A plea for peace. was: RE: DAML+OIL (March 2001) released: a correction

Jim Hendler wrote:

> >RDF needs to be integrated with the XML datatypes.  Shall I take it
> >that DAML+OIL isn't going to help out on this?
> >
>
> Jonas asks a good question - let me try to help out as best I can
>
> DAML+OIL did not have authority to change anything in RDF or to
> otherwise impact the RDF spec except by example.  We decided to take
> on the datatyping issue because it was important to our users.

	DAML+OIL is in an interesting position due to the nature of its editorial
board, and I imagine is in a strong position to impact RDF - hopefully in a
good way. Let me say that I believe the attitude of trying to pitch in and
arrive at solutions to problems is a highly constructive one.

	It has been distressing to me to see some of the DAML+OIL group bluntly
proclaim that RDF is ultimately a bad way to do xxx or yyy, with the strong
suggestion that the only reason they are working with RDF is as if someone
is actually holding a gun to their heads.

	I think everyone agrees that there are some problems in various areas of
RDF. However I think that statements such as:

"I strongly agree with Pat Hayes's characterization of RDF (and RDFS) as
based on a fundamentally flawed semantic model."
http://lists.w3.org/Archives/Public/www-rdf-logic/2001Apr/0000.html

are not supported by practical suggestions as to what might be fixed. Such
broad statements are not helpful, rather are merely inflamatory. Indeed I
could easily state that _no model anywhere_ I've ever seen is anything other
than semantically flawed. Such discussions have no end, occur frequently,
and have little impact. To make such a characterization simply negates the
usefulness of "semantic" as a term of discourse. Perhaps we should simply
agree not to use this word, nor the term equally ambiguous term "knowledge"
and define our shared goals in a more practical fashion e.g. "design an
agent that has the following characteristics ... and is able to successfully
complete the following tasks ..."

> Unlike Pat and
> some others, I do think RDF compatibility is a crucial aspect of the
> DAML languages, and to this end I have agreed to work with the W3C to
> help in whatever way I can to coordinate ontology efforts with other
> RDF efforts.

I think that the essential characteristic is perhaps not RDF compatibility
(as RDF exists today) rather integration with the distributed Web
architecture and that means that URIs are the heart of the real issue. I
think that whether the RDF model is triples or quads or foo is a side issue,
the real issues that need to be solved are:

1) imprecise perceptions about what URIs and resources are to begin with.
RFC 2396 is broad in its definitions and we might take this opportunity to
introduce more structure into the definitions of URI, URI reference,
resource, entity etc.

2) integration of RDF with other web standards in particular XML. Alot has
changed in the XML world since the RDF recommendation came out. XML has
gained wide adoption across industry, academia and plain ol' web hackers
(this last one perhaps to a less wide extent). If DAML+OIL is struggling
with whether to maintain RDF compatibility or not, the issues regarding
integration with XML and specifications such as XML Schema would appear
hopeless.

> If a working group is formed to attack the remaining
> issues of RDF (which might include the datatypes) I suspect that I
> and others on the US/EU committee would be happy to join and work on
> finding a clean solution.

Working together is in everyone's best interest. A strong core RDF standard
is good for RDF and for everyone who uses RDF.

>
> That said, as long as we maintain syntactic compatibility, it's not
> clear to me that the higher levels of the diagram alluded to above
> can never have their own way of doing things.

Exactly!

> If the RDF layer can
> find a good datatype solution, then DAML+OIL users are free to use
> it, and/or to build maps to it (or, in the best case, to have the
> languages coevolve to the same standard).  However, if one needs to
> express shoe sizes in an ontology, one needs some way to do it, and
> currently we couldn't use RDF's mechanisms directly - so we went
> ahead and created a way to use XML datatypes.

A noble effort. The problems with doing this are deeper than they appear at
first glance. This is something that absolutely needs to be solved for many
reasons. The XML community itself is demanding a coherent type system. IMHO
XML Schema needs a little tweaking in this area, and RDF may have it right
after all.

>
> So, Jonas, you shouldn't take it that " DAML+OIL isn't going to help
> out on this" but rather should ask the question "Will the RDF
> community join us in finding a clean solution we can all use?" -- and
> (at least some of) the DAML+OIL folks are happy to help out.
>

I think that you've correctly framed the DAML+OIL/RDF problem. I would like
the RDF world to see integration with XML in the same light. The only way
this is all going to work is if it works on the Web, the non Web experiment
has already been tried.

Jonathan Borden
Director
The Open Healthcare Group
http://www.openhealth.org

Co-chair ASTM E31.25 XML Healthcare Standards
http://www.astm.org

Received on Monday, 2 April 2001 01:47:14 UTC