On Mon, 16 Aug 2010 00:08:06 +0200
Alexander Teinum <atei...@gmail.com> wrote:

> > For timestamps that must be both human-readable and machine-readable, I
> > just told you: YYYY-MM-DDThh:mm:ss-tz:tz (the fractional-second
> > timezone should be optional).  (That paragraph wasn't entirely a
> > joke.)
> 
> Sorry, I misinterpreted what you wrote (“last standard that we’ll ever
> need” – the meaning of that and the rest of that paragraph did
> actually get swapped around in my head. ;)

“That paragraph” was referring to my second paragraph, and that really
is my favorite (to use, read, etc.) of the ISO 8601 formats.  I'm not
such a fan of the formats involving week-of-year, day-of-week,
week-of-month, fortnight-of-Mercury-year, etc..  (OK,
fortnight-of-Mercury-year is supposed to be a joke.)

> I did some research about different formats when working on some other
> calendar project this spring, and I arrived at the conclusion that all
> should switch to ISO 8601.

s/ISO 8601/a big-endian date format/

I do believe they put *every* big-endian date format in there.

>                            But I feel that it’s more sane than not to
> drop the number of seconds from flo’s output, since it will always be
> 00.

Definitely.  If you know you won't need a component of the format, drop
it before someone starts using it and relying on its presence.


Robert Ransom

Attachment: signature.asc
Description: PGP signature

Reply via email to