Re: New RDFa Primer Draft (calendar details)

On Mon, 2006-04-24 at 20:22 -0400, Ben Adida wrote:
> Dan,
> 
> Thanks for the comments.
> 
> >   <meta property="cal:dtstart" content="20060508T1000-0500">
> >
> > iCalendar doesn't allow 20060508T1000-0500 as a dtstart,
> > and neither does http://www.w3.org/2002/12/cal/ical .
> 
> [...]
> 
> >     <dtstart rdf:parseType='Resource'>
> >      <dateTime>2002-06-30T09:00:00</dateTime>
> >      <tzid>/softwarestudio.org/Olson_20011030_5/America/New_York</ 
> > tzid>
> >     </dtstart>
> 
> That's unfortunate, though I see why, in pure RDF land, this is a  
> good idea.
> 
> It might be interesting to think about simpler RDF properties for  
> these kinds of tools, RDF properties that mirror the MF properties  
> more closely. I'm thinking this is especially possible since the  
> vCard clients can already consume the literal we're thinking about.

I think you mean iCalendar clients. And no, they cannot handle
20060508T1000-0500. They only grok Z times, floating times,
or times with timezones where the all the timezone rule details
are copied into the .ics file.

> > The newer schema is http://www.w3.org/2002/12/cal/icaltzd
> >
> > It models timezones as datatypes, which is an idea that I'm not
> > sure is going to stick. But it looks like:
> >
> > <dtstart rdf:datatype="http://www.w3.org/2002/12/cal/tzd/America/ 
> > New_York#tz">2002-06-30T09:00:00</dtstart>
> 
> Yeah, that seems like an abuse of type inheritance....
> 
> > I see you cite the RDF calendar workspace; it might be better
> > to cite the IG Note:
> >
> > RDF Calendar - an application of the Resource Description Framework to
> > iCalendar Data
> > W3C Interest Group Note 29 September 2005
> > http://www.w3.org/TR/rdfcal/
> 
> Fixed the reference.
> 
> So, not sure how to proceed on this one. It seems like some simpler  
> RDF properties would be useful, especially if the existing tools can  
> already handle the literal we want. What do you think?

I think hCalendar+GRDDL works pretty well. I'm not optimistic about
encoding calendar info in XHTML using something more generic than
hCalendar.

I can imagine a bunch of RDF properties that mimic hCalendar
syntax, and maybe some rules to convert to the more general
purpose calendar properties. I've done something a little bit like
that when converting my PDA calendar to RDF and to .ics; it
didn't work out all that great; perhaps the main problem
with that particular approach was that RDF rules engines aren't
as mature as things like XSLT processors and python runtimes.


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Tuesday, 25 April 2006 01:53:53 UTC