[Remind-Fans] Remind using 12-hour time?

[Paul M Foster]

On Thu, Apr 10, 2008 at 01:27:38PM -0500, Paul Pelzl wrote:

> On Thu, Apr 10, 2008 at 12:58 PM, David F. Skoll <dfs at roaringpenguin.com>
> wrote:
> > J.P. Tuttle wrote:
> >
> >  > Oh.  Support for input in 12-hour would definitely be a useful feature.
> >
> >  Erhmm... really?  How would you write it?
> >
> >  REM AT 1:00pm ...
> >
> >  I guess I could modify the parser to look for a suffix "a..." or
> "p..." after
> >  a time and adjust appropriately.  I'm used to 24h clock so it's never been
> >  an issue for me. :-)
> 
> I have a mildly negative feeling about this.  If Remind uses *only*
> 24-hour time, then there is never any ambiguity.  If 12-hour times are
> supported, then the user has to be extra careful to always encode the
> AT timespec properly.  You can't even reliably drop the "a..." suffix
> because 12-hour time is not 0-indexed (e.g. 12:30am -> 00:30).
> 
> I'm a programmer, and I realize that this sometimes gives me a twisted
> worldview, but frankly 12-hour time sucks.  And I say this as an
> American who uses 12-hour time in casual use.
> 

I'm a programmer too, and I *hate* working with dates, because they're
so non-orthogonal. (Twenty-eight days in February *unless* ....) Twelve
hour time would be a similar (though not as severe) issue. Twenty-four
hour time is unambiguous under any circumstance. But more importantly...

Beware of feature creep.

Someone else mentioned the simplicity of *nix software interfaces. And
remind lends itself to having its output piped and manipulated. There
has to be some point where you, as a programmer, simply say, "Enough."
And for those left out in the cold because their feature is not
supported, pipe and manipulate.

Paul

-- 
Paul M. Foster