready access to it but this paper does a comparison of the two architectures: http://portal.acm.org/citation.cfm?doid=641746.641748. The abstract claims that "The two systems are based on similar principles and are similar in overall design. However, although the ICL2900 postdates the B6700 by seven years, the machines have independent origins and differ in many important details."


 * The other thing I would add is that it was definitely part of ICL's culture not to simply copy IBM. This was partly I think pure technical pride, partly business strategy that you couldn't beat IBM if you played the game by IBM's rules.


 * Mhkay (talk) 22:20, 10 September 2008 (UTC)

Requested screenshot
Someone has added a tag asking for a requested screenshot. Could he or she please explain why? What would you expect to see on a screenshot of a 1970s operating system? Half the time in the early days, you ran the system from punched cards, not from a screen at all.

Mhkay (talk) 23:39, 28 June 2009 (UTC)
 * File:ICL 7561 terminal.jpg is available, if that's any good? I know that it's connected to an ICL 2966 which is probably running VME (though I'd have to go and ask to find out to be sure of that) so could see whether the operators could put some sort of logon screen or console output on it. I do agree with the question, though, what use would it serve? —ClickRick (talk) 23:58, 28 June 2009 (UTC)


 * Any photos of the console output (logon screen or command-line interface) would be fine. Here are some examples of what I mean with 'screenshot': CP/M, 86-DOS, MS-DOS, Acorn MOS etc. Thanks! Ghettoblaster (talk) 00:06, 29 June 2009 (UTC)

IPA
Anyone care to write a related article on ICL's Information Processing Architecture? Including the underlying bearers X.25, OSLAN and CO-3, the relationship to OSI and the applications such as RSA (Remote Session Access), FTF (File Transfer Facility, based on NIFTP), DP (Direct Print), DAF/DTS (Distributed Application Facility/Distributed TP Service, ICL's equivalent of IBM CICS), RVC (Remote Virtual Console), ADI (Application Data Interchange) and CM (Community Management). —Preceding unsigned comment added by 217.154.240.195 (talk) 07:41, 4 September 2009 (UTC)

VAaaS
It might be an idea to mention the further progression of VME into a service called VAaaS - Link here http://www.fujitsu.com/uk/services/application-services/application-development/vme/index.html  Jazzman0570 (talk) 09:09, 18 April 2011 (UTC)

VME/K
VME/K is scarcely mentioned in this article, although it was provided to many customers running smaller 2900s. I'm sure that there is still a lot of knowledge of the role and history of VME/K out there and that the article would be improved by a section covering this OS. Surely there's much more that could be written about B too, and e.g. B-MEEP (microcoded 1900 and System 4 order code implementation). Was there a 1900 order code board (i.e. hardware) in some smaller 2900 systems? --TraceyR (talk) 12:19, 11 July 2011 (UTC)


 * No point bemoaning the absence of material here: just contribute it. (But not from memory, please - the tricky part is finding documentary sources that back up your recollections.) Mhkay (talk) 23:30, 13 July 2011 (UTC)


 * Well, I've got an original KCL design guide here, on on SFL (job control language) and design documents on the file system and backup software. I haven't read them in 40 years... Perhaps I should donate them to Bletchley or something... but I'm retiring in a few months, I might be able to do something with them then. Number774 (talk) 22:01, 5 January 2022 (UTC)

"VME is in many ways more modern in its architecture"
"As a creation of the mid-1970s, with no constraints to be compatible with earlier operating systems, VME is in many ways more modern in its architecture than today's Unix derivatives (Unix was designed in the 1960s) or Microsoft Windows (which started as an operating system for single-user computers, and still betrays those origins)." is one of those statements that I think a (US) court would call puffery; you can't call it false advertising, because it doesn't really mean anything. The first note is mostly false; the first thought about Unix was in 1969, and whereas the article puts the start of ICL VME in 1968. This article puts the first release of ICL VME as "mid-1970s", which is about the same time as Unix's 1974, so they were being designed and released at about the same time. The comment on Windows might be a little more fair, but it was designed in the early 1980s, and thoroughly redesigned for Windows NT in the early 1990s. It's also weird to say that a mainframe OS from the 1970s is more modern than a single-person OS from the 1980s, especially given the number of single-person PCs and cloud systems running in VMs that look like single application PCs from the inside.--Prosfilaes (talk) 06:59, 11 October 2021 (UTC)
 * Windows NT wasn't just a redesign. It was a rewrite from the ground up. (That's where I went after K was cancelled - ICL's PC servers) Number774 (talk) 22:08, 5 January 2022 (UTC)

Cost of upgrades
This is just my memory of chatter in the building at the time, but the story I heard was the the decision to go with B not K was driven by advice from the B side of the company. There were a few odd things (CAFS, DAP) which K didn't support. We were told that the sums included that HW upgrade from a system running K to get the same workload on B would need one more disc drive, and an extra meg of memory. A meg was a lot on those days, a cabinet 3ft square and 8ft high.

The story was that actually it needed doubling the memory. That was mostly right, as most K systems had 1MB.

But there was a slight problem that B needed about 10% more CPU, which was an issue on those systems that were using it all. Number774 (talk) 22:08, 5 January 2022 (UTC)