On 2001-08-20 18:13:08 kfogel wrote:
>"C. Scott Ananian" <cananian@lesser-magoo.lcs.mit.edu> writes:
>> Can I find support on the list for the idea of making *locale* the
>> dominant means of deciphering potentially ambiguous date expressions?
Here's a vote against that. See below.
>I'd also love to hear some others' opinions on this, especially non-US
>folks who (presumably) are most often bitten by unreasonable defaults
>or un-useful locale assumptions.
Thanks for asking :-)
Strong opinions follow; no offence is intended to anyone.
Dates like 3/4/01 are ambiguous, I hope we all agree. Now this
might not seem to matter, because it can be resolved on any given
occasion by things like the locale. But dates don't just get used
once. As soon as you have more than one person working on a project,
you have scope for confusion. And what happens if someone puts a
date like that into a script? The meaning of the script will change
depending on who runs it, and where! Now maybe that's "human error"
- but it's an error just waiting to happen.
>I personally always use DD-MM-YY or DD-MM-YYYY, even though my locale
>is U.S., because I think it's more logical.
A near miss...
> So how should Subversion
>behave to treat me right? :-)
Use the International Date Format, ISO 8601:
YYYY-MM-DD
First and foremost, it's unambiguous. All other advantages are
insignificant compared to this one, but here are some:
- it sorts correctly;
- it is consistent with time notation;
- it has constant length, so it can be formatted neatly.
There's more about it at http://www.saqqara.demon.co.uk/datefmt.htm
and doubtless many other places on the net.
Peter.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:36 2006