26 Aug
2007
26 Aug
'07
6:49 p.m.
SIP wrote:
Michael Grigoni wrote:
<snip>
>Yes, Seymour liked to code in machine language
right from the console
>and the O/Ses for his machines developed 'organically' from the bottom up
>in general, but wouldn't you consider SCOPE, KRONOS, NOS etc. operating
>systems? And as for hardware, the CDC peripherals and Hyperchannel
>offerings seemed to be the best in the industry on Cray CPUs...
Right, but Cray himself was TOTALLY against their
creation.
<snip>
He was certainly well-known for a conservative engineering approach,
avoiding bleeding-edge technology (e.g. Integrated Circuits, later LSI)
and pushing mature technologies to incredible performance, but I had not
heard of a philosophical objection to making peripherals and operating
systems even though it was evident he felt it was the domain of others.
Then, again, when Cray Research was faced with the
decision of whether
to build computers that people would buy or computers for the sake of
building computers, they understood that they needed to stay in
business, and decided on the path that would lead them to solvency.
This caused Cray to leave his own company and form Cray Computers.
<snip>
Interesting. I remember the public face of this was that he needed to avoid
sole-source procurement problems with big government contracts so he formed
Cray Computers to the mutual benefit of both companies.
His nephew was a classmate of mine and once invited me to a visit to his uncle's
lake home; I wish I had been able to go (sadly too late now).
Regards,
Michael