Re: small comments on the RDFa Primer

Hi Ivan,

Happy New Year to you, too!

> 1. I see that you have put in the role of the 'class' attribute leading
> to 'rdf:type'.
> 
> If I look at chapters 2.5 and 2.6, in the example at the end of 2.5 you
> have two very similar constructions with class "val:Vevent" and
> class="contactinfo" but, according to the generated RDf tripls in 2.6,
> the first one leads to a <#xtech_conference_talk> rdf:type cal:Vevent
> whereas the second one does not yield any similar constraction. Is this
> a bug or a feature? If it is a feature, I would have hard time
> explaining it to an unitiated...

It is a bug. It should still generate a triple, although with a local
scope so that it won't lead to any unintended deductions.

In this case, it should be something like:

<http://example.org/staff/jo> rdf:type <_contactinfo>

I'm not sure if we've agreed what the local prefix would be... Mark,
you've thought about this, can you jump in?

I'll wait for Mark's comments to fix this in the draft.

> 2. My comment on XMLLiteral is still pending[1]. I also see that there
> *is* a 'type' attribute that one can use to set the datatype (as an
> aside, I am not sure I like the word 'type' for this, it is misleading
> for an RDF person:-(. I did not realize that in my previous comment. In
> view of this and my arguments described in [1], let me rephrase my
> proposal: the default datatype should be 'xsd:string' (which would also
> mean stripping all the xml elements from the results if necessary); if
> one really wants xml, than setting the rdfs:XMLLiteral datatype
> explicitly is an option.

There was a bug here, it is supposed to be "datatype" according to our
syntax. I've updated this.

I am inclined to agree with you on the default datatype: it should just
be a string, except if you really want some XML. What do others think?

-Ben

Received on Thursday, 11 January 2007 23:16:07 UTC