Re: more musings on RDF vCard and iCalendar

Harry,

Harry Halpin wrote:
> Garret - any updates?
>   

I'm attaching version 2007-04-29 of my process document, which switches 
back to the W3C namespaces and adds RDF equivalents for value types and 
structured values. Things to note:

    * The more controversial structured value types are now included,
      including vcard:N and vcard:Adr. The controversial
      "value-flipping" is included---hopefully there will be enough
      debate to allow us to converge on a good decision in this area.
    * Certain iCalendar value types such as PERIOD and RECUR are modeled
      as structured values (and therefore as RDF classes).
    * The lexical form of UTC-OFFSET is inconsistent between vCard and
      iCalendar; the vCard form, which closely follows xsd:dateTime, was
      chosen.
    * Certain decisions are still TBD, such as the proper ontology into
      which to place directory:geo.


> I'd be willing facilitate a #swig discussion if people can pick a time
> where we can discuss all the different concepts.
>   

That would be great, but I think for such a discussion to be most 
productive we should first at least lay out the issues to be discussed 
on this list.

> Again, I think the goals should be maximum simplicity with the ability
> to, if needed, round-trip.

I'm not sure I agree with "maximum" as a modifier, as that would 
probably entail simply taking some vCard values verbatim. (e.g. I don't 
want <vcard:n>Stevenson;John;Philip,Paul;Dr.;Jr.,M.D.,A.C.P.</vcard:n>, 
even though one could argue that this is the "maximum simplicity with 
the ability to ... round-trip.")

> I also think we should involve the W3C
> Internationalization effort in order to make sure whatever new VCard RDF
> comes is properly international.
>   

I definitely agree, although I wouldn't want to slow down this effort 
too much with such inter-group involvement. My reasoning here is that 
we're not creating an entirely new domain model---we're only creating a 
new mapping to existing data. I'm afraid if we waited for i18n support 
for things like iCalendar RECUR, for example, that this effort would be 
delayed for a very long time. I'd prefer first to create a canonical 
mapping for the existing vCard/iCalendar model, and then add appropriate 
i18n support (e.g. lunar calendars).

Garret

Received on Sunday, 29 April 2007 20:08:52 UTC