[uf-discuss] Apple Data Detectors

Breton Slivka zen at zenpsycho.com
Tue Feb 5 21:42:07 PST 2008

I think what was intended, was rather than try to write a parser that
picks up most styles of natural language dates, as you suggest-
Instead write a parser that only picks up one or two standard styles
of dates. Much like the style guides that are used in academia for
writing standard forms of citations, and other things. Decide on a
freeform text format that you know a machine can pick up, and excludes
ambiguous date formats.

But you do raise a valid point here: Even if you do that, you invite
the assumption from authors, that since it can pick up this format of
date, or that format, then perhaps it will pick up THIS format as
well. So authors will write badly formed versions of this freeform
standard. There will be typos, too. All kinds of things can happen.

But much of these bad things can be aleviated by one of the other
suggestions in this thread: As-you-type validation.  As soon as you
type in "Feb" for instance, autocomplete style routines kick into
action, helping the author write the date in exactly the right format.
Then as they hit "publish" it becomes a microformat, proper, with
markup and all.

On Feb 6, 2008 3:58 PM, Michael MD <mdagn at spraci.com> wrote:
> > people write dates, addresses, etc on the Web or on their emails. Asking
> > people to write "Tuesday, February 5, 2008" in this order, with the
> > commas, etc. is very likely even simpler for normal people than writing
> you would *think* so - and it would certainly be nice .... but the behaviour
> or most people out in the real world does not suggest that this would be
> easy.
> Most freeform text dates I see out there are missing the year (how is a
> machine supposed to work out what year was intended?)...
> and a lot of them are in useless ambiguous formats like dd/mm/yyyy or
> mm/dd/yy - ) ... then there all those other variations... to many to list!
> I've experimented a bit with trying to parse freeform text dates ... the
> problem is as soon as its loose enough to pick up most of the common ways
> people write dates it then also starts to pick up a lot of other stuff as
> dates that were not intended to be dates at all!
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss

More information about the microformats-discuss mailing list