Re: comment on the fragment document: (inverse) functional and DL-Lite

Boris Motik wrote:
> Ivan,
> 
> Obviously the current document is a first draft, and can be improved in many ways.
> 
> Adding examples would obviously be a good thing. Describing the fragments in terms of RDF features may not be easy as they are not
> always symmetric (i.e., features can be used in one place but not in another).
> 

Hm. Do you mean it is difficult to do or that it is impossible to do?

> Regarding 4.3.1, the table mirrors the one used in http://www.w3.org/ TR/owl-semantics/rdfs.html
> 

Ah. That was not clear (and the typesetting being different it was 
misleading...) I would still think that, for example, the "CEXT I (x)=" 
at the end of the header cell is a bit strange there.

> Regarding the tables of rules, these are designed to be as easy as possible to understand and to help implementers to use them.
> 

I am sorry Boris, but this is not the answer to my question. Namely, 
what is exactly the relationship between this set of tables and the ones in

http://www.w3.org/TR/rdf-mt/#RDFSRules

TThis is clearly something that implementers have to understand. After 
all, OWL-R Full is an extension of RDFS. Some RDFS rules are repeated 
(eg, rule rdfs9 is the same as the first rule in Table 4), whereas some 
are not (eg, rdfs1, rdfs13, rdfs10). Should an implementation implement 
the RDFS rules, too, plus these? Or should the RDFS rules be disregarded 
(they are only informative, b.t.w.) in favour of these?

