ICL 190x computers
John Machin
generated music dragged my mind back to the late sixties when I
worked on an ICL 1905.
Here are some random recollections:
Music: the machine had an amplifier on the side of the console; this
emitted a peep whenever a branch, call, etc was executed. This
facility was heavily used and abused; the latter to the extent that a
45rpm record was produced of the machine playing (among other things)
"Colonel Bogie" and the national anthem.
Punched cards: We had a manual (i.e. no electric power) 12-key card
punch (manufactured by ICL Hindustan Ltd.) which was used for "just
one card" emergencies. There were many "proper" key-punch machines in
a cave in the mountains, the entrance to which was guarded by a
fearsome dragon to whom one submitted punching requests. Legend had
it that the machinery therein was operated by nubile maidens, but the
dragon slew all who attempted to verify the story.
A good trick, when one had made a mistake in hand-punching a card, was
to lay the card flat on the desk, put a chad (the small piece of card
punched out of the hole) into the hole, and rub one's thumbnail over
it to ensure a good fit. This usually lasted for two or three trips
through the card reader before it fell out and caused havoc.
The card reader would sometimes develop a problem in the area where
cards were supposed to flop into the stacker when they came out of the
upwardly-sloping read-path: the cards would be bent through an angle
of about 45 degrees and fly into the air like a cloud of bats coming
out at dusk.
Another problem with the card reader was really an operator problem:
when a job had to be aborted, the operator would press the "runout"
button; this would flush the two cards then in the read-path into the
reject stacker. This then gave the operator six choices of an order
in which to re-assemble the card deck from the three sources (input
hopper, reject stacker, main stacker). Empirical observation led to
the conclusion that the chance of the one correct order being selected
was much less than one-sixth. Victims of this behaviour were said to
have had the "two-card trick" played upon them.
If there are enough ICL 190x aficionados out there, maybe we could
have our own newsgroup ... for a real special-interest newsgroup, how
about comp.lang.PLAN? :-)
P.S. Is there anyone else out there who remembers what the DISTY,
SUSTY, and DELTY extracodes did?
Ian G Batten
> Recent threads whose subjects were punched cards and computer-
> generated music dragged my mind back to the late sixties when I
> worked on an ICL 1905.
Nostalgia time! I used a 1902 when I was a child as a hanger on at
Birmingham Poly, a 1906A running the dreaded MAXIMAP (#QMMA) via a
10cps teletype whilst I was at school and arrived at University just as
they were shipping out one of the last of the 6As.
#QMMA was a masterpiece of a timeshare system. The schools system had a
small directory (flat, with four character file names!) plus a set of
large scratch directories called COMMON1 through to COMMON9. The
problem here was that any other user could gun your stuff.
We found that the checking for valid filenames was done IN THE
UTILITIES. The valid character set was actually embedded in the binary.
Therefore, if you patched your copy of COPY so that $ was valid, you
could create files called XYZ$. Other users couldn't do anything to
this, because the library routines would complain about invalid
filenames. Such fun.
> SUSTY, and DELTY extracodes did?
SUSTY is SUSpend and TYpe, which allowed you to emit as message rather
than a two character code. DELTY is DELete and TYpe, in the same
fashion. I always enjoyed being able to explicitly wait for I/O to
complete with the SUSBY instruction.
And I still think of #UPPER and #LOWER when allocating memory.
ian
Ian Kemmish
>Recent threads whose subjects were punched cards and computer-
>generated music dragged my mind back to the late sixties when I
>worked on an ICL 1905.
>If there are enough ICL 190x aficionados out there, maybe we could
>have our own newsgroup ... for a real special-interest newsgroup, how
>about comp.lang.PLAN? :-)
>P.S. Is there anyone else out there who remembers what the DISTY,
>SUSTY, and DELTY extracodes did?
Sure do.
DISTY displayed a message on the console typewriter.
SUSTY suspended the program and typed a message (I think two
characters only) on the console, for the operator to react to
(Operating systems? What are they?)
DELTY did something else in the same vein.
I think DISTY let you type more than two characters, but made you suffer
for it in the way you specified the message.
WE use to (ab)use a 1904E at the local Tech, when we were at
Grammer School. I remember I was writing a chess program
which got up to three boxes of cards before the operators revolted.
To a teenageer, the female operators were almost as much of a pull
as the computer.
I still think the 1900 series was a better architecture to *learn*
assembler on than the 360.
