Model 40 Prototype?

[Quadibloc]

In an advertisement belonging to a series "IBM reports to the
industry", on page 28 of the November, 1964 issue of Computers and
Automation (now available on Al Kossow's site), in the first column,
under the heading "SYSTEM/360 includes universal code capabilities",
there is a small photograph of more of an unusual System/360 front
panel which also appears, with better resolution, in this photo of
Fred Brooks:

http://www.computerhistory.org/revolution/mainframe-computers/7/164

I originally thought, because of the darker background of the lights,
that this was a Model 30, but looking at some details of how the
buttons and switches are arranged, it seems more likely to be a model
40.

This could be an early development prototype of the 360 series, or it
could be an alternative remote style of front panel available for
different installations.

John Savard

[Dan Espen]

Quadibloc <jsa...@ecn.ab.ca> writes:

> In an advertisement belonging to a series "IBM reports to the
> industry", on page 28 of the November, 1964 issue of Computers and
> Automation (now available on Al Kossow's site), in the first column,
> under the heading "SYSTEM/360 includes universal code capabilities",
> there is a small photograph of more of an unusual System/360 front
> panel which also appears, with better resolution, in this photo of
> Fred Brooks:
>
> http://www.computerhistory.org/revolution/mainframe-computers/7/164
>
> I originally thought, because of the darker background of the lights,
> that this was a Model 30, but looking at some details of how the
> buttons and switches are arranged, it seems more likely to be a model
> 40.

Didn't the 30 and 40 both have much taller panels?

--
Dan Espen

[Nick Spalding]

Quadibloc wrote, in
<43b6f9a7-9ce6-447a...@o2g2000yqb.googlegroups.com>
on Fri, 17 May 2013 04:04:02 -0700 (PDT):

It doesn't ring any bells with me for either the 30 or 40, both of which
I had on my patch at one time.

I don't know what the last picture on the left side of that page is but
it certainly isn't one of the TROS tapes for the Model 40 or the
controller for the 2314, which look like this:
<http://ibmcollectables.com/gallery/whittemore-1/360_40_TROS>
--
Nick Spalding

[hanc...@bbs.cpcn.com]

On May 17, 7:04 am, Quadibloc <jsav...@ecn.ab.ca> wrote:

> This could be an early development prototype of the 360 series, or it
> could be an alternative remote style of front panel available for
> different installations.

When IBM announced System/360, some of the models were still under
design, and the front panels used for display for merely mockups.
Perhaps it was just a mockup in this particular case.

IMHO, IBM announced S/360 way too early. In the end, it worked out
well as the machine was a huge success. But it put a terrible strain
on IBM employees and their families* for several years to get the
machines and software completed and out the door.

IMHO, the early announcement--done because of losses to the
competition, was anti-competitive and a possible violation of anti-
trust laws. IBM already had a huge share of the IT marketplace and
the loss of some customers might have been better for the industry
overall. It might have saved IBM from its very costly anti-trust
actionts from the govt and CDC.

*Watson Jr's memoir touches on the impact of the strain. The S/360
history also mentions it in terms of staff defections later on to the
peripheral industry.

[Quadibloc]

On May 17, 8:17 am, hanco...@bbs.cpcn.com wrote:

> IMHO, the early announcement--done because of losses to the
> competition, was anti-competitive and a possible violation of anti-
> trust laws.  IBM already had a huge share of the IT marketplace and
> the loss of some customers might have been better for the industry
> overall.

For IBM to explain to the public what it was doing, and why it wasn't
releasing new and improved models of the 1401, for example, is just
sharing factual information. But, indeed, a product announcement is a
commitment to deliver a particular product - and that can indeed be
premature.

IBM could have made a more gradual transition to the 360, by using
some resources for keeping its older lines competitive. I suspect they
did not do that because I think there is evidence to suggest that the
little-known 7070 series was a failed attempt by IBM to rationalize
its product line-up the gradual way.

John Savard

[Quadibloc]

On May 17, 8:02 am, Dan Espen <des...@verizon.net> wrote:

> Didn't the 30 and 40 both have much taller panels?

Yes, and those panels had a big box of circuitry behind them. This is
what seems to be an alternative style of front panel, which is free-
standing.

