IIRC that was developed after 4.3 was released by Arthur David Olsen
(elsie!ado), and then incorporated into 4.3 via the patches that CSRG
sent out periodically.
You are probably right .. better memory. I knew it became widespread in a BSD stream, but I did not realize it was donated to CSRG.
Thanks.
But the key point is that the timezone DB development and inclusion in UNIX systems was much, much later in UNIX time and long after 1984 /usr/group standard, where the use of the TZ variable began to spread to make it easier for end users.
As Phil and I pointed out to Ken's original question, the V7-based systems compiled the TZ info (number of minutes west of GMT) into the kernel and supported a couple of primarily USA-based TZs in time(3) and the like. Which makes changing it for the local user a tad more complicated. Then again, you had the sources in those days, and at least the system administrator recompiled the core kernel from scratch. [I remember Joy once saying he thought rebuilding the entire system from the source at each site was a good thing because that way, binaries were not stale].
Anyway, the placing of the TZ string into a program's environment was pushed by the UNIX vendors, of course, because they were not releasing binaries.
Thus, by the time of the TZ=xSTdyDT convention, the time(3) family was a bit more flexible [i.e., most of Europe was easily supported] - but you still needed to know what to set it all to.
It was only much later, when the timezone DB code was created, that it became easy to set the timezone in a more worldwide scheme.