--
Ian D. Kemmish Tel. +44 767 601 361
18 Durham Close uad...@dircon.UUCP
Biggleswade ukc!dircon!uad1077
Beds SG18 8HZ United Kingdom uad...@dircon.co.uk
Henry Black
> have our own newsgroup ... for a real special-interest newsgroup, how
> about comp.lang.PLAN? :-)
>
anywhere, or maybe a 2903 still in existance?
> P.S. Is there anyone else out there who remembers what the DISTY,
> SUSTY, and DELTY extracodes did?
How about some rarer 1900 stuff:
SUSAR and the mark one subprogramming system.
F.D.S., WGTP execs, 1902/3 Dual programming, 1901A T.E.D.S.
1911 Haystacker, 7010 Uniplexer, I-O Generator, PATSY,
M.C.S., The 1004 link, 40 Column cards, The GE/Lundy MICR,
oooooo I could go on....
But does anyone know the real story behind the FP6000 (reputedly the
Ferranti-Packard Model 6000 computer and the grand-daddy of them all)?
I was never able to get to the bottom of that one.
If only we'd built a superscalar 1900 instead of getting mixed up with
that 8 bit EBCDIC nonsense.......
-------------------------------------------------------------------------------
Henry Black hen...@hpspd.HP.COM +1 415 857 6655 KK6JR@W6PW-3.#NOCAL.CA.USA.NA
Obviously my views and not those of my employer.
Mike Lawrie
>In article <2...@rand.mel.cocam.oz.au> s...@rand.mel.cocam.oz.au (John Machin) writes:
>> Recent threads whose subjects were punched cards and computer-
>> generated music dragged my mind back to the late sixties when I
>> worked on an ICL 1905.
The first computer music that I came across was on an ICT 1301, ca 1963.
The standard tune was Any Old Iron. This had to be hard-coded (ie hard
to code!) in numbers built into the program instructions, 12 columns
of card having two instructions in Fastfeed (?) format. Later came
Arne's compiler, that made music coding much simpler. The basic blip
to the speaker was produced by a successful test of an indicator (ie
program condition, such as 'test if zero').
>Nostalgia time! I used a 1902 when I was a child as a hanger on at
>Birmingham Poly, a 1906A running the dreaded MAXIMAP (#QMMA) via a
>10cps teletype whilst I was at school and arrived at University just as
>they were shipping out one of the last of the 6As.
The program was called MAXIMOP, not ...MAP. MOP stood for Multiple
On-line Programming. Minimop was ICL's idea of low-cost simple
interactive computing, Maximop was developed by Arthur Dransfield
of Queen Mary's College. It had a goodly number of enhancements,
and no bugs as far as we could tell.
Later versions of Maximop were supplied by ICL, but maintained at
QMC. There was an interface to ICL's George 2, to allow batch work,
as well as Maxibatch that did a similar thing. We ran all three very
successfully on an ICL 1902S, later on a somewhat unreliable 1904S.
Last execution was in 1983, when a Cyber took over.
Remember George? GEneral ORGanisational Environment, but I don't
remember the surname of the fellow George who developed it.
George 3 had bells and whistles, and had the reputation of
requiring all of the CPU cycles to run. George 2 was much simpler,
and to us was the only practical way to operate the machines if we
were to use an operating system at all.
And do you remember E3RM and 36RM, the ICL 190x Executive programs?
QMC had good expertise in E6RM as well, and added enhancements
that dovetailed well into Maximop features.
>#QMMA was a masterpiece of a timeshare system. The schools system had a
>small directory (flat, with four character file names!) plus a set of
>large scratch directories called COMMON1 through to COMMON9. The
>problem here was that any other user could gun your stuff.
This must have been a local decision to install Maximop this way,
it did not have to be so. But security was not one of its strong
points anyway. Withing Maximop we had pretty good security, but
unfortunately many utilities (eg XMED) could be run outside of the
operating system, and these could read the so-called private Maximop
files.
.....................................................................
Mike Lawrie
Director Computing Services, Rhodes University, South Africa
......................<cc...@hippo.ru.ac.za>..........................
Rhodes University condemns racism and racial segregation and
strives to maintain a strong tradition of non-discrimination with
regard to race and gender in the constitution of its student body, in
the selection and promotion of its staff and in its administration.
--
....................................................................
Mike Lawrie
Director Computing Services, Rhodes University, South Africa
.....................<cc...@hippo.ru.ac.za>..........................
