Re: draft response for LC comment 26 (a and b)

I guess I disagree as to whether the distinction of QL and RL is as  
clear as this, I think most users of RL care about databases (which is  
why I understand Oracle backs it) - as best I can tell the difference  
between QL and RL has to do with reduction to datalog and differences  
in worst case behaviors in certain situations.   I also think the  
description of EL will scare people off of DL (if EL is reasoning with  
good time performance, that means DL doesn't?)
   Frankly, in practice I'm finding it extremely hard to explain why  
three profiles are needed, and I would again propose that we consider  
moving the other documents to CR, but hold back the profiles document  
for further work on explanations and to avoid confusion that could  
lead to less adoption of DL/Full, which still seems to me to be where  
the most important OWL 2 extensions currently live.
   -JH


On Feb 16, 2009, at 11:07 AM, Christine Golbreich wrote:

>
>
> 2009/2/16 Ivan Herman ivan@w3.org
>
>  ....
> Concerning the rest of your comment and NF&R
>
> That said, I
> think part of the issue is that there is no clear understanding when  
> QL
> could be or should be used as opposed to, say, RL (or EL or the DL
> altogether for that matter). Neither the profile document nor any  
> other
> gives any help for that and my understanding is that this is Lilly's
> main concern...
>
>
>
> There were not really guidelines for OWL 1 species ... The Overview http://www.w3.org/TR/2004/REC-owl-features-20040210/#s1.3 
>  only introduced  the three (OWL DL, Lite, and Full),  just as NF&R  
> presently does for the OWL 2 profiles  http://www.w3.org/2007/OWL/wiki/New_Features_and_Rationale#F15 
> :_Profiles_OWL_2_EL.2C_OWL_2_QL.2C_OWL_2_RL
>
> That said, we should improve as much as we can.
> I may slightly reformulate the section below, to be more explicit
>
> (1) "Consequently, different profiles of language have emerged, been  
> requested by different types of users - ontologists, DBMS or rule  
> engine developpers -  which correspond to various application  
> scenarios:
> a scalable profile for large but (rather) simple ontologies that  
> enables good time performance for ontology (TBox/schema) reasoning  
> (OWL 2 EL).
> a profile that can easily interoperate with relational database  
> systems, useful for applications where scalable reasoning on large  
> datasets is the most important task (OWL 2 QL).
> a profile that can easily interoperate with rules engines and rule  
> extended DBMS, useful for applications where query answering is the  
> most important task (OWL 2 RL). "
> replacing it by something like
>
> (2) Ontology developpers may consider which profile best suits their  
> needs. The choice between the different profiles mainly depends on  
> the expressiveness required by the application, the priority to  
> reasoning on classes or data, the size and importance of scalability  
> etc. For instance, those who look for
> a scalable profile for large but (rather) simple ontologies that  
> enables good time performance for ontology (TBox/schema) reasoning   
> may prefer to chose OWL 2 EL.
> a profile that can easily interoperate with relational database  
> systems, useful for applications where scalable reasoning on large  
> datasets is the most important task may prefer to chose OWL 2 QL.
> a profile that can easily interoperate with rules engines and rule  
> extended DBMS, and useful for applications where query answering is  
> the most important task  may prefer OWL 2 RL. "
> Would that agrees you ?
> else, if it's not yet enough, could you/anybody suggest a suited  
> sentence to be added to discrimate between OWL QL and RL or what  
> else  ?
>
> (BTW NF&R overview of the profiles points to the Profile and to the  
> Primer for more extensive doc)
>
>  Thanks
>
> Ivan
>
> >> Given the sentence in Lilly's comment "... in particular,  
> identifying
> >> different subsets of OWL2 for developers with limited logic  
> background. ..."
> >> it might be welcome to add that profile checkers* are on the way  
> that will
> >> offer such functionality and allow them for checking just as they  
> did
> >
> > I added a sentence about profile checking to the response,
> >
> > http://www.w3.org/2007/OWL/wiki/LC_Responses/SS1a
> >
> >> Christine
> >>
> >> * as pointed out by
> >> http://lists.w3.org/Archives/Public/public-owl-wg/2009Feb/0035.html
> >
> > peter
> >
>
> --
>
> 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
>
>
>
> -- 
> Christine

"If we knew what we were doing, it wouldn't be called research, would  
it?." - Albert Einstein

Prof James Hendler				http://www.cs.rpi.edu/~hendler
Tetherless World Constellation Chair
Computer Science Dept
Rensselaer Polytechnic Institute, Troy NY 12180