Re: SHOULD vs MUST ignore @version

Not disagreeing, just adding...

My opposition last week was also motivated by the fact that having a requirement on using @version means a real load on parsers. In the meantime I looked at my own code and it is indeed relatively easy to factor out the 1.1 specific things. So it is not _such_ a load. In this sense a SHOULD on ignoring @version seems o.k. to me, even a MAY is fine. I think the only thing we want to avoid is a MUST on _using_ the @version...

Ivan



On Jul 14, 2010, at 14:05 , Toby Inkster wrote:

> Essentially what we want parsers to do is ignore @version -- i.e.
> always use the latest version of RDFa. 
> 
> Firstly, let's check that we've left enough space for future
> innovation. I think we have. If a future version of RDFa, say 2.0, is
> so incompatible with RDFa 1.x that version detection is needed, they
> can say that authors MUST include a particular profile attribute on
> their root element, say:
> 
> 	profile="urn:w3c:rdfa:2.0"
> 
> As it's deliberately unresolvable, an RDFa 1.1 processor will skip the
> root element. This effectively allows future RDFa Working Groups to
> re-establish versioning, so I think we have that covered.
> 
> So the versioning discussion only needs to apply to RDFa 1.0 and 1.1.
> 
> Now, let's assume that somebody *wants* to do version detection,
> despite us saying it's a bad idea. Their code might be something like
> this:
> 
> 	function rdfa11_processor ($doc) {
> 		if (check_version($doc)==1.0) {
> 			return rdfa10_parse($doc);
> 		} else {
> 			return rdfa11_parse($doc);
> 		}
> 	}
> 
> If we prohibit this and say it's not conforming, they just refactor
> their code:
> 
> 	function html_processor ($doc) {
> 		if (check_version($doc)==1.0) {
> 			return rdfa10_processor($doc);
> 		} else {
> 			return rdfa11_processor($doc);
> 		}
> 	}
> 
> Now they're still doing version detection, but their rdfa11_processor
> function is technically conforming, because the version detection
> occurs externally to it.
> 
> So my argument is that we make ignoring @version a SHOULD rather than a
> MUST, because if it's a MUST people can just work around it.
> 
> -- 
> Toby A Inkster
> <mailto:mail@tobyinkster.co.uk>
> <http://tobyinkster.co.uk>
> 
> 


----
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