Re: Syntax Document Editor's Draft update for review

Shane, Mark,

It looks better and better:-)

Here are my comments.

Section 4.2: Just a heads-up: *if* the discussion on URI-s lead to the
usage of a default namespace, then I presume that should become part of
the evaluation context. With an extension of the processing model in 4.3
to update, possibly, that URI, too

In the processing model:

Step 1, item 2. I am a little bit in the dark as for the usage of @lang.
AFAIK, the current DTD does not allow @lang and the validator reports an
error (at least last I tried...). It is fine with me either way
(although I have a slight preference for keeping @lang), but we should
be consistent with the DTD.

Step 2, item 1. To be ruthlessly precise, the sentence "Note that the
final value of the [current object resource] is an absolute URI, " is
incorrect, because it can also be a bnode!

Actually, the sentence "...unique identifier or [bnode] is created" may
be misleading. One could interpret it by saying that the implementatiom
may choose to use, eg, a uuid urn to create a real and unique URI,
instead of a bnode. Is this the intention?

If that was _not_ the intention then, well, the formulation is
misleading, it did misled me!:-)

Step 3, item 1. "Once the triple has been created, the [recurse] flag is
set to false." First of all, I think what should be said is "If any
triple has been created, the [recurse] flag is set to false." (there may
be more). However, the statement is not fully correct: if the literal is
set via a @content attribute, than the recurse flag must stay True!

Replacement text might be: "If any triple has been created and the
[current element] does not contain the content attribute, then the
[recurse] flag must be set to True".

Step 5. At this moment, this step seems to be a direct copy of what we
had before (setting the [chaining] flag at this point is meaningless:-).
I guess a note should be put here (just at the beginning of Section 5)
that the @instanceof question must be solved asap.  (In case the TC was
lost among all those emails, I think that [1] describes the different
alternatives and their effect on the processing model). The current
description is the same as alternative (1) in [1].

Step 6. Again to be precise: in the DOM terminology, afaik, everything
is a 'Node', including attributes. Ie, what you want to say here is
'element node'.

Section 5.1: you seem to have some formatting problem right before
Section 5.2; I presume some of the text there should not be in a <pre>...

Section 5.2.1.3: re exclusive canonicalization. I am not sure it
_should_ be exclusive canonicalization of the XML content. I realize
that using ex canonicalization is good if SPARQL queries are done
involving that literal, but it also puts a significant burden on
implementors because the usual XML tools (eg, minidom or PyXML in
Python) do _not_ have this facility. Generating an XML Literal means
simply copy what is in the original source verbatim. We should consider
restrict ourselves to that...

Section 5.2.2.4.1: I thought we would not include [_:xx] after all.
Having said that... It may be o.k. to use them if we do have the [..]
formalism... just a flag that this was a pending issue for a while...

Final comments:

- we do have now a profile for RDFa. I think this should be added to the
document
- do we plan to address the <ul>... issue in the next version? I am o.k.
with it, but we should not let it out of our sight for too long...

Ivan


[1]
http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Sep/0097.html

Shane McCarron wrote:
> 
> There are still some open issues, but Mark, Steven, and I have checked
> in all of our changes and I have put up a draft at
> http://www.w3.org/MarkUp/2007/ED-rdfa-syntax-20070912/
> 
> Please review this document in preparation for the next task force
> call.  We do not plan on checking in any more edits until after we have
> gotten comments and discussed them.
> 
> Thanks for your patience.
> 

-- 

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