Incidentally, there are many other things of interest in those old
issues of Computers and Automation on Al Kossow's site. An ad for the
Pacific Data Systems 1020 computer led me to look up its manual on Al
Kossow's site, and I found it had yet another version of hexadecimal,
which I added to my page at:

http://www.quadibloc.com/comp/cp02.htm

Also, I saw two ads from a series of comical "historical" ads by
Computape. I haven't yet seen the one I best remembered, featuring an
ancient Egyptian named Memamadun Ptolemy. That name, of course, is a
reference to

https://www.youtube.com/watch?v=MCSqKGytGe4

a popular song hit of bygone days.

John Savard

[hanc...@bbs.cpcn.com]

On May 17, 11:01 am, Quadibloc <jsav...@ecn.ab.ca> wrote:
> On May 17, 8:17 am, hanco...@bbs.cpcn.com wrote:
>
> > IMHO, the early announcement--done because of losses to the
> > competition, was anti-competitive and a possible violation of anti-
> > trust laws.  IBM already had a huge share of the IT marketplace and
> > the loss of some customers might have been better for the industry
> > overall.
>
> For IBM to explain to the public what it was doing, and why it wasn't
> releasing new and improved models of the 1401, for example, is just
> sharing factual information. But, indeed, a product announcement is a
> commitment to deliver a particular product - and that can indeed be
> premature.

As best as I can tell from the histories, the S/360 announcement was
seriously premature, and everyone in IBM knew it. Watsor Jr said he
was troubled during the presentation, knowing some machines on display
were mere mockups.

> IBM could have made a more gradual transition to the 360, by using
> some resources for keeping its older lines competitive. I suspect they
> did not do that because I think there is evidence to suggest that the
> little-known 7070 series was a failed attempt by IBM to rationalize
> its product line-up the gradual way.

The policy principle at IBM was to make a clean break from the past
and have a single unified architecture for the four worlds of compuger
(small, big, sci-eng, business). I think at the time they had about
six distinct architectures which meant that there were six separate
software groups and collections of peripherals.

Actually, I'm not sure that IBM's temporizing efforts were so bad.
For 1401 users who outgrew their machines there was the 1410 which
seemed pretty powerful. For 7090 users, there was the 7094. Even tab
machine users had some upgrades. So, I do think customers had some
upgrade options.

I don't know when the 1403 printer was upgraded from 600 lines per
minute to 1,000 LPM--if that was available when S/360 was announced or
came along a little later. But that by itself would've been a huge
boost in sales if made available on the older machines.

(I also wonder why the 1403 kept its number despite the upgrade while
other peripherals got new numbers, like the 1311/2311 disk.)

[John Levine]

>IBM could have made a more gradual transition to the 360, by using
>some resources for keeping its older lines competitive.

Wasn't the 7094 designed and built after IBM had committed to the 360
project?

--
Regards,
John Levine, jo...@iecc.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. http://jl.ly

[hanc...@bbs.cpcn.com]

On May 17, 11:47 am, John Levine <jo...@iecc.com> wrote:
> >IBM could have made a more gradual transition to the 360, by using
> >some resources for keeping its older lines competitive.
>
> Wasn't the 7094 designed and built after IBM had committed to the 360
> project?

"An upgraded version, the IBM 7094, was first installed in September
1962."

"In April 1964, the first 7094 II was installed, which had almost
twice as much general speed as the 7090 due to a faster clock cycle,
dual memory banks and improved overlap of instruction execution, an
early instance of pipelined design."

http://en.wikipedia.org/wiki/IBM_7090

[Quadibloc]

On May 17, 9:13 am, Quadibloc <jsav...@ecn.ab.ca> wrote:
> an
> ancient Egyptian named Memamadun Ptolemy.

Apparently that ad ran in 1958, while these ads were in 1964 and
following years, so I must be mistaken.

John Savard

[Peter Flass]

On 5/17/2013 10:17 AM, hanc...@bbs.cpcn.com wrote:
> *Watson Jr's memoir touches on the impact of the strain. The S/360
> history also mentions it in terms of staff defections later on to the
> peripheral industry.
>

I was disappointed with the very small number of pages Watson devoted to
the development of the 360. I feel that it was one of his most
important accomplishments at IBM, and deserved a bit more coverage.

--
Pete

[Peter Flass]

I don't think "little-known" is accurate, but maybe that's just my
experience. Certainly the 1401 sold a lot more machines, but the 707x
was (again IME) widely used at big installations with a scientific
workload - universities, NASA, etc.

There were two reasons they didn't do as you suggest. First, they
didn't want to detract attention from the introduction of the 360 by
making it "and another system is..." Probably more importantly I don't
think they had the resources to do both. As it was the 360 caused major
strains. Haanstra wanted to produce an updated 1401, but got shot down.
Probably at another company he would have been fired, as it was he got
pushed aside.

