Re: ISSUE-55: Re-enable @profile in HTML5 (draft 1)

Maciej Stachowiak writes:

> On Oct 12, 2009, at 1:13 AM, Jonas Sicking wrote:
> 
> > I was thinking about the per-subtree-version-indicator when I wrote
> > my post. While I agree that it makes it easier for content producers
> > that produce a snippet of markup, it adds significant complexity to
> > the language. Just look at the amount of complexity xml:base or
> > xmlns produce.
> 
> I agree, it can be complicated to process a document with mixed
> versions, if versioning triggers different processing. But if you have
> global versioning only, then you can't safely use the versioned
> feature syndication feeds or mashups.

If entirely different attribute names are used for an incompatible
second version then you avoid complexity of mark-up's meaning changing
by wrapped by an element with a particular indicator.

Effectively that would make RDFa2 be a different language from RDFa1,
just as each of them are different from microdata.  There will be no
conflict, because user agents won't have to know they are in any way
linked.

> Overall, I think this is an argument against versioning as a trigger
> for different processing behavior.

Quite.

Smylers

Received on Monday, 12 October 2009 11:09:23 UTC