TZ and TERM per process; V5N16

[Doug Gwyn <gwyn>]

[This discussion originated in mod.std.unix.]

> 					It is (so far) independent of
> astronomical or political changes in the "real" time.  We'll see how
> this works once we start talking in terms of relativistic distances
> [;-)].

I realize Joe was being facetious, but this is actually an important
problem in an optimistic future with networked computers located on
distant planets or space stations.  Leslie Lamport of SRI published
a CACM article a few years ago which addressed the issue of
synchronizing clocks in such a distributed network.  I think
Lamport's scheme works when relativistic non-simultaneity is a
significant factor.

Simpler schemes, such as designating one host (e.g. the Royal
Observatory) the official timekeeper, do not work in the general
case.