Re: Proposal, a new class of Web Names

Jonathan Rees wrote:
> Well well well, small world.  I was assigned to track this question
> for the TAG.  I was in correspondence with Ben and Manu and didn't
> know you were looking at this too.

aye, that was some time ago actually, following from me "having a go" at 
iand on the pedantic-web mailing list [1] - but obviously I'm now also a 
member of the RDFa WG and thus knew you'd talked to manu and also 
discussed the issue on via mails and on the telecon yesterday just after 
you spoke.

[1] http://bit.ly/fkb1nm

further ironically, it's possible that the chain of events, mails and 
discussions which arose after that mail consequently ended up in me 
being a member of this tf!

> http://www.w3.org/2001/tag/group/track/actions/509
> 
> I've linked my action to your writeup and to RDFa issue 84, so maybe
> we'll be able to knit this together.
> 
> Since 3986 is authoritative for fragid semantics I'd be sure to use
> that as a starting point in any spec-conscious writeup - I was
> surprised you don't even refer to it on that page. You could say that
> you choose to ignore what it says, since it's borderline nonsense, and
> that you use the (nonauthoritative) AWWW interpretation instead (which
> is also borderline nonsense, but nonsense that's more useful for RDF
> purposes). That's similar to what I asked Manu to do. Here inside W3C
> you've got to either follow the RFCs, change them, or own up publicly
> to the fact that you're at variance. I suspect AWWW is not quite as
> direct about this as it ought to be.

TBH, after discussing it in the telecon yesterday (this isn't an 
official rdfa wg response of course) - the general sentiment was that 
everybody had tried to stay as far away from these issues as possible, 
in the hope that people would just do things correctly themselves, and 
so that should AWWW or RDF change then we don't need to change RDFa - 
but also that if you (/the tag) really wanted us to drop a note in there 
to say "something" (like what you propose) on the subject, then we'll 
add a note to the spec (or cookbook was another suggestion).

> If you wanted to fix the problem, or at least part of it, note that
> the +xml and html media type registrations are both in the process of
> being revised. The new +xml makes things considerably worse for RDFa.

ack I could argue both ways on things given the current specs, because 
going by the words in the recs, an @id corresponds to an element found 
in a representation, and @about in RDFa is joined to another string to 
make an RDF URI Reference, such that

  GET /foo.html
  Host: example.org
  -> returns html containing..
  <div about="#bar" id="bar">

then the @id has the value of "bar" - a client side indirect reference 
specified as being in the scope of the representation.
and the @about has the value of <http://example.org/foo.html#bar> - a 
name, an RDF URI Reference.

It's only when you have "follow your nose" (which we don't define) that 
the two can be confused.. but that's confused by the current text of the 
recs, in reality they both /should/ refer to the same thing.. in the web 
names proposal they do.. they're both ( 'http://example.org/foo.html' , 
'bar' ) and thus must be consistent.

Best,

Nathan

Received on Tuesday, 15 February 2011 17:17:30 UTC