Re: Suggestion: HTTP Timezone header

On Thu, 23 Mar 1995 14:26:53 -0700, Paul Hoffman wrote:
>>A suggestion for HTTP: that clients send a Timezone header in the request,
>>describing the timezone of the user.
>On first glance, this seems useful. However, I'm not sure I see what kind
>of script would need to know the timezone of the user. Most time-dependant
>CGI scripts I can think of will always be talking about the time in the
>zone in which the server resides. Maybe you want to be able to reply with
>something like "At xx:yy (your time) you said...", but that seems like a
>stretch, particularly if the user doesn't set the timezone correctly. Toss
>in various daylight savings times and so on, and I think you introduce more
>uncertainty than you want.

The applications I have in mind are bulletin board or bug-tracking systems.
Individual files contain messages from users, stamped with dates/times in
either server time or GMT, which is no problem. However, I also provide an
'index' of recent messages sorted by date, so the users can check for any
updates since they last read the messages. Having that list presented in the
user's local timezone would make it much more intelligable; otherwise they
would have try to figure out whether they have already seen a posting
at '13:25 CET' when they know they last checked the system at 5:15 PST.

>It seems safer to base all times on the server's (hopefully correct) time.
I think you are assuming some penalty value for error which may not apply
in my applications. And I don't believe that _all_ time information would
suffer from an attempt at presenting it in the users timezone. I certainly
wasn't suggesting that use of the Timezone information be mandatory.

 David Robinson.

Received on Friday, 24 March 1995 02:55:00 UTC