Re: Comments on draft spec (intro wordsmithing, ns rule bug)

On Sun, 2007-02-11 at 22:01 +0100, Danny Ayers wrote:
> re. http://www.w3.org/2004/01/rdxh/spec
> Revision: 1.206
> 
> Overall I think it's pretty close to what's required to go to Last
> Call, but there are quite a few not-major editorial changes I'd
> suggest (below).


Thanks for your detailed review; this 1st reply just covers
your comments up to section 3...

I applied most of your edits verbatim, resulting in
  1.207  2007/02/12 18:19:49
Exceptions are noted below...


> *** Table of Contents
> ...
> 4. The GRDDL profile for XHTML
> 5. GRDDL for HTML Profiles
> ...
> - could those headings be made a bit more explicit (to sound less
> similar)? Seems potentially confusing.

Yes, that was a case where the TOC was out of sync with the body.
The 4 main sections are now:

  <li><a href="#grddl-xml">Adding GRDDL to well-formed XML</a></li>
  <li><a href="#ns-bind">GRDDL for XML Namespaces</a></li>
  <li><a href="#grddl-xhtml">Using GRDDL with valid XHTML</a></li>
  <li><a href="#profile-bind">GRDDL for HTML Profiles</a></li>

> *** 1. Introduction: Data and Documents
> I think "dialect" needs a little pinning down somewhere, maybe here:
> [[
> There are many dialects of languages in practice among the many XML
> documents on the web.
> ]]
> =>
> [[
> There are many domain-specific languages ("dialects") used in practice
> among the many XML documents on the web.
> ]]

I made that edit, though the result is a little choppy.
I'm not inspired with anything better.

> (Dunno : are iTunes and Audioscrobbler different XML languages?
> Different dialects of the XML language? Different domain languages
> expressed in XML? The same domain language expressed in different
> dialects?)
...

> The RDF/XML example doesn't declare the dc namespace.

good catch.

> Also, might it be preferable to show this as Turtle/N3, to make it
> clearer that is isn't just some weird kind of canonical XML
> expression? (Or at least add another sentence to stress the point).

Hmm... what's wierd about it? It's totally ordinary RDF/XML, no?

I don't see sufficient reason to prefer turtle at this point in
the document.



> [[
> GRDDL stands for Gleaning Resource Descriptions from Dialects of
> Languages. That is, GRDDL provides a relatively inexpensive mechanism
> for bootstrapping RDF content from uniform XML dialects; shifting the
> burden from formulating RDF to creating transformation algorithms
> specifically for each dialect.
> ]]
> =>
> [[
> The publishers of the XML above could also provide the same data in
> RDF using RDF/XML or one of the other RDF syntaxes. GRDDL provides a
> relatively inexpensive mechanism for bootstrapping RDF content from
> uniform XML dialects; shifting the burden from formulating RDF to
> creating transformation algorithms specifically for each dialect.
> ]]

I made that edit, though it means the only place we expand
the GRDDL acronym is in the abstract. The document
is supposed to stand without the abstract (and vice versa).


> *** GRDDL Primer
> This should maybe moved to a "Related Documents" block, somewhere near
> the start.
> [[
> ...It develops on a number of examples...
> ]]
> strike "on"
> 
> *** GRDDL Use Cases
> This should maybe moved to a "related documents" block, somewhere near
> the start.
> *** GRDDL Specification
> - redundant?

OK, I changed that to "Preface and Companion Documents" and
subordinated the use cases and primer under it.

> [[
> The use cases document[usecases] collects a number of use cases
> together with their goals and requirements for GRDDL. These use cases
> also illustrate how XML and XHTML documents can be decorated with
> microformat, Embedded RDF or RDFa statements to support GRDDL
> transformations in charge of extracting valuable data that can then be
> used to automate a variety of tasks.
> ]]
> =>
> [[
> This document [usecases] collects a number of use cases together with
> their goals and requirements for GRDDL. It also illustrates how XML
> and XHTML documents can be decorated with microformat, Embedded RDF or
> RDFa statements to support GRDDL transformations to enable the
> extraction of valuable data that can then be used to automate a
> variety of tasks.
> ]]

No, I prefer "The use cases document...".


> !! Maybe add:
> [[
> *** GRDDL Recursion
> The mechanisms of GRDDL which will be described in the next section
> enable the interpretation of either an individual XML/HTML document,
> or a whole class of documents as RDF. In the latter case the
> association between individual documents and the appropriate
> transformations is done indirectly through linkage (dereference over
> HTTP). The documents specifying which transformations to apply may
> themselves need transformation from an XML/HTML dialect to RDF to make
> this information available. Hence in the general case recursion
> through multiple layers of linked documents may be required to
> complete a faithful RDF rendition of a source document. This recursion
> should be transparent to end-users of GRDDL-enabled systems, but is
> something implementers of GRDDL-enabled systems will need to consider.
> ]]

I haven't made that edit; at least not yet. I still wonder
whether the intro is a good place for it.


> *** 3. Using GRDDL with XML Namespace Documents
> 
> the green box on namespace transformations doesn't seem quite right -
> [[
> If
>     * an information resource NSDOC, identified by a URI NS,
> represented by an XML document with root node NODE with a GRDDL result
> that includes a triple whose
>           o subject is NSDOC, whose
>           o predicate is the property
> <http://www.w3.org/2003/g/data-view#namespaceTransformation>, and
> whose
>           o object is TX,
>     * and an information resource IR has an XML representation whose
> root element's namespace name is NS,
> 
> then TX is a GRDDL transformation of NODE.
> ]]
> Presumably this needs an "there exists" inserting,

The variables are all implicitly universally quantified, per pretty
normal usage.

>  but aside from that -
> second clause:
> "and an information resource IR has an XML representation whose root
> element's namespace name is NS"
> what has this to do with the right-hand side of the implication?

Eek... yes, that's garbled... I meant...

[[
If 
      * an information resource NSDOC, identified by a URI NS has a
        GRDDL result that includes a triple whose 
              * subject is NSDOC, whose
              * predicate is the property
                <http://www.w3.org/2003/g/data-view#namespaceTransformation>, and whose
              * object is TX,
      * and an information resource IR has an XML representation with
        root node NODE and with a root element associated with a
        namespace name NS,
then TX is a GRDDL transformation of NODE
]]

[more on the rest later; checkpointing here...]

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

Received on Monday, 12 February 2007 18:22:36 UTC