In the interest of consensus and progress on datatyping...

Putting on my editor's hat, and taking off my proponent's hat,	
as a means of trying to facilitate progress towards consensus
and closure of the datatyping issues, I would like to propose
the following approach.

I have restructured the current datatyping specification into
two parts. The first part attempts to capture the core aspects
of datatyping which it appears the WG has reached agreement on
and which, for better or worse, could be adopted as-is as the
solution recommended by the WG for RDF Datatyping, even if not
addressing all desiderata identified thus far, leaving the
remainder to be resolved by future working groups and/or the
RDF community and marketplace.

The second part attempts to define extensions, refinements,
and optional machinery which may still be considered by the WG for
adoption and inclusion into the core proposal defined in part
1, but may also simply be left as issues which need further
investigation; the latter part of the document providing a
suggested roadmap for further research and discussion, rather
than simply remaining silent on such issues (i.e. capturing
at least what the WG was able to achieve thus far, even if
full consensus was not attained).

In a nutshell, Part 1 contains:

   * intro and general definitions of datatype classes

   * local/explicit typed non-XML literals

   * single typed literal nodes rather than bnode idiom

   * global datatyping assertions (constraints) of
     locally/explicitly typed literals and discussion
     of datatype clashes (NOT implicit global datatyping)

   * minimal model theory covering only the above

   * implications of adopting only Part 1 as-is (simply a
     heads-up to developers/users, no arguments against
     adoption of only Part 1 as-is)

   * use case examples based only on Part 1 as-is

Part 2 contains:

   * typed XML literals

   * typed literals as subjects

   * inline literal semantics (tidy versus untidy)

This last section in Part 2 takes each of the tidy and untidy 
approaches in turn, and defines how the core in Part 1 would
be augmented with regards to RDF/XML and graph representation
(if needed), extensions to the core MT, and implications and
considerations if adopted, including any changes/enhancements
to the use case examples presented in Part 1. Part 2 does not
(intentionally, at least) advocate either approach, but tries
to simply present them in terms of the core defined in Part 1
so that either the present WG, a future WG, researchers, or
developers can understand and weigh the two alternatives.

Note that none of the top level sections in Part 2 are co-dependent,
and any can be accepted or rejected for inclusion into Part 1
independently of any other.

--

It is my hope that this restructuring will speed progress towards
a positive closure on datatypes, without discarding or losing
any of the work that has been expended thus far by the WG on
the remaining unresolved issues.

The WG should, I hope, be able to vote on acceptance of Part 1
as soon as it has been suitably reviewed, and then can decide
if and how much of Part 2 to consider before closing datatypes
at this time.

When an official WD is produced, Part 2 will likely be relegated to
non-normative comments/recommendations for future work.

A static copy of the restructured specification can be retrieved from

http://lists.w3.org/Archives/Public/www-archive/2002Aug/0111.html

The online editorial version will be kept, for the time being, at

http://www-nrc.nokia.com/sw/rdf-datatyping.html

I consider the material in Part 1 to be sufficiently
complete for the WG to review.

The material in section C of Part 2 is not yet considered complete nor
ready for extensive review, but I will continue to work on it. I just wanted 
to bring the WG's attention to this approach and the restructured
document. 

One final note: I have made a distinction between the authorship of
this restructured specification and the official WG datatyping WD,
which this document is expected to become, due to the repeated comments
about whether the other WD editors have or have not fully reviewed its
content. I.e. I bear sole responsibility for all content in the above
mentioned document. If/when this document becomes a WG WD, the other
editors are welcome to share in the effort and responsibility of its
production. Hopefully this is satisfactory to all and will make clear
where all blame for inaccuracies, errors, or omissions is due.

Cheers,

Patrick


--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Thursday, 29 August 2002 05:55:38 UTC