W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > April to June 2004

Re: Use Case: Searching for educational learning materials

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Wed, 23 Jun 2004 05:57:58 -0400
Message-Id: <CE57C834-C4FB-11D8-AED7-0003939CCC42@isr.umd.edu>
Cc: kendall@monkeyfist.com, RDF Data Access Working Group <public-rdf-dawg@w3.org>
To: Dave Beckett <dave.beckett@bristol.ac.uk>

On Jun 22, 2004, at 7:07 AM, Dave Beckett wrote:

> On Sat, 19 Jun 2004 10:01:43 -0400, Bijan Parsia <bparsia@isr.umd.edu> 
> wrote:
>> On Jun 19, 2004, at 9:04 AM, Kendall Clark wrote:
>>> On Sat, Jun 19, 2004 at 11:28:38AM +0100, Steve Harris wrote:
>>> You could use the N3 syntax, seperating the literal from the datatype
>>>> and
>>>> language with ^^ and @, eg.
>>>> 	(foo, rdfs:label, ?x@jp)
>>> Hmm, that's an idea.
>> I ran this by our FLA guy as one option from the start. This is
>> certainly how I would do it in SWI Prolog, at least for the common
>> case.
> (FLA guy?)

Fujitsu Labs of America. This is a requirement generated from their 
Task Computing environment.

> I think this a bad idea since it requires looking inside a RDF concept


> which as far as I am aware, is not usually indexed by the common
> RDF/OWL systems in that manner, although this does not preclude it.

Does it need to be, in the common case?

> Adding this requirement to a DAWG QL I think would make it less
> likely that this part would be implemented.

Ah. Hmm. Well, I won't argue this point further as it's not my prefered 
syntactic solution and if I'm dealing with a heck of a lot of lang 
filtering, I can beat on my vender to do the right sort of indexing 
regardless of syntax (I think).

>> The other possibility was having a filter predicate for AND clauses
>> (e.g., langEq). The latter was prefered.
> This is what I would prefer, selecting the language aspect of an RDF 
> literal
> after SELECTion.  This is what SeRQL does if I recall correctly, there
> is a language() predicate for the constraint section -

Ah, good, there is an existing predicate. We will be implementing this 
in the next month or so.

> so it is already
> implementable, and implemented.  If some system did index on RDF
> literal language it clearly could use this to advantage here.

Yes. Ok. I think we agree.

>> [snip]
>>> I think Yoshio mentioned something back in the early days of the WG
>>> about being sensitive to i18n issues, which this feature request
>>> does. (If anyone needs to hear a use case, I can provide one related
>>> to task computing in multilingual environments.)
>> Or portals, or even ontology browsers. Anything that makes significant
>> use of rdfs:label and wants to internationalize.
> I've built a few of those and it would haven been easy to look at the 
> RDF literals
> returned and try to select those with an appropriate language (this is 
> not a
> straight compare, but can  involve matches such as 'fr' with 'fr-ca')

Do you mean in application code? That is, of course, a possibility, 
but, IMHO and to our customer, not a very attractive one.

Bijan Parsia.
Received on Wednesday, 23 June 2004 05:58:09 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:00:44 UTC