Ivan


> Regards,
> 
>  Boris
> 
>> -----Original Message-----
>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org] On Behalf Of Ivan Herman
>> Sent: 06 March 2008 12:33
>> To: Boris Motik
>> Cc: public-owl-wg@w3.org
>> Subject: Re: comment on the fragment document: (inverse) functional and DL-Lite
>>
>> Boris,
>>
>> yep, our mails crossed, I just realized where I got it wrong in:
>>
>> http://www.w3.org/mid/47CFDB45.20902@w3.org
>>
>> so this issue is solved:-)
>>
>> I guess what this revealed, however, and besides my own stupidity, is
>> that it is still fairly difficult to read the document as is. Do not
>> take me wrong: all my attempts and comments are just to try to find some
>> ways of improving it...
>>
>> Though your approach is clearly more concise, it makes it fairly
>> difficult for a casual reader to have a clear picture of what is and
>> what is not allowed in each fragment. This is made even more complicated
>> by the fact that, I believe, many readers of this document may come in
>> without trying to understand the full OWL 1.1 (after all, they try to
>> find the minimum they can use!) and, as it stands now, they do have to
>> have a thorough understanding of the full functional syntax before even
>> having a hope to understand what these fragments are. This may be a lot
>> to ask for, eg, RDF(S) users who try to find a minimal way to make use
>> of better reasoners. (Although it may well be that an overview in this
>> direction should be in another document and not this one. I am not sure.)
>>
>> In any case, I think a table listing the features (in RDF) and adding
>> marks on what can and cannot be used (or under what circumstances) would
>> be great. Also, adding examples, like Bijan just did for the functional
>> syntax, would be *very* helpful.
>>
>> It would also be good to have some clear 'elevator pitch' for each of
>> the fragments, probably in the introduction. Something which say "if
>> your application/ontology is roughly like that, try to check whether
>> this and this fragment works for you, and this will make you happy":-)
>>
>> While I am at it, two more, purely editorial comments, if I may:
>>
>> - I wonder whether the first table in 4.3.1. did not use the wrong
>> typesetting. Ie, on the right hand coloumn the header cell should simply
>> say 'then' and the rest belongs to the table below
>>
>> - there is an overlap between the Tables in 4.3.2 and the RDFS
>> entailement rules in
>>
>> http://www.w3.org/TR/rdf-mt/#RDFSRules
>>
>> I think it would be important to clarify how those two rule sets relate
>> to one another. I would expect (but I may be wrong) that the rule set in
>> this document is real superset of the RDFS rules although maybe some of
>> the datatype-related rule sets are not in this one. (Also, formally, the
>> RDFS entailement rules are informative, not normative...).
>>
>> Actually, to make the relationship easier to swallow, I wonder whether
>> it would not be better to follow the editorial style of the RDFS rule
>> tables. It is only typesetting but, again, it may help the reader who
>> moves from one document to the other. Just a thought...
>>
>> Thanks!
>>
>> Ivan
>>
>>
>>
>> Boris Motik wrote:
>>> Hello Ivan,
>>>
>>> Please let me explain the way in which the fragments have been defined. In the introduction, we
>> said that, for each fragment, we
>>> show the productions that are different from the complete OWL 1.1 language. We did this in order to
>> avoid duplication of the
>>> material.
>>>
>>> Please note that the complete OWL 1.1 language contains the following production. Effectively,
>> <objectPropertyAxiom> is responsible
>>> for OWL 1.1 allowing for (inverse)functional object properties.
>>>
>>> objectPropertyAxiom :=
>>>     subObjectPropertyOf | equivalentObjectProperties |
>>>     disjointObjectProperties | inverseObjectProperties |
>>>     objectPropertyDomain | objectPropertyRange |
>>>     functionalObjectProperty | inverseFunctionalObjectProperty |
>>>     reflexiveObjectProperty | irreflexiveObjectProperty |
>>>     symetricObjectProperty | asymetricObjectProperty |
>>>     transitiveObjectProperty
>>>
>>>
>>>
>>> In the DL-lite section, however, we redefined this production to the following one:
>>>
>>> objectPropertyAxiom :=
>>>     subObjectPropertyOf | equivalentObjectProperties |
>>>     disjointObjectProperties | inverseObjectProperties |
>>>     objectPropertyDomain | objectPropertyRange |
>>>     symetricObjectProperty
>>>
>>> Effectively, this disallows the (inverse)functional object properties in DL-lite. In other words,
>> starting from the <ontology>
>>> nonterminal, there is now no way to get into the <functionalObjectProperty> nonterminal; hence,
>> these are not allowed. The
>>> production <functionalObjectProperty> can thus be seen as "dangling".
>>>
>>>
>>>
>>> We used this approach throughout the document: whenever some production is not redefined, we do not
>> repeat it. Admittedly, while
>>> this makes the definition of each fragment more concise, it makes it also rather difficult to see
>> what exactly is and what isn't
>>> supported by the fragment. But for this purpose, we presented an overview of the supported features
>> before giving the grammar
>>> productions. We listed the supported and the not supported features.
>>>
>>>
>>> Now you are right in that we forgot to list the (inverse)functional object properties in the list
>> of the missing features of
>>> DL-lite. Thanks for that; I've updated the document. If you find further omissions, please let us
>> know and I'll add them to the
>>> appropriate section.
>>>
>>> Regards,
>>>
>>>  Boris
>>>
>>>> -----Original Message-----
>>>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org] On Behalf Of Ivan Herman
>>>> Sent: 06 March 2008 11:04
>>>> To: boris.motik@comlab.ox.ac.uk; bcg@cs.man.ac.uk
>>>> Cc: public-owl-wg@w3.org
>>>> Subject: comment on the fragment document: (inverse) functional and DL-Lite
>>>>
>>>> Boris, Bernardo,
>>>>
>>>> I went through
>>>>
>>>> http://www.w3.org/2007/OWL/wiki/Fragments_Proposal
>>>>
>>>> again today. One thing that I may have missed: I tried to see if I can
>>>> use (inverse)functional properties for DL-Lite or not. I did not find
>>>> any reference to those neither in 3.1 nor in 3.2. Again, I may have
>>>> missed something...
>>>>
>>>> I also have some more editorial comments, but I will send them in a
>>>> separate mail
>>>>
>>>> Ivan
>>>>
>>>> --
>>>>
>>>> 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
> 

-- 

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