For the Old Computer types in here

Incomplete as yet, but:

Aka Multibodge

My first job was on VME/K development. It should have been canned years before. We'd just about got it working and reliable (SV18) when ICL finally did pull the plug. On a system with a larger order book than installed base... I never went to West Gorton, though Kidsgrove was a frequent trip, all the way from Bracknell.

I'm still developing S/W now. The machines have a thousand times more RAM, and run a thousand times faster. But now they fit in my pocket!

Andy

VME/K was a mess. The last version we used was SV12. SV13 was a big change and it was wholly unsuitable for us. So we moved to the homebrew operating system, which increased the MTBF from 20 hours to about 2000! And was more functional.

You might like to read my anecdotes:

some of which are VME/K (or 2900) related.

Also, have you read the 'ICL Anthology' books (which are the source of the story of the middle manager and his car)?

Don't forget Elliott Automation and English Electric Leo Marconi (EELM) which were also force-merged into ICL.

ICT also got the data processing parts of English Electric and Elliot Automation and Marconi, as the government tried to merge all the data processing expertise into one British company in the hope it would create a company which could compete in the world.

Likewise, all the process control and military computing from English Electric and Marconi was merged into Elliott Automation, which became Marconi Elliott Computer Systems briefly, and then GEC Computers a year later. GEC later also absorbed the process control and military computing from Ferranti (I think that's all that was left of Ferranti by then).

The agreements around these government driven mergers forbid ICT (ICL) from building process control systems, and forbid Elliott Automation (GEC) from building data processing systems for some years afterwards, so they would both concentrate on their areas of expertise and not try to compete with each other.

Correct!

and the coating fell off in pieces.

With all that government interference is it any wonder that the UK computer systems manufacturing has failed!

Did you also discover that Honeywell DDP 516 interface hardware had both design and construction faults?

The standard DDP516 had a 8K byte Ram.

The govt. interference was bailing out failed companies.

I did more on hacking the CPU than anything else.

And, going back a bit, The British Tabulating Machine Co. and Powers Samas!

It was already failing as far back as the mid-60s. There was a fragmented collection of companies who were each failing to produce competitive systems. The government response was two-fold - try and consolidate the industry in an effort to catch up with the Americans, and meanwhile allow the top universities and essential government departments to buy American. I.e. even by the mid-60s it was recognised that a buy-British policy meant the UK's industrial and scientific progress was being held back by substandard equipment.

I've never used an Elliot 803 although I did once win a programming contest on the one at TNMoC. But I do have a manual for it and it's an appalling mess. Absolutely no comparison with IBM manuals of the same vintage or even earlier, for clarity and usability.

I started with a bootstrap loader an assembler and Honeywell diagnostic programmes that in many cases did nothing. Most of the diagnostics didn't use interrupts. In most cases the hardware was wrong and either didn't trigger interrupts or triggered them at the wrong time. I wrote a simple OS and drivers for most devices before getting on with what I really wanted to do