W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2009

Re: [TF-ENT] RDFS entailment regime proposal

From: Ivan Herman <ivan@w3.org>
Date: Fri, 25 Sep 2009 16:37:26 +0200
Message-ID: <4ABCD5A6.3040007@w3.org>
To: Birte Glimm <birte.glimm@comlab.ox.ac.uk>
CC: SPARQL Working Group <public-rdf-dawg@w3.org>


Birte Glimm wrote:
> I got the message :-)

Great!

> 
> 2009/9/25 Ivan Herman <ivan@w3.org>:
>> Hi Birte,
>>
[snip]

>>
>> - I have a question v.a.v. your first design alternative, ie, excluding
>> axiomatic triples from the set of query answers. If my data has a
>> container, ie, I do have
>>
>> ex:A rdf:_1 ex:Q .
>>
>> in my data and I use the query
>>
>> WHERE { ?x rdf:type rdf:Property. }
>>
>> does it mean that I will not get back a ?x->rdf:_1 mapping? After all,
>>
>> rdf:_1 rdf:type rdf:Property .
>>
>> is an axiomatic triple and the way I understood what you wrote this
>> would be excluded.
> 
> Yes. If we exclude allaxiomatic triples, it does not matter whether it
> occurs in your data or not. That option assumes that you "know" the
> axiomatic triples and that they always follow. It is not the design
> option that I have chosen, but it is another alternative. My currently
> prefered option would return ?x->rdf:_1, but not ?x->rdf:_2 unless
> rdf:_2 occurs somewhere in your data.
> 

yep, that is fine

>> I may misunderstand something but if this is the case, I do not think I
>> like it:-(
> 
> That is good to know. It is not my favourite choice either, but I
> thought I mention it as an option. Once I have some more feedback,
> I'll either remove the design choice options and stick to my choice or
> change my choice if it seems that there is a better way to do it that
> I overlooked or there is a majority of people who like another way of
> doing it.
>

Just to give some more 'meat' to my vague remark of not liking it... I
think the following query is perfectly legitimate and useful in
applications:

Data:

ex:A a rdf:Seq;
  rdf:_1 ex:Q;
  rdf:_2 ex:R;
  ...

Query:

SELECT ?x  ?r
WHERE {
  ex:A ?r ?x .
  ?r rdf:type rdfs:ContainerMembershipProperty .
} ORDER BY ?r


which would give me back the element of the container in order, without
knowing the size of the container in advance, provided we have RDFS and
using your chosen rule that will use only those rdf:_i-s that are in the
dataset. But if all the axiomatic triples are excluded, then this would
not work.

(Actually, this is actually the only way I know to query the
constituents of containers in SPARQL...)

[snip]
>>
>> P.S.1: Would it be possible to transfer this wiki page to the Working
>> Group Wiki? There were some trivial misspelling and my fingers were
>> itching to change those, but, well, you guys are not on the Group pages...
> 
> I should have CVS access soonish and I thought I move all that is in
> WebOnt to the CVS. Would that help or should we maintain a wiki page
> too? I have not yet seen what is in the CVS, will it show up in the
> wiki too?
>

Well, I do not thinks we have cast-in-concrete modus operandi now but, I
think I would not touch the CVS content because that is the editor's
draft (ie, under his/her control), whereas the wiki can be used to
record random thoughts, too. Ie, you have some sort of a scratch pad
that you guys use, and that is fine, but that content would not
necessarily end up in the editor's draft and, eventually, the working draft.

But, as I said, there is no real rule here, so it is also (mainly!) the
choice on how *you* want to operate...


>> P.S.2 As an aside, you may want to look at [1] to see if that is a
>> problem with your mails.
>>
>> [1] http://lists.w3.org/Archives/Member/chairs/2009JulSep/0080.html
> 
> I think my change to the config file helped. I rather get some spam
> than having to unspam a lot of important mails and recieve them with
> delays...
> 
> Birte
> 

Cheers, and have a nice week-end!

Ivan


>> Birte Glimm wrote:
>>> Hi all,
>>> whoever is interested in RDFS entailment: I would be very happy about
>>> comments and suggestions for the RDFS entailment regime as outlined
>>> in:
>>> http://wiki.webont.org/page/SPARQL/OWL
>>>
>>> Ignore the OWL part for now. It is not at all finished and will
>>> certainly change! Axel, can you check whether that would work with
>>> what you have in mind for RIF? My RIF knowledge is far too limited to
>>> judge that :-(
>>>
>>> There are a view design choices that I have listed at the end of the
>>> section on RDFS, which could be an alternative to the currently
>>> proposed way of restricting the answers sets to a finite size. Please
>>> let me know prefer any of them or have any other suggestions. Ideally,
>>> a monotonic behavior for queries would be very nice, but that is not
>>> easy to achieve when solution sequences are possibly infinite without
>>> restrictions.
>>>
>>> I would like to get access to the CVS and work in a real template. Who
>>> can arrange that?
>>>
>>> Cheers,
>>> Birte
>>>
>>> PS: Our email server is still behaving strangely regarding W3C
>>> messages and I do not even get my own messages, so I might get your
>>> answers with some delay only :-(
>>>
>> --
>>
>> Ivan Herman, W3C Semantic Web Activity Lead
>> Home: http://www.w3.org/People/Ivan/
>> mobile: +31-641044153
>> PGP Key: http://www.ivan-herman.net/pgpkey.html
>> FOAF: http://www.ivan-herman.net/foaf.rdf
>>
> 
> 
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf

Received on Friday, 25 September 2009 14:38:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:08:28 GMT