Re: Presentation Syntax - why?

[Please note, this message is in reply to one from Danny Ayers that 
was posted to public-webont-comments -- his message is included below 
-- I have moved it here to avoid confusion with Last Call comment 
replies on the public-comments list]

At 11:47 AM +0200 6/17/03, Danny Ayers wrote:
>At times it seems that the real activity of the WG can only be seen like the
>structure of DNA was through X-Rays. A strange crystal that has just
>appeared on a plate is the OWL Web Ontology Language XML Presentation Syntax
>Note [1]. There doesn't appear to be any documentation of the role of
>languages like this, and it's hard to place it in context.
>
>Reading the documentation for OWL it is clear that the intention was for it
>to be a layer directly on top of RDF and RDFS, yet what the AS&S describes
>needs a significant translation from OWL->RDF [2] and is currently only
>mappable in the RDF->OWL direction through quite a convoluted procedure [3].
>
>The impression this gives is that AS&S has in large part been constructed as
>an entirely new language, with the RDF(S) considerations being retrofitted
>late in the day. How much truth there is in this isn't really important,
>what is important is that the roadmap has become smudged.
>
>I believe significant clarification is required around certain issues, in
>particular those that lead to the Presentation Syntax. It appears that this
>is a concrete representation of the AS&S, but for what purpose? If the
>underlying model used by the AS&S is compatible with the RDF graph/triples
>model, then why not use RDF/XML? Or is there such an air gap between the RDF
>and OWL layers, that the OWL can fly free with it's own model, syntax and
>serialization?
>
>On a practical level, the question is simple if a developer wishes to build
>a Semantic Web application, where do they start? If they start with RDF now,
>will they need a rework to be able to include OWL features without the
>struggle of [3]? Or if they start with OWL AS&S will they lose the
>compatibility with existing RDF data without building tools to carry out the
>transformations of [2]?
>
>I realise this is relatively early days for OWL, it just seems from what's
>being delivered that sometimes there are triples being asserted that are
>dark for anyone outside of the WG.
>
>Cheers,
>Danny.
>
>[1] http://www.w3.org/TR/2003/NOTE-owl-xmlsyntax-20030611/
>
>[2] http://www.w3.org/TR/2003/WD-owl-semantics-20030331/mapping.html#4.1
>
>[3] http://wonderweb.semanticweb.org/owl/parsing.shtml
>
>----
>
>http://dannyayers.com



Danny  - since this document isn't a Last Call document (and isn't 
currently on the path to becoming one), I will take the liberty of 
interacting without a Working Group mandate.  So these comments can 
be considered my own opinion, not those of the group.
  Frankly, I don't understand your point of view as raised in this 
message.  Most of the work I do in RDF is done using N3, because I 
find it easier to use.  My tools translate that N3 to RDF/XML, which 
is then used directly or turned into triples by some other tool. 
WOuld you argue against use of N3 or RDF/XML or Ntriples because they 
are all variants on using "pure" RDF triples (as are stored into the 
underlying RDF DB).
    OWL (Full) is a vocabulary extension to RDF (c.f [1] - my slides 
from the W3C track at WWW).  It can be written directly in RDF/XML, 
it can be written in N3, and now, thanks to the document you are 
complaining about, there is an XML presentation that more directly 
corresponds to the abstract syntax we use in proving the semantic 
properties of OWL (and particularly the OWL DL profile of OWL).  This 
presentation syntax comes with an XSLT that maps it into RDF/XML [2], 
and this is how we would expect it to interact with other OWL (and 
RDF Core and RDFS) tools. Thus,this syntax is just another way to 
produce OWL documents for people who have a different tool set.
    XML and RDF have different models, but many XML schemas can be 
mapped into perfectly reasonable RDF/XML, and this is one of them. 
It guarantees  that documents that validate against this schema, and 
go through this XSLT, end up as legal RDF documents consistent with 
the OWL DL profile.  This is why we stress its role as a presentation 
syntax - like N3 or Ntriples it is another way to look at RDF 
documents
   Other people are working on UML presentation syntax for OWL (the 
OMG has released a call to produce a two-way mapping between UML-2 
and OWL), a graphical presentation syntax, a prolog front-end.  All 
of these things lead to further adoption of OWL (and thus RDF) and I 
do not understand why you think they could be bad things.

  I also must state that I am personally upset at your charge that the 
activity of this working group is in any way mysterious.  Guus and 
Ihave worked very hard as chairs, spurred on by Dan Connolly as team 
contact, to make sure that everything the WG did was in public, and 
that every decision we made was open to anyone who wanted to track 
it.  The issue about whether to have an XML presentation has been in 
our publicly available issues list since Oct 2002 [4].  The WG 
decided to have this document as an "appendix" in Dec 02, but later 
decided a separate note made more sense.   Further, our entire mail 
archive is open to the public, and if you search on "XML presentation 
syntax" you will find close to 150 messages dating back to May 2002 
-- so you've had over a year to "study the chrystalline structure of 
our DNA -- we've made it damned easy to do!

   -Jim Hendler
p.s. I might also suggest you read the article Bijan Parsia and I 
wrote in XML Journal [5], perhaps it can help clarify my position on 
why there is no contradiction to having an XML presentation syntax 
for an RDF vocabulary.



[1] http://www.w3.org/2003/Talks/0522-webont-hendler/
[2] http://www.w3.org/TR/2003/NOTE-owl-xmlsyntax-20030611/owlxml2rdf.xsl
[3]
http://www.w3.org/2001/sw/WebOnt/webont-issues.html#I5.17-XML-presentation-syntax
[4]
http://www.w3.org/2001/sw/WebOnt/webont-issues.html#I5.17-XML-presentation-syntax
[5]
http://www.mindswap.org/papers/XML-J-Oct2002.pdf




-- 
Professor James Hendler				  hendler@cs.umd.edu
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  *** 240-277-3388 (Cell)
http://www.cs.umd.edu/users/hendler      *** NOTE CHANGED CELL NUMBER ***