Jon Knight wrote:
>Well I orginally had RFC822 style dates (with times and timezones) as the
>default but the ANSI one seemed more popular with other meta2ers. I must
>admit I'd prefer one with times in as well, and I'd also personally prefer
>one with a freely redistributable definition (which counts out the ISO
>standards).
Let's distinguish between "freely redistributable definition" and "freely
redistributable standard". The last time I checked the ANSI homepage, they
were charging for standards :-)
Now if it's a definition you want, Markus Kuhn has done a great write-up of
ISO 8601 - see <http://www.ft.uni-erlangen.de/~mskuhn/iso-time.html>. His
page contains pointers to other write-ups, some quite entertaining.
One challenge of ISO 8601 is that it allows so many options. For instance,
the following are all legal (note that the time is optional):
19970220
1997-02-20
19970220T203512
1997-02-20T20:35:12
The time specification handles time zones and Coordinated Universal Time
(UTC).
>But we've also got to make sure that the default is one that
>people will be able to generate correctly, and the ANSI does seem pretty
>simple to generate.
ANSI.X3.30's "19970220" format is one of the formats allowed by ISO 8601.
>If having the ANSI badge is a problem, there's always
>ISO.31-1:1992 which has a similar format to the ANSI.X3.30-1985 standard.
>Those needing more precise dates can alwars use a qualifier to specify it
>of course.
>
>Ah standards, eh? One for every day of the week and then some to spare...
>:-)
The version I've heard is: "The good thing about standards is that there are
so many to choose from".
>Tatty bye,
>
>Jim'll
Misha
|