Timezone

Tim Daneliuk tundra at tundraware.com
Mon Jul 30 14:00:02 EDT 2001


"John W. Baxter" wrote:
> 
> In article <3B64B34C.6D8BEBE6 at tundraware.com>, Tim Daneliuk
> <tundra at tundraware.com> wrote:
> 
> > Actually, this raises an interesting question, which perhaps the resident
> > geniuses
> > can 'splain to me.  Here in Chicago, we're currently 6 hourse or 21,600 sec
> > *behind*
> > UTC.  Why does time.timezone return 21600, not -21600?  It seems that for
> > arithmetic reasons, you'd want to know both the magnitude and direction of
> > this vector...
> >
> > Inquiring minds wanna know ;)
> 
> Actually, you're currently at 5 hours from GMT...you're 6 during the
> winter.  time.timezone appears to be returning the zone without regard
> to silly adjustments like daylight time.  As per the library reference:
> "The offset of the local (non-DST) timezone, in seconds west of UTC
> (i.e. negative in most of Western Europe, positive in the US, zero in
> the UK)."  I think you need to look at the "dst flag" in
> time.localtime()'s result to learn the current status.  [It looks as if
> the C and Unix folks have forgotten about "British Double Summer Time"
> from WW-II, although Mac OS 9 and below record a DST delta rather than
> a flag, which could have handled that.]
> 
> If you were 6 hours the other way from GMT, you would get -21600.
> Basically, one needs a convention.  (Naturally, given two choices of
> convention here, the industry uses both in different places.)
> 
> You either get the number of seconds to add to GMT to give local time,
> or the number of seconds to add to local time to give GMT.
> 
> (There are also sub-conventions...it may be given in hours and minutes,
> instead, as it is in email headers...yours for that purpose is -0500
> currently.)
> 
> Is it any wonder that thing happen like the loss of an early moon
> unmanned visitor because the ground thought a negative yaw meant "go
> right" while the craft thought it meant "go left?"  [Or vice versa.]
> 
>   --John

Thanks John.  That's what's great about standards - if you don't like
one, invent your own ;)
-- 
------------------------------------------------------------------------------
Tim Daneliuk
tundra at tundraware.com



More information about the Python-list mailing list