Re: Small @profile question

On Sep 10, 2010, at 12:17 , Nathan wrote:

> 
> Cool, and thanks for your replies, they've really helped me nail my concerns :) I think getting to the crux of it,
> 
> When parsing an RDFa Document, if a CURIE is unresolvable then
> (1) discard the would-be triple
> (2) emit a Warning and (1)
> (3) emit an Error and (1)
> 
> I guess if I was building a parser within a generic scraper then I'd instinctively go with (1), if I was building a parser in a library or a user agent implementation then I'd go with (2), and if I was building an RDFa validator (which checked the would-be RDF Graph) then I'd go with (3).
> 

The document already contains some reference to that[1], though the spec leaves the error/warning reporting strategy to implementations. Ie, (1) means that errors/warnings are not reported back to the users.

[1] http://www.w3.org/TR/2010/WD-rdfa-core-20100803/#processor-status


> As yet I'm personally undecided as to whether the benefits outweigh the potential problems, and have yet to fully consider Marks recent clarification on his profile related objection. At the same time though I'm very aware that it's late in the day, that I'm not bringing a better proposal to the table which addresses these potential problems and the ones @profile hopes to solve, thus unsure if I'm actually doing any good!
> 
> I can say though, that given the current status where @profile is in RDFa Core 1.1, then I'd be keen to see:
> - a note to implementers that stipulates a warning should be emitted if a profile cannot be retrieved or a curie cannot be resolved

As I said, this is already there

> - an !Important note to authors that makes them aware that if a profile cannot be retrieved then triples which rely on said profile will not be included in the graph produced by parsing the document.

Yes, that is something for the RDFa primer. Ben, are you listening?:-)

> And I'd hope some first-choice RDFa validator (probably on w3c) was implemented such that an error was produced if a curie cannot be resolved in an RDFa document.
> 

Well, at the moment, we do not have a _validator_. I have my own tool[2] which does something like that but it should probably be stricter if we wanted it to be a validator. You know, issues of time, free cycles and that sort of stuff:-(

Cheers

Ivan

[2] http://www.w3.org/2007/08/pyRdfa/Shadow.html


> Best,
> 
> Nathan


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