Nick Felisiak
>
>Remember George? GEneral ORGanisational Environment, but I don't
>remember the surname of the fellow George who developed it.
Felton, I think.
>George 3 had bells and whistles, and had the reputation of
>requiring all of the CPU cycles to run. George 2 was much simpler,
>and to us was the only practical way to operate the machines if we
>were to use an operating system at all.
There was also George 4. This had *virtual memory* and only ran on
1906's.
The 1900 range didn't end with the 2903 (which someone referred to
in a previous posting). It lived on in the ME29, which was a microcoded
machine based on the Stanford EMMY architecture (mid - late 1970).
Cute machine - I wrote a pile of microcode for it. There were (of course)
various models in the range. The first two were just (or mainly)
different microcode. The top of the range had an 'accelerator'; in
other words, they plugged some 1900 instructions into it as a co-processor!
The 1900 order code also ran on the smaller (microcoded) 2900s. In some
cases, faster than the 2900 order code did!
Interesting. I've just noticed I typed 'order code' rather than 'instruction
set'. Strange how the right terminology automatically got recalled!
Nick
--
Nick Felisiak ni...@spider.co.uk
Spider Systems Ltd +44 31 554 9424
Joseph M. Newcomer
:-)
joe
Steve Linton
The 1A just ran E3RM and programs from paper tape and cards, the '2T
eventually ran MINIMOP and a cluster of 8 terminals. You could watch the
time-slicing in progress: first my ALGOL60 compilation produced a page
of listsing then his ALGOL60 compil...
We got to recognise the characteristic sounds (on the console speaker) of
the system utilities. The COBOL compiler made a particularily unpleasant
noise, I recall.
One particularly interesting OS feature we came across was the 3/4 memory
switch. If a particular wire (the one and only yellow wire in fact) was
present on the backplane the bootstrap loader would disable the top 1/4
of the supplied memory. Thus the 48K -> 64K upgrade simply consisted of
cutting the yellow wire. List price: 6000 pounds.
Geoff Lane
Hmmm, lets see... A non-IBM operating system running on non-IBM compatible
hardware... No it must have been for some other reason :-)
Seriously, George 3 was, for its time, a very nice OS. It had a
very powerfull macro language (IMHO better than most Unix shell
languages even now) at a time when IBM JCL was begining to take over
the world.
Although not totally open, the sources for the OS were not impossible
to get hold of and as a result there were a number of hacked about
versions in various CS departments. It was even possible to modify
the hardware - Manchester University CS had a 1904S which had been
modified to perform paging (at the time a very rare facility - IBM had
not yet ``invented'' virtual memory!)
--
Geoff. Lane. Janet: zza...@uk.ac.mcc.cms
UTS Sys Admin, Manchester Computing Centre, Oxford Rd, Manchester, M13 9PL
k...@wongaman.demon.co.uk
I was a Plan man (XPLT) and went on to work for ICL supporting DME.
Bob Eager
summers working as a programmer and/or operator for the University of
Sussex, also for the Advance Linen Company (roller towels, etc.)
I didn't see John's original post for some reason; only this followup.
k...@wongaman.demon.co.uk
Yes smaller 2900s supported 1900 programs via DME or concurrently with VME under CME. Low end 3900s supported 1900 type programs under CME*
Bob Eager
> The top of the range ME29 was the ME29/54 which had the hardware
> decoder. The other modes where ME29/45 & ME29/37. TME was a good
> operating system, allowing traditional 1900 programs to run and offering
> advanced facilities (RAPID database, TP & MAC). And it supported
> filestore and a decent JCL. A non released version of TME was on the
> drawing board; I think it was called SUPREMO.
> Yes smaller 2900s supported 1900 programs via DME or concurrently with
> VME under CME. Low end 3900s supported 1900 type programs under CME*
software decode; I found a bug in the hardware decode which ICL would not
accept (or could not understand). I ended up writing my own microcode
patch to work round it.
We didn't run VME. Or DME. Or CME!
http://www.tavi.co.uk/icl/bob.htm
Dave Garland
> On Tue, 14 Jul 2015 07:54:46 -0700, kas wrote:
>
>> On Monday, April 22, 1991 at 9:49:22 AM UTC+1, John Machin wrote:
> I didn't see John's original post for some reason; only this followup.
