Re: Overview and Abstract Data Model - new document

Dave,

I have some agreement, some not, but overall I think we're converging...

At 11:37 AM 7/30/02 +0100, Dave Beckett wrote:
> >
> > >2.2
> > >   [W3C] XML schema datatypes were never in the original RDF design,
> > >   need to clarify somehow.
> > >
> > >   Cite something about W3C XML Schemas.
> >
> > I'm not sure what can usefully be said here, other than what is already
> > said.  I think the fact that use of XML schema datatypes was not detailed
> > in the original RDF design is not really relevant here.
>
>
>   [[The design of RDF is intended to meet the following goals:
>
>       * A simple data model
>       * Formal semantics and well-founded inference
>       * Extensible URI-based vocabulary
>       * XML-based syntax
>       * Use XML schema datatypes
>       * Anyone can say anything about anything
>       * A basis for legally binding agrements
>       * Universal expression of ground facts
>   ]]
>   -- 
> http://lists.w3.org/Archives/Public/www-archive/2002Jul/att-0056/01-Overview.htm
>
>
>No, RDF was not designed to use XML schema datatypes.  This is
>incorrect.  We are revising RDF to provide a datatyping mechanism
>that is compatible with using XML schema datatypes but it does not
>mandate their use.  Better:
>
>   Provide a datatyping mechanism that incorporates W3C XML Schema Datatypes
>
>(or some contraction)

OK, I'll work that in.

> > >2.2.4
> > >   cite the normative rdf syntax WD here not stripedrdf
> >
> > I looked into this, and come to the view that the normative syntax draft
> > doesn't really illustrate the point I'm trying to make.  Dan's document 
> does.
>
>
>   [[2.2.4 XML-based syntax
>
>    RDF has an XML-based serialization form which, if used
>    appropriately, allows a wide range of "ordinary" XML data to be
>    interpreted as RDF [STRIPEDRDF].
>   ]] -- 
> http://lists.w3.org/Archives/Public/www-archive/2002Jul/att-0056/01-Overview.htm
>
>The section is titled about the XML syntax.  Change the reference to
>refer to an RDF Core WD that describes the syntax - either the primer
>WD and/or the syntax WD.  The STRIPEDRDF document is useful for
>thinking about the syntax from a historical perspective but is not
>the document to point people at.
>
>And why "ordinary" and "appropriately" ?  This paragraph is a mess.
>Please rewrite it to describe that there is an XML syntax rather than
>concentrate on this corner case of using a profile of the syntax as
>"ordinary" XML (whatever that means).  The way we recommend to use
>RDF in XML at present is to use the syntax we define which is
>normative, now updated for best XML practice and should be linked here.

I agree that I should lead in with a broader mention of the XML syntax, and 
cite reference to the syntax specification.

But what you describe as a "corner case" is something that I have found to 
be crucial to getting agreement to incorporate RDF into commercial 
products.  As such, I think it's an important feature of RDF looking 
forward, regardless of the historical reasons.  (Even CC/PP implementers 
are now trying to wriggle out of using RDF because they cannot process 
CC/PP profiles using familiar XML based tools.)

I think there are two points of debate here:

1. we clearly disagree about the importance of this aspect of RDF, so I 
must appeal to wider WG input for a decision (either way).

2. Regarding the wording, I accept that it is inadequate, and am fully 
prepared to revisit that.  A possible approach might be to rework the 
section to:
(a) start with a general reference to the XML syntax and cite its 
specification,
(b) refer to the striped syntax section of the normative syntax draft, then
(c) to indicate by reference to DanBri's document an example that shows how 
XML can be constructed to be RDF compliant.

> >
> > >2.4.3
> > >   [XML-AS-RDF] - no reference, not sure what you want to cite here.
> >
> > I've cited a work-in-progress of my own, but I'd prefer something more
> > widely known.  RSS might be a candidate?  Can any RSS cognoscenti comment?
>
>Yes, replace with RSS 1.0 which is a well known and used RDF/XML
>syntax profile.

OK, I'll dig out that.

> > >   N-Triples rather than notation3?  Reverse setence to emphasis
> > >   the normative syntax :)
> >
> > In this case, I meant Notation3.  (Otherwise, I've reworked the 
> sentence as
> > you suggest.)
>
>
>   [[
>   2.4.3 XML serialization syntax
>   ]]
>
>Yet still does not mention the normative syntax despite citing your
>work [XML-AS-RDF] reference and actually mentioning that there is a
>normative syntax!
>
>Please add a reference to the syntax WD to this section.

Agreed - my oversight.


> > >2.5
> > >   It would be nice to see something about XML namespaces, RDF schema
> > >   document location and so on, how they have typically been used.
> > >
> > >   I noticed CC/PP says things about this, beyond what RDF core does.
> >
> > That feels to me more like detail material for the tutorial.
>
>Well, in that case the words in the CC/PP WD are beyond what RDF sanctions.

Quite possibly...

>see section 2.5 of this member only link to draft:
>   http://www.w3.org/Mobile/CCPP/Group/Drafts/WD-CCPP-struct-vocab-20020716/

OK, I'll stop resisting, but I'm still not clear what you want me to 
say.  There is no section 2.5 in that reference.  The nearest text I see is 
this:

[[
There is a reasonable expectation that a designated (globally unique) 
namespace will have associated semantics, including schema-related 
semantics. Thus, there is a convention that a namespace URI is associated 
with a corresponding schema document, though the specific mechanism for 
determining such an association is not formally defined. (The RDF Schema 
specification does say that the namespace identifier is also used as a 
schema identifier.)
]]

> > I'm not sure what I could usefully add here.
>
>Mention what you might conventionally find behind a namespace URI.

OK.  I can add something about that.

>Mention how an RDF schema document can be found for a property,
>possibly.

Apart from as above, I don't know how.

> > Currently, we *are* normative on these.  (A note has been added to 
> section 4.
> > 1)
>
>No.  Charmod is not a normative reference for the syntax work - we do
>not make such references on external working drafts and are not
>extending the normative references for the syntax.

This is currently subject of separate WG debate.   I propose we follow WG 
consensus.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>

Received on Tuesday, 30 July 2002 13:02:31 UTC