Precise Expansion Patterns (was: Re: [uf-discuss] Hcalendar in bbc.co.uk/programmes)

Fri Dec 14 16:21:41 PST 2007

On Sat, 2007-12-15 at 12:33 +1300, Paul Wilkins wrote:
> On Dec 15, 2007 12:18 PM, Martin McEvoy <martin at weborganics.co.uk> wrote:
> > On Sat, 2007-12-15 at 10:13 +1300, Paul Wilkins wrote:
> > > I think you're right. It should be P3M23S instead.
> >
> > I dont think thats its possible to drop the *T*
> 
> You're right, the T has to be there to resolve any ambiguity between
> month and minute.
> 
> PT3M23S is ISO 8601 notation for 3 minutes and 23 seconds.
> 
> > you could perhaps do PT3:23 ? which seems more accurate
> 
> Then it won't be able to be parsed as an actual time.

hmm are you sure? 3:23 is expressed as whole plus a decimal fraction

* http://en.wikipedia.org/wiki/ISO_8601

> 
> <snip>
> 
> > No I do not have enough wisdom to say what is right or wrong.
> > I do NOT however believe that machine data should be displayed in a
> > people area such as @title, I think machine data can be stored elsewhere
> > in a document such as in the <head> in a list of <link>'s or <meta>'s
> 
> The data format is machine data, as are geo-coordinates as well.
> The machine data has to remain with the human data so that
> discrepancies between the two aren't allowed to occur.
> 
> When thinking about other time formats, consider how 3 seconds would
> be marked up, or even days and weeks.

"PT1M" 1 min

"PT00:03M" 3 seconds decimal fraction

"PT3S" 3 seconds

"P0001-03-22T12:33:00" 1 year 3 months 22 days 12 hours 33 minutes and
zero seconds.

Thanks

Martin  

> 