Re: [TUHS] 516-TSS, was re: Joe Condon

Dennis Boone writes: Well, if I could find my org chart I would probably find it with the 516-TSS docs. Here's a little bit of what I can remember... Our machine had a whopping 12K 16 bit words of core memory, and a 500K disk. I think that at some point we added another 4K of expensive but faster DRAM. This sticks in my mind because it was fragile, and when we moved from building 2 to the newly constructed building 7 someone in the group decided to hand-carry the memory but got caught by the union movers who lectured him on taking food out of the mouths of their families. I believe that much of the work on 516-TSS was done by Carl Christensen and Heinz Lycklama. Since Heinz is still alive you could ask him if he remembers anything about it. The 516 architecture included an index register. This was used as the base address register for the VM system. I recall that there were special acrobatics that I had to go through if I actually needed to use the index register, probably involved turning off interrupts or something. The system was multi-tasking, and was the main machine that the explorer scouts used. This is how I got involved before my youthful exuberance and the fact that I was cutting school and hitchhiking up to the labs so that I could play with it on non-scouting days got me hired. The coolest thing about the system was "the loop" which I think was designed by Dave Weller and Sandy Fraser. It was an early LAN. I remember that it had special repeaters using the hot tech of the time, 74S00s. Since it was a loop configuration everyone had in and out jacks in their offices and labs. When it was meeting time and people weren't showing up Joe would unplug his to take the network down. I also remember one day when the error rate went through the roof; turned out some guy in a lab down the hall left some cover off of his cyclotron. Comforting. One of the main things on the loop were the Glance-G graphics terminals, an early vector graphics display. That's what I was working on. If I remember correctly they had 74S181 bit slice ALUs and 1101 DRAMs. Lots of cool stuff was done on them including graphics editors, a display that showed the phone network routing around faults which, of course, it no longer does. The graphics that I did was used by Jim Kaiser for his digital filter work. I have some really vague memory of a Glance-G in the UNIX lab. Some other departments had their own systems. I eventually moved to area 20 to work on a 516-TSS based integrated circuit test system where we had things like wafer steppers and measurement equipment on the loop. The main high-level language on the system was called FSNAP, where the F was for floating point and SNAP was from the SNAP language. I remember adding pseudo-codes and statements to the language to support IC test. Other than that the system was nothing special. Only had a single-level file system. It was a great system for me a the time, but once we got our PDP-11 and UNIX I didn't look back. Oh, and we had a facility that let a modem on our system call the computer center to submit printer and card punch jobs and such. It had the ability to get called back if we submitted compute jobs. This may have been the predecessor to the UNIX gerts command. There's a special place in my heart for the computer center calling facility. I had finished up a large project and sent the whole thing up to the computer center to get it all printed and cards punched. I had sent so much stuff that I really annoyed the summer student who was working in the computer center. We spent a lot of time together after she calmed down. I tracked her down a few years ago and we're still friends. Jon