--
Pete

[hanc...@bbs.cpcn.com]

On May 18, 8:13 am, Peter Flass <Peter_Fl...@Yahoo.com> wrote:

> I was disappointed with the very small number of pages Watson devoted to
> the development of the 360.  I feel that it was one of his most
> important accomplishments at IBM, and deserved a bit more coverage.

I don't have the book (Father, Son, & Co) handy, but I think he gave
it some coverage. He mentioned the strain on employees and the
families, the rift between him and his brother when he replaced his
brother with Learson to expedite the project, the big money tied up in
work-in-progress parts, high cost of 'clean room' factories, and the
premature announcement and mockups. He may have mentioned the
competitive threat from Honeywell's "Liberator". He mentioned
visiting a development office and a programming manager--who had a cot
there-- throwing him out so they could get some work done.

Around 1965 Fortune magazine did a piece on IBM and Watson Jr, and I
think they covered some of the S/360 travails. The same author
assisted Watson with the memoir.

_Perhaps_ Watson Jr wasn't that involved in the gory details of the S/
360 project, leaving that to his staff. If there was a problem,
Watson's role was to shuffle people and/or resources around to fix
it. From the S/360 history, we know some of those measures failed
miserably. For instance, they upped SLT chip production targets--the
manager said that wasn't possible, they fired him, and chip production
got worse.

[Walter Bushell]

In article
<bd180db1-8e74-427d...@h13g2000yqe.googlegroups.com>,

And 4 more index registers. Did the FORTRANs of the time use them?

--
Gambling with Other People's Money is the meth of the fiscal industry.
me -- in the spirit of Karl and Groucho Marx

[Quadibloc]

On May 18, 9:28 pm, Walter Bushell <pr...@panix.com> wrote:

> And 4 more index registers. Did the FORTRANs of the time use them?

I remember at least one version of FORTRAN allowed arrays having up to
seven dimensions, in contrast to a previous version which was limited
to three. Since you could only use one index register at a time in an
instruction, I'm not sure how there could be a direct connection...

John Savard

[Walter Bushell]

In article
<6e9e941f-30db-4eb2...@h13g2000yqe.googlegroups.com>,

Seven dimensions was considered more than enough. 4^7 is already half
of memory of the beast. 32K of 36 words total and it would provide
basically the computer support of thousands of engineering
professionals and support. OTOH people spent their careers plotting
graphs from computer listings which can today plop out automatically
and people make simple spread sheet mistakes that plunge the whole
frikking global economy with hideous ruin and combustion down to deep
perdition there to dwell in adamantine chains. . .. Today we could
easily work with 10^9 no hu hu.

[Shmuel Metz]

In <kn7r1c$sfj$1...@dont-email.me>, on 05/18/2013

at 08:19 AM, Peter Flass <Peter...@Yahoo.com> said:

>I don't think "little-known" is accurate, but maybe that's just my
>experience. Certainly the 1401 sold a lot more machines, but the
>707x was (again IME) widely used at big installations with a
>scientific workload - universities, NASA, etc.

The 7070 was a very nice machine, but it was more suited to a
commercial workload. The 709x was generally a more reasonable choice
than the 707x for a scientific workload.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to spam...@library.lspace.org

[Quadibloc]

On May 18, 8:34 pm, Shmuel (Seymour J.) Metz

<spamt...@library.lspace.org.invalid> wrote:

> The 7070 was a very nice machine, but it was more suited to a
> commercial workload. The 709x was generally a more reasonable choice
> than the 707x for a scientific workload.

That's true; the 707x _was_ decimal, after all. But my point is that
IBM didn't think that this completely precluded effective use for
scientific computation; both the LARC (from Univac) and the NORC (from
IBM), after all, were decimal.

John Savard

[hanc...@bbs.cpcn.com]

On May 18, 11:28 pm, Walter Bushell <pr...@panix.com> wrote:

> > "In April 1964, the first 7094 II was installed, which had almost
> > twice as much general speed as the 7090 due to a faster clock cycle,
> > dual memory banks and improved overlap of instruction execution, an
> > early instance of pipelined design."
>
> >http://en.wikipedia.org/wiki/IBM_7090
>
> And 4 more index registers. Did the FORTRANs of the time use them?

My guess is at the time more people used assembler to program the
machine than a higher level language.

