Re: DSThesia
UNIX had (and in most cases still has) a way of coping with timezones that while it is not as sophisticated as the Olsen tzdata system (which was actually implemented on UNIX before Linux was even around), is actually quite simple to implement, if not quite so automatic.
The TZ variable can contain the normal timzone name, the DST timezone name, the offset from GMT/UTC, the DST offset and the days of the year that DST comes in and goes out (both in terms of the Julian date of the switches, and the weeks within a month, the day of the week, and even allows you to specify exactly what time the switch happens).
It required the system administrators to know a little about their environment, time zone names, offset and when DST started and finished, but this worked well across the '90s and on into the current day. It actually also puts the representation of time for a user into their own hands, as TZ is an environment variable that users themselves can change.
I worked on SVR2 systems back in the mid eighties that had the Julian dates for the DST shift hard-coded into the C library, which was a pain, but it soon started to become a little more flexible in SVR3 and later, so by the time it got into '90s, the more sophisticated handling of the TZ variable had become normalised, and I saw it both on Digital UNIX, HPUX and several other UNIX systems of the time.
I first saw the Olsen system on Digital UNIX and HPUX, but this soon got adopted in other UNIX systems.