Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site ucbvax.UUCP
Path: utzoo!linus!decvax!tektronix!ucbcad!ucbvax!daemon
From: daemon@ucbvax.UUCP
Newsgroups: fa.info-vax
Subject: Re: VMS Time Representation and Time Zones
Message-ID: <821@ucbvax.UUCP>
Date: Wed, 21-Sep-83 02:36:37 EDT
Article-I.D.: ucbvax.821
Posted: Wed Sep 21 02:36:37 1983
Date-Received: Thu, 15-Sep-83 13:52:46 EDT
Sender: daemon@ucbvax.UUCP
Organization: U. C. Berkeley Computer Science
Lines: 29

From GEOFF5@SRI-CSL  Wed Sep 14 23:36:04 1983
Your justification for using local time instead of GMT would seem
to justify precisely the opposite.  If I've got a distributed system
that spans a time zone boundary, I want all nodes to have a common
mapping from time to "time representation"; I don't want to have to
look at every message that crosses the boundary and modify the time
rep.

Local time and GMT are, in one sense, equally valid views of the same
thing.  Neither is "time"; both are "views of time".  When there is
no "correct" reference point, one picks an arbitrary one (which, by
world-wide agreement, is GMT/UTC).

Most computer uses of "time" are satisfied by an arbitrary, but absolute
and universal scale, or by elapsed time. When local time is need, one
can use a conversion function to map from a "time" value and a time zone
(and a "DST is observed" bit) to provide local time.  You'd use this
only at the highest (user-friendly) levels, say to print a directory
listing or to know when to turn on the porch light.  There's no need
to store representations of local time in internal format, except for
the infrequent case when a command such as "turn on the porch light
at 7pm" is compiled; even there, late binding is really the right way
to go.  There's every need to store a representation of universal time,
unless your computer system isn't ever going to be used on a network.

The problem of getting loosely connected computers to agree on what
time it is is another one entirely; there has been some careful theoretical
work done on it in the past few years.
-------