Re: vCard/iCalendar RDF process document 2007-04-06

Harry Halpin wrote:
> 1) Mapping hCard/Simple VCard use-cases. This is what Norm does.
>   

I receive vCards in my email all the time. I don't remember ever 
receiving an hCard. I had never heard about it until this discussion.


> 2) Mapping full VCard with round-tripping. i.e. Garrett Also, to express
> preferred names, multiple phone numbers of different types, etc., this
> requires much more extensive use of class names and various
>   

I have a lot of user data I need to store. I wanted a robust, 
interoperable way of doing that, so I chose RDF+vCard. I want to be able 
to store *all* my data, not just some of it. I could care less about 
hCard. (I'm sure it's a fine specification, it just doesn't influence my 
reasoning regarding storing vCard information.)

> The distinction seems to be:
>
> 1) Properties with a range of only literal values for vcard attributes.
> 2) Properties with a range of (multiple) resources distinguished by
> their use of classes.
>
> These aren't mutually exclusive. While I haven't sat down and hammered
> it out, one could just remove the range constraints for the vCard RDF
> Schema so one could use the same ontology to do both, as Garett
> mentioned.  But it makes processing a bit more tricky. And a third
> option would be to:
>
> 3) Attempt not to use classes and subclasses, but instead use
> sub-properties, and so one I think gets the best of both 1) and 2).
>   

That's a nightmare. You know above where you say that my proposal makes 
processing a "bit more tricky"? Well, this is many light years more 
tricky. Let's say I want to look for an email address. First I have to 
check for vcard:email. Then I have to check for vcard:preferredEmail. 
Then I have to check for vcard:workEmail. Then vard:preferredWorkEmail. 
How does this help make thing simple?

> However, I am also pro-keeping the spec for VCard separate from
> iCal/RDFCalendar.

I'm pro "separate" as in "componentized", not "separate" as in 
"independent".

> Garrett - I suggest you e-mail your Directory
> suggestion to the RDFCalendar task force [1].

Will do.

> As for a #swig discussion, I'm out of town (as is almost all W3C Staff
> such as Dan Connolly and others who I know are interested in this for
> WWW2007 next week, so I suggest 1:00 PM EST (which should map decently
> well to people in Europe) in #swig for a meeting.
>   

What date? Now that we've hashed over some of the issues here, I can 
create an agenda.

Garret

Received on Wednesday, 2 May 2007 16:18:02 UTC