I was a student, but working as a 360/TSS programmer and PDP-10/20 programmer in the CMU computer center. I also did some hacking for the EE department mostly for free for them – as EE was my major. My girlfriend at the time was working for an EE prof and she got me to bring UNIX over from CS (EE had a 11/20's and a prototype 11/34 [not a 34A]) and CS of course had numerous 11's as they were building C.mmp and writing Hydra. [CMU/CS actually ran UNIX on the “pre-Hydra” 11/40E; the E was a CMU/3 River's writeable microcode modification - very cool).]
Anyway, I had access to the only 11/45's on campus in the computer center where I worked – which was important because V5 and V6 really wanted to be booted on an 11/45. Dan Klein, who was math major, but like me, had a similar job to mine, working for the CS Dept to pay his bills. There was a lot of helping each other because the equipment we had was both common and complementary – so all the systems programmers tended to have “guest accounts” in both worlds. Since Dan and I were the common threads with access to all versions of the hardware were sort in the middle of moving this stuff around.
Note: CMU/CS had already done the V5 to V6 upgrade. EE never had V5, although any of us that had CS accounts has briefly used it there.
In the fall of 1976 (I believe), Ted Kowalski shows up for his OYOC year in the EE dept after his summer at Bell Labs, and graduating U Michigan (he was once Bill Joy's housemate at Michigan in a strange twist of fate). He brought with him some stuff, including a photocopy of what would become K&R and copy of phototypesetter C, and a copy of UNIX/TS. He also brought his cool (but unfinished) program he had started to write originally at U Mich – fsck. [Hacking on fsck under the tutorial guidance of tjk was my first real introduction into UNIX file system internals]
Ted brought back to AT&T the 11/34 support and eventually 11/34A and RK07 support that Dan and I would later do. He also spoke of the upcoming 7th edition. Ted wanted to get the 34 support added back into the main line at AT&T, because it was clear that was going to be system most people bought from then on. We keep asking when it was going to be released. While CMU was not the first V7 licensee, thanks to Ted, we were very early.
In 1977/78 Dan and I would start to work from CMIR – CMU's Industrial Research arm which had a new group start up a person who originally had an appointment in EE (Ron Krutz) - CMIR was the “Mellon Institute” part of Cargnie-Mellon Univ). They worked with industry and even took on profit stuff. Dan and I brought the Kowalski/Cole/Klein UNIX over from EE to be the basis and added to write a driver for the newest DEC disk (CMIR had serial #3 RK07 and a very early 34A). Ted was a little nervous about the license, as CMIR was doing a lot of stuff in the for profit part and the AT&T University Research License forbade that.
The folks at CMIR promised to fix that, but it seemed like promises for a long time – hence Dan and mine work slow down. My memory is that actual commercial license was purchase after we got V7, but that's purely memory. I'd need to look at the system we ran. Certainly by the time I left for Tek, 1979 V7 was running at CMU.
But the point is: Ted brought TS to CMU as a post V6 system. Some of the CMU changes stuff made it back to V7 via Ted.