Re:The search for the lost Cray supercomputer OS (2014)
by beat on Friday May 01, 2015 @10:16AM
(#9471396)
Back in the '90s, I worked under a manager who, as an engineer, had written the linker for Cray Unix. He's retired now, but we're still in touch. I forwarded this story to him - hopefully he can help some. He went back to Cray not long after we worked together, so he might know someone still around from those olden days.
George was a great manager to work for, too... he's the one who taught me the phrase "It's better to ask for forgiveness than permission". Back in his Cray days, he was told to write a single-pass linker for arbitrary theoretical reasons by the architects. He went ahead and wrote a double-pass linker anyway, because he felt it was the Right Thing To Do. Later, customers praised the stability and debugability of Cray Unix, relative to other Unix of the era (old SunOS == suXX0r), so his "forgiveness rather than permission" paid off.
I actually worked with a whole crew of ex-Cray people in the late '90s, after Silicon Graphics bought them out and started running them into the ground. Smartest team I've ever worked with.
George was a great manager to work for, too... he's the one who taught me the phrase "It's better to ask for forgiveness than permission". Back in his Cray days, he was told to write a single-pass linker for arbitrary theoretical reasons by the architects. He went ahead and wrote a double-pass linker anyway, because he felt it was the Right Thing To Do. Later, customers praised the stability and debugability of Cray Unix, relative to other Unix of the era (old SunOS == suXX0r), so his "forgiveness rather than permission" paid off.
I actually worked with a whole crew of ex-Cray people in the late '90s, after Silicon Graphics bought them out and started running them into the ground. Smartest team I've ever worked with.