W3C home > Mailing lists > Public > www-rdf-rules@w3.org > June 2005

Re: SWRL

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Mon, 13 Jun 2005 10:44:13 -0400
Message-Id: <9C10BCDC-DC19-11D9-8498-0003939CCC42@isr.umd.edu>
Cc: <www-rdf-rules@w3.org>
To: <tim.glover@bt.com>

On Jun 13, 2005, at 7:59 AM, <tim.glover@bt.com> wrote:

> Hello,
>
> I am taking this opportunity to comment as a member of the public on 
> the SWRL proposal. I apologise if my comments are ill informed - if 
> so, please ignore them :)

We never ignore such comments :)

> My general comment is that this document in particular and the XML 
> community in general seem too pre occupied with syntax and parsing, 
> which are are non-issues.

I've asserted this in the past and generally agree with the spirit. 
However, 1) this is the world in which we live and 2) there are some 
advantages to XML derived formats (not realized by RDF/XML). The 
Grammar/type formalisms are language independent and fully (mostly) 
internationalized and far more compositional than traditional grammar 
formalisms (in that extensibility of the language is more or less built 
in and expected).

>  General purpose parsers for general purpose grammars have been around 
> for 40 years, and are freely available in all common languages (I use 
> CUP for java).
>
>  In particular, whats gained by the awkward abstract syntax?
[snip]

That's a different issue. It's because the semantics in this document 
are not complete, but rather an extension of that in the OWL Semantics 
and Abstract Syntax document. Abstract syntax was needed *there* 
because the two "concrete" syntaxes (RDF graphs...themselves abstract, 
and RDF/XML) were very...hmmm...tricky to define the kind of semantics 
of the direct model theory effectively.

Re: embedding strings in XML, this has been proposed for SPARQL (which 
is primarily defined in terms of a human readable syntax). The grammar 
is significantly complex, fwiw. Also, such embedding into XML is not an 
embedding at all. You can't schema check, transform, xpath or xquery 
query, data bind, etc. it in any useful ways. You cannot define 
profiles (via more restrictive schemas). Etc.

These are want the XML community is after at the moment, afaict. They 
don't seem all that nutty to me (having come from your position, FWIW).

I guess, to sum up, I think that the XML community *isn't* *just* 
obsessed about parsing (though the early days rhetoric sometimes seemed 
that way), but with *processing*. That is, having some degree of common 
processing models and methodologies.

Cheers,
Bijan.
Received on Monday, 13 June 2005 14:44:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:53:12 GMT