On 19/04/2018 17:07, Martin Gregorie wrote:
> On Thu, 19 Apr 2018 16:42:39 +0100, mm0fmf wrote:
>
>> Time does strange things to memory... I worked at a company that had a
>> pair of 1905's in 1983. One just ran in batch mode and the other was
>> used interactively during the day. There were plenty of the funny orange
>> cased ICL VDUs about the place. The 1905's sat in the air-conditioned
>> room along with an LSI 11/23+. I was writing stuff for assorted Apricot
>> PCs and the LSI 11 and got to the room regularly to load 9trk tapes etc.
>> I remember them being in quite dark grey cabinets unlike the orange on
>> the VDUs and the creamy-beige of the LSI 11.
>>
> I've never seen a 1905, the biggest 1900 I ever saw or got my hands on
> was a 1904S at British Steel's Battersea labs. I was sysadmin on that for
> a bit under a year in 1977.
>
>> I can remember being there when one of the 1905's was being rebooted and
>> watching the questions being answered on an ASR33 console to boot the OS
>> (GEORGE III?).
>>
> Almost certainly George 3, a surprisingly good and advanced OS for its
> age (it was a late 60s design that I used/maintained from about 1971 to
> 1978. It was one of the first OSes to have a hierarchic filing system.
>
>> I can clearly remember one of the 1905's
>> had 256k of memory (can't recall if it was core)
>>
> Probably ferrite core. That would be 256Kwords - 1900s all had 24bit word
> addressing and, for character strings packed four 6-bit ISO characters
> into each word.
>
>> and whilst the bootup
>> guys were working, there was a panel open and there was a wall of lights
>> (incandescent bulbs). There were few lit whilst the questions were
>> answered and then after return was hit, the room lit as the lights came
>> on. Some flashing and it all went dark. Lather, rinse, repeat a few more
>> times till it stayed up and the jobs were re-queued.
>>
> As I said, I never used a 1905 - it was just a 1904 with a hardware
> floating point processor, and was dropped after the 1905E - only the
> later S and T series machines had semiconductor memory.
>
>> Having read the comments, I'm wondering what I was looking at? If it
>> wasn't the CPU cab for the 1905 because there were almost no lights,
>> what was it?
>>
> Almost certainly and engineering panel and possibly specific to the
> 1905/6 models since those were the only models to have hardware FP (and
> the 1906 has hardware registers too: everything else used the first few
> words of each program as its registers.
>
>> I really don't think it was 2900 series
>>
> They were all burnt orange, even the 2903, a desk-sized box that ran on
> an office environment. It was actually a 2900DFC (Disk File Controller)
> running firmware that emulated a 1900 and ran bog standard 1900 software
> under George 1S (a modified version of the single-streaming George 1 OS.
> George 1 had nothing at all in common with George 3.
>
>> Answers on the back of an unused punch card.
>
> :-)
>
>
Now you said S and T series and 1905_S* has jumped out of assorted
neuron triggers. It could well have been semiconductor memory because
another guy there said something along the lines of "each cab contains
256k x 6bits of memory".

But oh, 35years does things to your memory. I never really paid much
attention, it was an awful place to work but I needed the cash and did
have my own LSI 11 to play with!

--- SoupGate-Win32 v1.05