Craig Latta <Craig.Latta at NetJam.ORG> wrote:
> 
> Hi--
> 
> 	The right thing, it seems to me, would be a group of servers which answered the current offset from UTC for a given latitude and longitude. That, in combination with current UTC time via the Network Time Protocol, would yield reasonably accurate local time. Tracking local UTC offsets seems like a perfect thing to delegate to a network service, instead of making lots of machines do it for themselves (which is likely to be complicated and error-prone).
> 
> 	Does anyone know of such a protocol and servers providing that service? The closest thing I can find is http://www.astro.com/atlas/atlquest-eng.html.
> 
> 


I still think the two queries should be merged into one somehow;
otherwise, you might query the time zone right before daylight time
takes effect, and then query the time after the shift, and end up with a
time which is off by an hour.  It's unusual, but it will probably happen
to someone at some point, and it will be a little embarassing when it
does.  It probably won't hurt anything, but still, let's try and do it
right if we can.

Also, many OS's already have reasonable understanding of timezones
(although their databases aren't always so great).  It seems a shame to
just ignore that information.  Also, Squeak in the future may well have
a good timezones library, in which case the appropriate thing to access
off the network is the database of timezones, not the current offset.


Lex