For complex programs, the extra index registers likely helped speed
execution and streamlined code.

[hanc...@bbs.cpcn.com]

On May 18, 10:34 pm, Shmuel (Seymour J.) Metz

<spamt...@library.lspace.org.invalid> wrote:

> The 7070 was a very nice machine, but it was more suited to a
> commercial workload. The 709x was generally a more reasonable choice
> than the 707x for a scientific workload.

I don't know the actual production numbers of the 709x series vs. the
707x series.

But my guess is that the 7090 was a more popular machine for large-
size business applications. It seems more has been written about it.
For instance, didn't the early SABRE system run on it?

In 1963, what mainframe computer would be typically found in the
largest insurance companies and banks? Or, would much of their work
still be on tab machines and manual clerks with adding machines or
book-keeping machines?

[Dan Espen]

I was over at Continental Life in 1964.
(Just visiting.)
They had some kind of 7xxx.
I worked at a small subsidiary of Continental that used a 1440.

--
Dan Espen

[Quadibloc]

On May 19, 3:22 pm, hanco...@bbs.cpcn.com wrote:

> My guess is at the time more people used assembler to program the
> machine than a higher level language.
>
> For complex programs, the extra index registers likely helped speed
> execution and streamlined code.

Remember, though, that FORTRAN for the IBM 704 was one of the first
really well-optimized compilers.

I was just checking, but there don't appear to be restrictions on how
deeply DO statements could be nested that were changed when more index
registers were added.

John Savard

[Nick Spalding]

hanc...@bbs.cpcn.com wrote, in
<73d9423f-1f87-4dad...@z8g2000yqd.googlegroups.com>
on Sun, 19 May 2013 14:27:57 -0700 (PDT):

> On May 18, 10:34�pm, Shmuel (Seymour J.) Metz
> <spamt...@library.lspace.org.invalid> wrote:
>
> > The 7070 was a very nice machine, but it was more suited to a
> > commercial workload. The 709x was generally a more reasonable choice
> > than the 707x for a scientific workload.
>
> I don't know the actual production numbers of the 709x series vs. the
> 707x series.
>
> But my guess is that the 7090 was a more popular machine for large-
> size business applications. It seems more has been written about it.
> For instance, didn't the early SABRE system run on it?

It did, at Briarcliff NY. I was given a tour round it in 1963. Two
memories, first was the loud Klaxon sound when it crashed and the other
was the string of four 1301 disks that had to be powered up in sequence
owing to the large power drain when the very inefficient motor housed
inside the spindle started up.

> In 1963, what mainframe computer would be typically found in the
> largest insurance companies and banks? Or, would much of their work
> still be on tab machines and manual clerks with adding machines or
> book-keeping machines?

--
Nick Spalding

[Peter Flass]

On 5/20/2013 4:03 AM, Nick Spalding wrote:
> hanc...@bbs.cpcn.com wrote, in
> <73d9423f-1f87-4dad...@z8g2000yqd.googlegroups.com>
> on Sun, 19 May 2013 14:27:57 -0700 (PDT):
>
>> On May 18, 10:34 pm, Shmuel (Seymour J.) Metz
>> <spamt...@library.lspace.org.invalid> wrote:
>>
>>> The 7070 was a very nice machine, but it was more suited to a
>>> commercial workload. The 709x was generally a more reasonable choice
>>> than the 707x for a scientific workload.
>>
>> I don't know the actual production numbers of the 709x series vs. the
>> 707x series.
>>
>> But my guess is that the 7090 was a more popular machine for large-
>> size business applications. It seems more has been written about it.
>> For instance, didn't the early SABRE system run on it?
>
> It did, at Briarcliff NY. I was given a tour round it in 1963. Two
> memories, first was the loud Klaxon sound when it crashed and the other
> was the string of four 1301 disks that had to be powered up in sequence
> owing to the large power drain when the very inefficient motor housed
> inside the spindle started up.

I believe there was always a pretty big drain when disks were powered
up. IBM added a sequencer and EPO to start them up serially.



--
Pete

[Dave Pitts]

On 05/18/2013 09:28 PM, Walter Bushell wrote:
> And 4 more index registers. Did the FORTRANs of the time use them

Yes, The IBSYS Fortran IV compiler, IBFTC, did when you specified the M94
option. The option also enabled the 7094 double precision floating point
instructions.

--
Dave Pitts PULLMAN: Travel and sleep in safety and comfort.
dpi...@cozx.com