Fwd: In case you were not on IRC...

This is a last call comment on the RDFa Core document; not really mine, I just forward this after having tested Mike's preliminary version of the RDFa validator.

Begin forwarded message:

[skip]

> From: "Michael[tm] Smith" <mike@w3.org>
> Subject: Re: In case you were not on IRC...
> Date: February 2, 2012 17:33:44 GMT+01:00
> To: Ivan Herman <ivan@w3.org>
> 
>> [11:02]  <ivan> It also decides that a @property="" is an error
> 
> That's because the current RDFa Core 1.1 spec requires that.
> 
> That spec says that the value of the property attribute is:
> 
>  "a white space separated list of TERMorCURIEorAbsIRIs"
>  http://www.w3.org/2010/02/rdfa/sources/rdfa-core/Overview-src.html#s_syntax
> 
> The empty string is not a valid term or CURIE or absolute IRI.
> 
> If the intention is that the empty string should be allowed as the value of
> the property attribute, then the RDFa Core 1.1 spec needs to instead be
> updated to say that the allowed value for the property attribute is,
> "a white space separated list of *zero or more* TERMorCURIEorAbsIRIs".
> 
> Incidentally, that's true for the values of other attributes as well; the
> spec needs to explicitly say either "zero or more" or "one or more" for
> them. Lack of that language makes it ambiguous.
> 

We do have cases when we used attributes with an empty value, like @typeof or @datatype. These, and maybe @property="" may make sense in some corner cases, possibly changing the effect of chaining. Though they are not very good practice, I think we should not disallow them.

The proposal is to make a change on the definition of those datatypes to allow for zero or more of the XXX types.

Thanks

Ivan

----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf