W3C home > Mailing lists > Public > public-owl-wg@w3.org > March 2008

Re: A proposal for the fragments document

From: Ivan Herman <ivan@w3.org>
Date: Wed, 05 Mar 2008 14:03:40 +0100
Message-ID: <47CE9A2C.9060006@w3.org>
To: Boris Motik <boris.motik@comlab.ox.ac.uk>
CC: public-owl-wg@w3.org


Boris Motik wrote:
>>
[snip]
> 
> The intention here was not to refer to the RDFS fragment from the member submission; rather, we wanted to say that DL-lite captures
> the intersection between RDFS and OWL 1.1 DL. Would that be any better? 
> 

Yes.  I mean, the statements would then say:

"it is, therefore, a proper super-set of the intersection between RDFS 
and OWL 1.1 DL."

But... isn't that statement also true for EL++ as well as for OWL-R? If 
so, than it becomes a meaningless statement, because it does not 
differentiate DL-Lite from the other two...

Maybe I am missing something here.

>> 2. In the spec of EL++ and DL-Lite 'facts' is listed as features
>> provided by EL++. Eg, I can use SameIndividual or, in RDF speak:-)
>> owl:sameAs. This is not present in the OWL-R DL spec; actually, there is
>> no (handy) list of feature overview. It took me a certain time (and
>> looking at the OWL-R Full version) to find the owl:sameAs is indeed
>> usable (and this is probably by far the most widely used owl term used
>> out there these days...).
>>
>> I guess this is really editorial: the description of OWL-R DL should
>> follow the same pattern as the other two...
>>
> 
> In the member submission, each fragment was defined in full, by repeating all the supported productions of the language. This turned
> out to be quite error-prone (the document did contain some omissions), and it also resulted in a cluttered definition.
> 
> Therefore, in this new document we deliberately chose to define each fragment as a "diff" to the original spec. In other words,
> everything that we didn't redefine applies as usual.
> 

But you did not do it in a consistent way! As I said: you refer to the 
facts for EL++ and DL-Lite, but not for OWL-R...

> Now I fully understand that this might result in a slightly inaccessible spec. We can address this drawback in two (nor necessarily
> disjoint) ways:
> 
> - We can expand the overview section for each fragment and be more explicit and precise about what is and what is not supported.
> 
> - We can fall back to the original style of definition and repeat all the "unchanged" productions.
> 

Personally, I would favour the former. I expect lots of users of these 
fragments who (a) are not sophisticated users of DL and the like but (b) 
would be very grateful and happy to use DL reasoners for their simple 
vocabularies, if possible. At present, the document essentially asks 
them to understand the full functional syntax which may be much more 
than what they can or are willing to do; even for myself, I would prefer 
to have a list of OWL/RDF terms that I can or cannot use, and that is it...

Actually, it would be nice to add one more table somewhere, listing the 
RDF terms (or family thereof) and add coloumns for the three fragments 
denoting which or those are usable... There should be a quick and easy 
way for me, as stupid and not-logic-expert-user to decide which feature 
I can use in my RDF encoding and which I cannot.


> Depending on what the group deems more suitable, we can adopt either (or both) approaches. We should be able to modify the documents
> without much ado.
> 
>> 3. In the fragment complexity table you refer to OWL DL in the top row.
>> Is that OWL1.1 DL or the 'old' DL? I am a bit confused with all the
>> acronyms:-(
>>
> 
> Yes, this was meant to be OWL 1.0 DL. In fact, the table should also contain a row for OWL 1.1 DL. I'll add them in a bit.
> 

Great.


> Thanks again for the comments; I really appreciate them.
> 

Sure. Thanks!


Ivan


> Regards,
> 
> 	Boris
> 
>> Ivan
>>
>>
>>
>> Ivan
>>
>> Boris Motik wrote:
>>> Hello,
>>>
>>> In the past couple of days Bernardo and I have come up with a proposal document for the OWL
>> fragments story [1] and we thought that
>>> it would be useful to send it around before tomorrow's discussion. Let us know how you feel about
>> it.
>>> Regards,
>>> Boris & Bernardo
>>>
>>> [1] http://www.w3.org/2007/OWL/wiki/Fragments_Proposal
>>>
>>>
>>>
>> --
>>
>> Ivan Herman, W3C Semantic Web Activity Lead
>> Home: http://www.w3.org/People/Ivan/
>> 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/
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf


Received on Wednesday, 5 March 2008 13:03:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 5 March 2008 13:03:52 GMT