>
Kerr Mudd-John
--
Bah, and indeed, Humbug
lawr...@cluon.com
> On Monday, April 22, 1991 at 9:49:22 AM UTC+1, John Machin wrote:
[many lines of deletia]
> Type. I also remember a 160 x=7 which accepted an MS= from the
> console. I was a Plan man (XPLT) and went on to work for ICL
> supporting DME.
alt.folklore.computers.usenet.meta
:)
NK1G
Bob Eager
Kerr Mudd-John
Stephen Wolstenholme
>A followup to an a.f.c. post that is 24 years old --
>alt.folklore.computers.usenet.meta
I worked for ICL for over 30 years but never on 190x
Steve
--
Neural Network Software for Windows http://www.npsnn.com
EasyNN-plus More than just a neural network http://www.easynn.com
Stephen Wolstenholme
>On 2015-07-16, Stephen Wolstenholme <st...@easynn.com> wrote:
>> On Tue, 14 Jul 2015 22:12:33 +0200, lawr...@cluon.com wrote:
>>
>>>A followup to an a.f.c. post that is 24 years old --
>>>alt.folklore.computers.usenet.meta
>>
>> Some topics should be resurrected occasionally.
>>
>> I worked for ICL for over 30 years but never on 190x
>
>1904S Autocoder system to a 2900 (can't remember which one - a small
>one) and RPG2. This was ca. 37 years ago. The main things I recall from
>then were (i) how much I loathed RPG2 & (ii) when we parallel ran
>the new system, the results were different from the old one. After a lot
>of "headless chicken" behaviour, it turned out that the old system was
>the one that was wrong.
>
>Does anyone write payroll systems any more?
Only my own. I pay myself!
A long time ago a payroll programmer I knew found an inconsistency
with a built in test. It was some sort of check that had to balance
with the pay packets coin inserts. He gave the problem to one of the
systems programmers who found it was a real mag tape check-sum being
incorrectly calculated by a tape deck controller. I was the engineer
at the time. We dug all the way down into the controller microcode
that did the calculation and discovered it was a design fault. The
designer had retired and no one else had any idea how the get the
problem fixed. We gave the problem back to the payroll programmer
saying he would have to make an allowance. He wasn't a happy man but
after bribing him with beer he asked to see how we had decided it was
a microcode problem. Within a few minutes he came up with a microcode
patch that fixed his payroll. Now that's a payroll programmer!
jmfbahciv
A lot of problems were solved with buying someone a beer.
/BAH
Charlie Gibbs
> At my current employers we needed to get a change made to a piece of
> IBM360 Assembler. The last Assembler programmer (and the one who wrote
> this) retired several years ago. I took a quick look at it and worked out
> what the change needed to be, even though I've never worked on a 360
> or seen Assembler before, although I have programmed in several assemblers.
>
> For some reason they wouldn't let me make the change. After all, the code's
> only called for every single job that runs on our production mainframe.
Ain't politics grand? Even though the Powers That Be have
prohibited a fix, they no doubt won't stop complaining and
making others' lives miserable because the system doesn't
work right.
--
/~\ cgi...@kltpzyxm.invalid (Charlie Gibbs)
\ / I'm really at ac.dekanfrus if you read it the right way.
X Top-posted messages will probably be ignored. See RFC1855.
/ \ HTML will DEFINITELY be ignored. Join the ASCII ribbon campaign!
Charlie Gibbs
>>
>>> A followup to an a.f.c. post that is 24 years old --
>>> alt.folklore.computers.usenet.meta
>>
>> Some topics should be resurrected occasionally.
>>
>> I worked for ICL for over 30 years but never on 190x
>
> 1904S Autocoder system to a 2900 (can't remember which one - a small
> one) and RPG2. This was ca. 37 years ago. The main things I recall from
> then were (i) how much I loathed RPG2 & (ii) when we parallel ran
> the new system, the results were different from the old one. After a lot
> of "headless chicken" behaviour, it turned out that the old system was
> the one that was wrong.
> Does anyone write payroll systems any more?
My philosophy is that payrolls should never have been put
on computers, because computers are logical and payrolls aren't.
Dan Espen
> On 2015-07-16, Huge <Hu...@nowhere.much.invalid> wrote:
>
>> On 2015-07-16, Stephen Wolstenholme <st...@easynn.com> wrote:
>>
>>> On Tue, 14 Jul 2015 22:12:33 +0200, lawr...@cluon.com wrote:
>>>
>>>> A followup to an a.f.c. post that is 24 years old --
>>>> alt.folklore.computers.usenet.meta
>>>
>>> Some topics should be resurrected occasionally.
