> In reading histories of IBM and the Bell System, my impression is that
> the companies were pretty distant from each other even though both
> were developing very similar technologies. Early on, both Bell and
> IBM were developing ever better ways of using relays to "think" in
> sophisticated ways, then using electronic components. (IBM obviously
> did go to Bell Labs to learn about the transistor).
The Bell System had Bell Labs, a research organization that did pure
research which turned into developments which Bell rapidly
incorporated into their products.
IBM had the T.J. Watson Center, a research organization that did pure
research which turned into developments which IBM's competitors
rapidly incorporated into their products and IBM ignored for the most
> Anyway, the Bell Labs history says Bell did make use of the IBM
> System/7 as part of the switching network. The S/7 was a process
> controller machine, kind of a sideline of IBM's normal business line.
> Anyway, Bell used the S/7 to replace AMA (long distance message
> accounting) machines. Even here the S/7 was eventually replaced with
> a PDP machine.
> Would anyone know if there was some sort of hostility between Bell and
> IBM in the 1950s and 1960s? Or, am I just missing that there was a
> lot of collaboration?
No, but I do know that IBM was not very good at the whole "small
computer" thing. The minicomputer revolution escaped them completely
for the most part, which is why DEC systems wound up being
incorporated into switching systems. I am sure that if IBM had made
minicomputers that actually worked well a decade before the Series/1,
the phone company (and a lot of other companies) would have bought
> Perhaps the lab histories of both companies prefers to focus on the
> company's own developments and ignore those elsewhere. The IBM
> history does give credit to semi-conductor makers. I sense Bell
> wanted to build everything it used for itself rather than buy finished
> products in the market.
Yes, which is why those DEC minicomputers later got replaced with AT&T
3b2 and 3b20 systems. Bell had a very strong impulse toward vertical
and horizontal integration.
> I also wonder if the commercial computer components of the 1960s (ie
> System/360 SLT chips and core memory) were adequate for the speed
> demanded by electronic switching. The Bell history suggests Bell had
> to develop its own gear because it needed faster speed and memory
> available in the commercial world on a cost- efficient basis. I
> believe an ESS of 1965 had quite a bit of memory and would compare to
> the largest commercial computers of that day.
Well, the early ESS systems were very far from general purpose
computer systems. There was a whole lot of combinational logic inside
there. As general purpose computers got cheaper and more powerful,
ESS systems evolved toward using them for control.
"C'est un Nagra. C'est suisse, et tres, tres precis."