W3C home > Mailing lists > Public > www-rdf-interest@w3.org > September 2003

abbreviated RDF/XML was Re: Why does OWL have an XML presentation syntax?

From: Matt Halstead <matt.halstead@auckland.ac.nz>
Date: Fri, 05 Sep 2003 15:01:53 +1200
Message-ID: <3F57FCA1.1020603@auckland.ac.nz>
To: Jonathan Borden <jonathan@openhealth.org>
Cc: "Www-Rdf-Interest@W3. Org" <www-rdf-interest@w3.org>

I certainly wasn’t very clear in my first email, especially the title.

I followed the following thread and got some insight into some reasons 
for the decision

http://lists.w3.org/Archives/Public/www-rdf-logic/2003Jun/0046.html

I am surprised my initial searches didn’t turn this one up.

I think my problem comes down to understanding the value of abbreviated 
forms of RDF/XML in documents based on extensions of RDF, such as OWL. 
One problem is that there doesn’t seem to be a way to deconstruct an OWL 
RDF/XML document into triples and put it back together the way it was. 
While there is no loss in the interpretation of the RDF graph 
alternative forms have, an IDE could end up showing a different 
structure than was originally made. However, thinking about this some 
more, that perhaps is the role of another namespace and associated 
constructs.

A question I have about abbreviated forms of RDF/XML is that at some 
levels of abbreviation, one cannot represent constructs of the language, 
for example an instance of a single kind of class can use the class 
identifier as an element name instead of using rdf:Description and 
rdf:type, but if the instance is of a Boolean combination of classes, 
then you can only use the rdf:Description and rdf:type form. So my 
question in the last email was confusing, it was not about the 
presentation syntax, which would seem to give a way to express all 
constructs at the same level, but abbreviated forms of RDF/XML.

The thread I mentioned above w.r.t the presentation syntax highlighted 
the XML Schema problem of RDF/XML. This is part of what made me wonder 
about why we write down RDF/XML constructs as is done in for example : 
http://www.w3.org/TR/2003/WD-owl-ref-20030331/, you need to deconstruct 
them into RDF/XML descriptions and then into triples to process them. 
Example : i.e. why use abbreviated forms such as

<owl:Human rdf:about="#William_Jefferson_Clinton">

<owl:sameAs rdf:resource="#BillClinton"/>

</owl:Human>

instead of full RDF/XML descriptions when describing the syntax and 
semantics of a language.

The reason why I ended up thinking about this is because I am currently 
trying to put together an OWL ontology for a mathematical modelling 
language that has been developed by the group I work with. The structure 
of the XML is pretty concrete – i.e. a specific DTD or XML Schema can be 
used to validate the use of the syntax in a document. What is pretty 
obvious to me is that the Schema can be thought of as representing one 
possible abbreviated form of an RDF/XML description. So if I instead 
interpret this language as an RDF based language with its OWL Schema, 
then I can no longer say this XML Schema they use to validate their 
models is complete – it may be sound, but not complete. My next thought 
was that evolving the language within the confines of this one XML 
Schema may be quite limiting, what we really want to do is develop the 
language using an abstract syntax and semantics devoid of XML 
limitations, and perhaps also limitations of using abbreviated RDF/XML. 
The conclusion was, why use the current XML form at all – i.e. it’s 
simply an abbreviated RDF/XML representation. Hence my question about 
why use abbreviated RDF/XML representations of OWL to explain OWL or 
ontologies based on it.

regards
Matt


Jonathan Borden wrote:

>
> Matt Halstead wrote:
>
>>
>> I'm not actually sure what the value of OWL/XML syntax is, I am 
>> finding it hard to find a place where an OWL/XML document is useful 
>> for a machine. If it is meant to be useful to humans and not machines 
>> then I guess I'd rather look at KIF or abstract syntax.
>
>
> The _exchange syntax_ for OWL is RDF/XML. Several folks thought it 
> might be useful to have different _presentation syntaxes_ for OWL, one 
> of which is XML, another UML etc. You clearly don't find the XML 
> presentation syntax useful, which is fine. I suspect that the folks 
> who've developed the OWL XML presentation syntax do find it useful for 
> their purposes, which is also fine. If by KIF you meen an s-expression 
> based syntax, there is nothing preventing you from developing one for 
> OWL.
>
>> I'm always learning something new about the core intentions of XML, 
>> so someone pointing out why there is an OWL/XML syntax would help me.
>>
> XML is a specification, so it doesn't itself have intentions. I am 
> sure that among the millions of people who use XML there are a variety 
> of intentions regarding how it should be used. There is an OWL XML 
> syntax simply because some people thought it would be useful to have 
> one. The WebOnt WG felt that if there were going to be several 
> presenation syntaxes for OWL, that one of these should be XML. I think 
> it is useful if only as an example of how a specific XML syntax might 
> be more compact than an equivalent RDF/XML syntax. In the case of OWL, 
> the RDF/XML although more verbose, isn't all that more verbose. Just 
> my opinion.
>
> Jonathan
>
Received on Thursday, 4 September 2003 23:02:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:02 GMT