>>>
>>> I worked for ICL for over 30 years but never on 190x
>>
>> I helped to port & rewrite my then employers payroll system from a
>> 1904S Autocoder system to a 2900 (can't remember which one - a small
>> one) and RPG2. This was ca. 37 years ago. The main things I recall from
>> then were (i) how much I loathed RPG2 & (ii) when we parallel ran
>> the new system, the results were different from the old one. After a lot
>> of "headless chicken" behaviour, it turned out that the old system was
>> the one that was wrong.
>
> I always referred to it as "decapitated fowl mode".
>
>> Does anyone write payroll systems any more?
>
> There must be someone, but thank God it isn't me anymore.
> My philosophy is that payrolls should never have been put
> on computers, because computers are logical and payrolls aren't.
Back in the early days, most companies ran payroll in house.
I supported a payroll system where the printing and file reading was
written inhouse but I'd periodically get a new COBOL calculation routine
from an outfit named "ALLTAX" (pretty sure).
I'd get a bunch of cards in the mail, insert them into the master deck,
and run a test.
Now the whole process is farmed out.
--
Dan Espen
Freddy1X
> On 2015-07-16, Charlie Gibbs <cgi...@kltpzyxm.invalid> wrote:
>> On 2015-07-16, Huge <Hu...@nowhere.much.invalid> wrote:
>> prohibited a fix, they no doubt won't stop complaining and
>> making others' lives miserable because the system doesn't
>> work right.
>
> I told him to (i) charge for a day's work & (ii) think of a number, change
> up to the next unit of measure and then triple it.
>
>
(iii) Wait for the check to clear and have him write back that your solution
would have worked as well.
Freddy,
told 'ya!
--
See disclaimers on product box.
/|>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>\|
/| I may be demented \|
/| but I'm not crazy! \|
/|<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<\|
* SPAyM trap: there is no X in my address *
Stephen Wolstenholme
>On 2015-07-16, Freddy1X <fred...@indyX.netx> wrote:
>> Huge wrote:
>>
>>> On 2015-07-16, Charlie Gibbs <cgi...@kltpzyxm.invalid> wrote:
>>>> On 2015-07-16, Huge <Hu...@nowhere.much.invalid> wrote:
>>
>> ( cuts )
>>>> Ain't politics grand? Even though the Powers That Be have
>>>> prohibited a fix, they no doubt won't stop complaining and
>>>> making others' lives miserable because the system doesn't
>>>> work right.
>>>
>>> Oh, we got the retired developer back to do it. Took him about 10 minutes.
>>> I told him to (i) charge for a day's work & (ii) think of a number, change
>>> up to the next unit of measure and then triple it.
>>>
>>>
>> (iii) Wait for the check to clear and have him write back that your solution
>> would have worked as well.
>
That's when you start to wonder how you ever had the time!
jmfbahciv
> On 2015-07-16, Dan Espen <des...@verizon.net> wrote:
>
> [39 lines snipped]
>
>>
>> Now the whole process [payroll] is farmed out.
>>
>
> I suspect virtually everyone outsources it these days.
ADP.
/BAH
jmfbahciv
> On 2015-07-16, jmfbahciv <See....@aol.com> wrote:
>
> [40 lines snipped]
>> A lot of problems were solved with buying someone a beer.
>
Good. With today's PC insanity, I figured the usual methods of solving
the pesky bugs went into the bit bucket.
>
> Something that is nearly always overlooked in outsourcing deals is just how
> much work gets done "out of band", because Sid knows Joe, when beer changes
> hands.
Or yelling over the office wall.
/BAH
Bob Eager
> On 2015-07-16, Freddy1X <fred...@indyX.netx> wrote:
>>
>>> On 2015-07-16, Charlie Gibbs <cgi...@kltpzyxm.invalid> wrote:
>>>> On 2015-07-16, Huge <Hu...@nowhere.much.invalid> wrote:
>>
>> ( cuts )
>>>> Ain't politics grand? Even though the Powers That Be have prohibited
>>>> a fix, they no doubt won't stop complaining and making others' lives
>>>> miserable because the system doesn't work right.
>>>
>>> Oh, we got the retired developer back to do it. Took him about 10
>>> minutes.
>>> I told him to (i) charge for a day's work & (ii) think of a number,
>>> change up to the next unit of measure and then triple it.
>>>
>>>
>> (iii) Wait for the check to clear and have him write back that your
>> solution would have worked as well.
>