W3C home > Mailing lists > Public > public-grddl-comments@w3.org > April to June 2007

Comments on GRDDL W3C Working Draft 2 March 2007

From: Dave Beckett <dave@dajobe.org>
Date: Wed, 4 Apr 2007 14:13:09 -0700 (PDT)
To: public-grddl-comments@w3.org
Message-ID: <Pine.LNX.4.63.0704041412440.19753@xyzzy.dajobe.org>

Sorry this is past the 30 Mar deadline, that was a busy
deadline for myself too.  These are personal comments.


Gleaning Resource Descriptions from Dialects of Languages (GRDDL)
W3C Working Draft 2 March 2007


S1 no comment

    "grddl:transformation attribute whose value is an IRI reference,
    or list of IRI references"

That's a whitespace-separated list or space-separated?  The example in
S2 could be read as a newline and a lot of ws

I think I've only implemented spaces

Looking further, the Normative Statement says: "space-separated"
but the Mechanical Rule (Informative) uses
    (?Vnorm "[ \t\r\n]+") fn:tokenize [ list:member ?REF ].
which I can't really decode but seems to be whitespace, not a space.

So these are inconsistent and strictly, only a single space is allowed:
the normative part does not say "spaces", the plural, either.

Can you have leading and trailing whitespace(s)?

"To associate a GRDDL transformation with a whole dialect, have the namespace document include the grddl:namespaceTransformation property."

That phrase made me double-take.  You mean, have the namespace document
have a GRDDL result that includes a triple with the
grddl:namespaceTransformation property.

Here the whitespace-separation of terms is defined by HTML4 and
is rather vague since it says right now you can have 1 URI and
later many, but doesn't tell you how.

Your normative definition of 'metadata profile name' here is
space-separated again and the informative, whitespace.

S5 no comment

S6 no comment

The long example in this section could be marked informative.

S8 no comments

What is the sentence
    "The namespace document includes RDF data about the terms in the GRDDL
    Vocuabulary, but these RDF data do not include any triples whose predicate
    is grddl:profileTransformation."
for?  Is it saying that it won't define profile transforms.  What about
it defining namespace transforms?

[spelling: vocabulary]

Also see General comment below

S Appendix Transformations for Styling versus data extraction
suggest this is deleted, or made informative or explained
further.  It's not saying to ignore xml-stylesheet but
saying it isn't well supported.  It kind of overlaps
with the S8 security considerations which also wanders
into implementation advice.

S Issues
I assume will be removed from REC


I don't know where this would go, but can you add something like 
this, if it is true:
    http://www.w3.org/2003/g/data-view is used as an identifier
    The results of resolving this URI, or the grddl result of
    this URI are not required to be read by GRDDL agents.

If this document may change, maybe you should say so.  Usually
things like this (XML namespaces, schema documents) refered to
in RECs are very slowly changing, if at all.

If you want this in the form of a question:
    Are the contents you get when resolving the URI
    http://www.w3.org/2003/g/data-view part of the GRDDL recommendation?

The comments in the XHTML you get back tend to suggest not, there
are lots of todos and reference to things in progress.
Received on Wednesday, 4 April 2007 21:13:31 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:52:28 UTC