ietf-nntp Section 11.5 - NEWNEWS

[Andrew Gierth]

>>>>> "Charles" == Charles Lindsey <chl at clw.cs.man.ac.uk> writes:

 Charles> I don't think we need to care if the server is out by
 Charles> several minutes even, since it should not concern us. That
 Charles> is more than enough to work out what the server's time zone
 Charles> is.

There is no way to work out what the server's time zone is, and none
is needed.

(me, I'm not convinced that the timezone support in my server is
actually working, and I'm sorely tempted to turn it off entirely and
have the server pretend it's always in GMT. Given that my users are
scattered all over the world, any choice of timezone for the server is
arbitrary at best. A case could probably be made for recommending this
approach in the standard.)

 Charles> The only command that cares about time is NEWNEWS. If we ask
 Charles> for all articles received on the server since 001116 031219,
 Charles> then what we actually want is "all articles received since
 Charles> the last time I asked you, which was at 001116 031219
 Charles> according to my reckoning". The problem is that the time
 Charles> quoted in the NEWNEWS came from the clock on MY machine,

it shouldn't - a sensible client uses DATE to get the server (GMT)
time, then passes to NEWNEWS the _server's_ time that it recorded the
last time it did NEWNEWS, possibly even subtracting a fudge factor.

However, it is desirable for the server's clock to be accurate enough
for the other approach to work too.

But all bets are off if you don't specify the GMT flag to NEWNEWS (or
NEWGROUPS) - these days, it's dangerous to assume that the server is
in the same timezone, because it often isn't. (for example, the UK
provider madasafish.com use corp.supernews.com as their news server,
which is in the US Pacific timezone.)

-- 
Andrew.