W3C home > Mailing lists > Public > wai-xtech@w3.org > April 2003

Re: [Fwd: [closed] chas-01 request to add aboutEachPrefix]

From: Al Gilman <asgilman@iamdigex.net>
Date: Wed, 09 Apr 2003 10:38:51 -0400
Message-Id: <5.1.0.14.2.20030409102607.0275a4e0@pop.iamdigex.net>
To: wai-xtech@w3.org

At 09:51 AM 2003-04-09, Charles McCathieNevile wrote:

>Hi Folks,
>
>This issue got rejected by the RDF core group.
>
>My feeling is that  aboutEachPrefix per se is indeed a nasty hack with 
>lots of problems. On the other hand there is a huge need to be able to 
>discuss a class of objects without enumerating every member of the class 
>in advance. For example, pages produced by an organisation, or pages 
>aproved by an individual for publication. (In fact there is no reason why 
>these would all begin with the same fragment of URI anyway).
>
>It may be that this is only possible by using OWL. This would mean that we 
>cannot expect to meet these use cases except with systems which have the 
>additional machinery, as far as I can tell. I would like to ask the grop 
>for an explicit clarification of whether this is the case, and if so I 
>will actually request something to this effect go in the primer (including 
>at the least a link to an explanation of how it can be done using the full 
>glory of the Framework)...

We have a more imminent and deeper need than that, I believe.

If I catch your drift

- it takes more than OWL, it takes the functions of the Logic group applied to
the declaratives of the OWL to do what one needs, and...

- we need this NOW for lexicons, in particular a Glossary entry is already
something of the pattern-class that you are discussing, and that RDF does not
support natively.  It binds some further information to a set of instances;
where what we have by way of definition for the set is an 
instance-recognition pattern, not an enumeration of identities of the 
affected instances.

Al

>cheers
>
>Chaals
>
>-------- Original Message --------
>Subject:        [closed] chas-01 request to add aboutEachPrefix
>Date:   Mon, 07 Apr 2003 16:44:13 +0100
>From:   Dave Beckett <dave.beckett@bristol.ac.uk>
>To:     Charles McCathieNevile <charles@sidar.org>
>CC:     www-rdf-comments@w3.org
>
>
>
>Charles,
>
>You made a last call comment captured in
>
>  http://www.w3.org/2001/sw/RDFCore/20030123-issues/#chas-01
>
>The RDF Core WG has resolved
>
>  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Apr/0128.html
>
>to reject this comment on the grounds that the original reasons for
>removal remain valid:
>
>* We had good feedback that it lived badly in RDF and as such was
>   little implemented or used.
>
>* It broke layering by looking inside URIs rather than dealing with
>   them as identifiers.
>
>* Felt more like a RDF/XML syntax thing than part of the RDF model
>   and as such it mixed badly with triples.
>
>* Worked badly with rdf:bagID - the interactions were never well
>   understood.
>
>* Experienced practitioners recommended against using it.
>
>Please reply to this email, copying www-rdf-comments@w3.org
>indicating whether this decision is acceptable.
>
>Dave
>
>
Received on Wednesday, 9 April 2003 10:38:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:15:38 GMT