Re: GRDDL Spec Review (editor's draft Date: 2006/11/16 14:17:13)

On Fri, 2006-11-17 at 13:48 +0100, Fabien Gandon wrote:
> Below is an updated version of my comments for "Gleaning Resource 
> Descriptions from Dialects of Languages (GRDDL) editor's draft $Date: 
> 2006/11/16 14:17:13 $"

Thanks; the resulting change summary is:

http://www.w3.org/2004/01/rdxh/spec
Revision 1.164  2006/11/22 15:23:06  connolly
- "algorithms" in abstract
- strike "GRDDL ... separates content from its authoritative meaning"
- replace "adorn" by more plain/international English
- fix typo in xsd example code
- remove DOCTYPE from an example


Details:

> *Abstract :*
> "and for linking to an algorithm, typically represented in XSLT,"
> I would say "linking to one or more algorithms" since several
> transformations can be given.

I changed it to "linking to algorithms". "one or more" feels
overly wordy in this context.


> "(...) more recently known as microformats"
> Does not GRDDL intend to be broader than microformats?

Yes, but the sentence is about the GRDDL primer, which is, at
least currently, about GRDDL and microformats.


> *1. Introduction: Data and Documents *
> "There are dialects of XHTML, XML and RDF"
> I can understand dialects for XML and RDF but I find it surprising to
> talk of dialects of XHTML and to put them at the same level.

It seems OK to me.

> "RDFa and microformats offer simple (...)"
> Shouldn't we mention embedded RDF too?

I have since struck that sentence.

> "While this breadth of expression is quite liberating"
> This sounds like the expression of RDFa or microformats themselves; I
> would rather say something like:  "While the diversity of these dialects
> is quite liberating"

I think it's more clear with the RDFa/microformats sentence struck.


> *Resource Descriptions*
> First paragraph: I would include the triple itself in the text e.g.
> (#TheStand, #hasAuthor, #StephenKing)
> or
> (http://.../TheStand, http://.../hasAuthor, http://.../StephenKing)

It is in there:

   ... a triple whose subject is "The Stand," whose
   predicate is "has the author," and whose object is "Stephen King."


> The link with the second paragraph is not natural when reading the text
> linearly. Something is missing may be to mention that the RDF/XML cannot
> easily be inserted in other dialects and therefore that there is a need
> for an extraction of RDF from these dialects.

I have perhaps read it too many times; it seems OK to me.

> "The use of XSLT to generate XHTML (...)"
> I find this sentence useless in the context of this spec and it sounds
> "gratuitous" to me in the sense that nothing in this context is related
> or backing-up this claim. IMHO of non native speaker, its phrasing is
> not really in the style I would expect in a spec.
> 
> "separating structured content from its authoritative meaning (or
> semantics)"
> I don't follow this sentence:
> A - whereas the migration from just HTML to XML+XSLT+XHTML+CSS did shift
> the initial HTML representation to a separated representation (XML) +
> XSLT + (XHTML + CSS) the GRDDL proposal does not change the initial
> representation dialect;
> B - XHTML is itself a source for GRDDL and I find it confusing to say
> that GRDDL separates the content from its meaning while the
> representation remains XHTML.
> C - separating content from authoritative meaning sounds odd when one of
> the point I heard several times in the GRDDL group is that the GRDDL way
> to explicitly specify the transformation in the source actually ensures
> that the extracted RDF will have the meaning specified by the
> authoritative source of the structured content; in other words to me,
> this sentence makes GDDL sound like scrapping which I would personally
> like to be addressed in GRDDL but which I also now know not to be in the
> scope of GRDDL.

Yes, that does seems sorta backwards. I tried to reword it a bit
but ended up just striking a couple sentences, leaving:

<p>GRDDL is a mechanism for <b>G</b>leaning <b>R</b>esource
<b>D</b>escriptions from <b>D</b>ialects of <b>L</b>anguages.  That
is, GRDDL provides a relatively inexpensive mechanism for
bootstrapping RDF content from uniform XML dialects; shifting the
burden from formulating RDF to creating <a
href="#txforms">transformation algorithms</a> specifically for each
dialect. GRDDL works by associating transformations for an
individual document, either through direct inclusion of references or
indirectly through profile and namespace documents. Content authors
can nominate the transformations for producing RDF from their content
and use GRDDL to refer to them. </p>


> *2. Adding GRDDL to well-formed XML*
> "by adorning the root element with a"
> I am sure that it sounds wonderful for native speakers (and I like it 
> too) but I think the  sentence would be more internationalized if we 
> were to replace this expression by a platonic "by adding to the root 
> element a" ;-)

OK.


> *3. Using GRDDL with XML Namespace Documents*
>   *Using GRDDL with an XML Schema namespace document*
> In the example code I would add the rdf namespace.

It's there:

      <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">

> There is a "d" missing in the prefix of the last closing tag
> "</xsD:annotation>"

OK. good catch.

> 
> *Multiple transformations in XHTML*
> I would replace "decoding algorithm" by "GRDDL transformation"

it gets boring to repeat that so many times, but OK.


> NB: Some of the HTML examples have the <!DOCTYPE ... > clause while 
> others haven't IMHO we should choose once for all if we include it or 
> not; personally I would remove it as it clutter the example for no real 
> added value.

I removed it from the profileTransformation example, but the Joe Lambda
is in a section about the constraints of DTDs, so I left it in.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Wednesday, 22 November 2006 15:24:33 UTC