Re: Fragments - specific proposal

On Dec 8, 2007, at 4:59 AM, Carsten Lutz wrote:

> Hi Jim,
>
> On Fri, 7 Dec 2007, Jim Hendler wrote:
>>
>> In light of the conversations and presentation at the f2f as best  
>> I could follow - I propose the following which I believe is  
>> actually within the spirit of consensus
>
> I strongly disagree. You are addressing here the issues of
>
> 1. whether there should be an official OWL Light
> 2. whether there should be one or several OWL Lights
> 3. what precisely it/they will be.
>
> Based on the discussions at and around the meeting, I think it is fair
> to say that all these issues are highly controversial. My  
> impression is
> not at all that what you propose is in the spirit of consensus.

consensus means finding a compromise in the middle of a lot of  
opinions, that is what I'm striving for


>
>> Summary (details below)  I propose we redefine OWL Lite to be  
>> somewhat less than is currently in the language, that we name a  
>> new subset with the intent of adding just enough OWL to RDFS to  
>> match current "low end" needs, and I propose that we create an  
>> easily findable Working Note and Wiki page with named subsets of  
>> OWL that are provably polynomial based on Bernardo's current  
>> document.
>
> Could you explain what is the difference to the OWL Light from the 1.0
> standard? You claimed yourself that that was broken. What is it that
> makes the new approach different?
>
> It is clear that different people have different ideas of what  
> "broken"
> means here. To me, the old OWL Light was computationally broken  
> because
> it wasn't really easier than full OWL DL. The same applies to the  
> fragment that appeared on Zhe's slides, i.e., it seems ExpTime-hard.

I don't personally believe the problem with the current OWL Light was  
that it was broken because it was exp time hard, but agree with you  
that it was broken because it wasn't really easier than OWL DL.  The  
real issue is what we mean by easier - my personal feeling was we  
made OWL Light too feature rich.  Oracle's OWL Prime is less rich,  
and I'd be willing to take out more.

What I do believe is we should continue to have a fragment named OWL  
Light.  Personally, if I could make it whats in RDFS 3.0 or similar  
to that I'd be much happier, but it was clear to me the group did not  
agree.


>
>> 1 - OWL Lite
>> I propose we redefine OWL Lite to be closer to the current  
>> language coverage that is in Oracle's OWL Prime.
>> For the DL version of this, we will have a model theory and the  
>> restrictions needed to prove whatever it is we wish to prove about  
>> it (completeness, decidability, etc)
>> There will be a Full version which is defined as the subset of the  
>> OWL Full Semantics (however we end up defining those)  
>> corresponding to the same language constructs (note: this may well  
>> require removing some language features that Oracle and others  
>> currently support from this named subset)
>
> I vote against this. I have sympathy for having *an* (not *the*) OWL
> fragment that is motivated by database applications. But I absolutely
> disagree with your approach of making this particular fragment, which
> is not even remotely understood, the main OWL Light, and all the other
> fragments just other fragments that are mentioned in some docs. From
> my perspective (which may be different from yours), this would simply
> mean repeating the mistake that was made with the original OWL Light
> (which was a failure).

so make a counter proposal - what do you want in it?  I already said  
that if we could express EL++ in a way that was easily understood I'd  
be willing to explore that being OWL Light - my proposal is that we  
DESIGN the fragment, not necessarily pick an existing one.   But if  
you want something else, I think it is incumbent on you to make clear  
what your motivations are -- I did so on the fragments page on the  
WIki, and continue to invite you and anyone else interested to  
outline other proposals -- it's easy to argue against things, let us  
know what you are arguing for and why.

>
>> 2 - RDFS 3.0
>> I propose we name a subset called RDFS 3.0 which is less than OWL  
>> Lite - aimed primarily at universals - i.e. named classes and  
>> properties, no restriction statements involved.
>> There should be a version of this which is provably polynomial  
>> within certain restrictions (at least no redefinition of the  
>> language features, possibly
>
> Then it would IMHO be appropriate if some of the supporters of RDFS
> 3.0 would state precisely what this tractable fragment is and prove
> that it is tractable. Otherwise, I feel I am discussing a ghost.

again, the goal is to DESIGN something - but note that I do not find  
polynomial to be a big deal - there's a lot of languages out there  
that are "effectively" polynomial even though they have worst case  
behaviors that are not.  Most of the most commonly used OR  
techniques, for example, either have that behavior or have  
exponential behavior with very low exponent.  So I care more about  
things like learnability, ease of parsing, etc - however, this WG  
made it clear they don't think those "Soft" reasons are enough - so  
again, my compromise is to offer space to design a polynomial set in  
the region of the most used languages.


>
>> those DL restrictions required to keep this true - ie only  
>> inversefunctional datatypes used as keys per Uli's presentation,  
>> or whatever).  This version could be called something like RDFS  
>> 3.0 DB (if we get it close to datalog) or DL (if that's the best  
>> we can do)
>
> So there will be different versions of RDFS 3.0. I would like to raise
> the question whether the OWL workgroup is the right place to do this.
> Both in spirit and from the name, RDFS 3.0 seems to be much closer
> to RDFS than to OWL. So shouldn't it be standardized in an RDF group?

There is no RDF group right now.  We own it.  But if you don't like  
the name, go ahead and give it something else.  But it is the most  
used fragment of OWL out there, bar none (FOAF falls in this, DOAP  
falls in this, SKOS is being moved towards this, etc. etc -- there's  
literally billions of RDF triples pointing to these ontologies - so  
when I say "used" i mean Web-scale used.

>
>> The language would likely be widely used (or more correctly, is  
>> already widely used) -
>
> I again point out that there are fragments that *are* widely used
> already, which seems much stronger than "likely be widely used". These
> fragments you want to send to a WG note. Again: I have sympathy for
> what you are trying to do. But let me follow Jeremy and emphasize a
> letter from our acronym OWL (though a different one): O is for
> *O*ntology languages. I point out once more that very important
> ontologies such as SNOMED, NCI, Gene ontology, and the medical
> ontology produced by Sebastian for Siemens Health are written in
> EL++. All of these ontologies are large-scale, professionally
> supported, and used in real (even commercial) applications. Can you
> say this of your (or any other) fragment? I cannot see that you do
> justice to this when moving "all other fragments" to a WG note.
>

Carsten - I think you and I use "widely" at different orders of  
magnitude, but that said, if EL++ could be defined cleanly in terms  
that non-logicians could understand - while it is nice to say things  
like "Concrete domain" to a group of DL folks, a bunch of us non- 
logicians have no idea what it means - this is my main problem with  
the current fragments document, it is tied to technical definitions  
that I, and recall I supported one of the worlds top DL groups for  
the past few years, cannot fully understand --  us "hackers" also  
have a real place in the real world, and if we define OWL and its  
fragments in a way such folks cannot understand, then we get what we  
deserve with respect to OWL not getting used...

> greetings,
> 		Carsten
>
> --
> *      Carsten Lutz, Institut f"ur Theoretische Informatik, TU  
> Dresden       *
> *     Office phone:++49 351 46339171   mailto:lutz@tcs.inf.tu- 
> dresden.de     *

"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