Ok, then it seems like we are all agreed, no? (01)
1) You need an expressive language for ontologies, the more expressive in
general
the better (with some qualification about expressive limits and tractability;
Description Logics as a KR thread has tried to find a tractable sub-First Order
Logic that is as close to FOL as you can get).
2) It's useful to have multiple targeted syntaxes, depending on your needs. In
OWL, we have an XML target (because the W3C and we build on XML and RDF/S and
these languages are increasingly popular). I think Common Logic (ISO-KIF?) has
multiple syntaxes too, no? (02)
But to me the real issue is: what is the base language of representation for an
ontology? If you choose wrong, you may be able to express it in a syntax you
like,
but not be able to express the semantics you'd like. (03)
Leo (04)
Adam Pease wrote: (05)
> Mike,
> I probably phrased my point too strongly. I'm not opposed to expressing
> an ontology in an XML-based form, but we should be clear on why that is
> needed. We translate SUMO into DAML (which is built on RDF and XML, which
> means that we generate valid XML), and we do that so people who have
> DAML-based tools can use SUMO.
> XML is not a good thing in and of itself, but good only to the extent
> that it serves some beneficial and well-articulated purpose connected to
> the issue at hand, which in this case I take to be the storage, inspection,
> manipulation and use of ontologies.
>
> Adam
>
> At 10:42 AM 11/19/2002 -0800, Uschold, Michael F wrote:
> >Adam made two points:
> >1. the language needs to be very expressive, such as KIF
> >2. it is unnecessary to wrap the expressive language in XML syntax.
> >
> >My message only adressed 2, and I disagreed. I agree with 1.
> >
> >Mike
> >
> >
> > -----Original Message-----
> >From: owner-ontology_site22@xxxxxxxxxxxxx
> >[mailto:owner-ontology_site22@xxxxxxxxxxxxx] On Behalf Of Leo Obrst
> >Sent: Tuesday, November 19, 2002 4:59 AM
> >To: ontolog@xxxxxxxxxxxxxxx
> >Cc: 'Christian Fillies'
> >Subject: Re: [ontolog] ontology tools and an ontology repository?
> >
> >I understand Adam's point, but will also underscore Mike's: i.e., yes, all
> >of
> >the Semantic Web languages do/will have an XML serialization, for exchange,
> >etc. But it's also the case that only a specific intepreter will be able to
> >make use of the RDF/S or OWL embedded or serialized in XML, i.e., an RDF/S
> >or
> >OWL interpreter. Perhaps that's obvious. What is not obvious is the internal
> >language used and how things are represented internally. You might view the
> >ontology as being represented in KIF and stored in a relational db, which
> >has
> >import/export converters to the exchange language of choice. Or the ontology
> >is
> >just a set of marked up documents stored as files. Or both (i.e, native
> >RDF/S
> >or OWL ontologies in a relational db) with its own query language (RQL,
> >etc.)
> >
> >I will say that the language of representation does matter, i.e., you won't
> >be
> >able to express rich semantics in a languge that does not allow the
> >expression
> >of rich semantics. So I think Adam's point comes around to: KIF/Common Logic
> >is
> >the most expressive language semantically and all other translations will
> >lose
> >information (by definition, though I don't have a proof, I do know a bit
> >about
> >the various languages expressivity levels). So you don't want to express
> >your
> >ontologies in a rich language, then have them converted to specific targets
> >(e.g., Horn Clauses, etc.), but in those conversion processes you will lose
> >info -- unless of course you convert to an even more expressive logic, e.g.,
> >a
> >second or higher order language.
> >
> >Leo
> >
> >"Uschold, Michael F" wrote:
> >
> > > Adam,
> > >
> > > While I'm no expert on this, I think I disagree. Having an XML
> > > serialization of KIF or whatever language you choose is a good idea - it
> > > saves you from having to build a parser and makes it more widely
> >available.
> > > Just about everything these days has an XML serialization, e.g. RDF,
> > > DAML+OIL, OWL etc. At the very least, there should probably be an export
> > > to XML option.
> > >
> > > Do you think everyone else who thinks it is worthwhile to have an XML
> > > serialization is (select one or more):
> > > 1. wasting their time
> > > 2. has different needs than is being discussed for Leo's ontology
> > > repository.
> > >
> > > In either case, please explain your reasons.
> > >
> > > Mike
> > >
> > > -----Original Message-----
> > > From: owner-ontology_site22@xxxxxxxxxxxxx
> > > [mailto:owner-ontology_site22@xxxxxxxxxxxxx] On Behalf Of Adam Pease
> > > Sent: Monday, November 18, 2002 6:14 PM
> > > To: ontolog@xxxxxxxxxxxxxxx; ontolog@xxxxxxxxxxxxxxx
> > > Cc: Christian Fillies
> > > Subject: RE: [ontolog] ontology tools and an ontology repository?
> > >
> > > Monica,
> > > XML is a general-purpose syntax, that can at some level accommodate any
> > > information, just as an RDBMS can. However, the computational semantics
> >of
> > > a more expressive language will be lost.
> > > For example, "Every horse has a head." is expressed in KIF as
> > >
> > > (=>
> > > (instance ?X Horse)
> > > (exists (?Y)
> > > (and
> > > (instance ?Y Head-PartOfBody)
> > > (part ?X ?Y))))
> > >
> > > KIF defines what "=>", "exists", "and" etc mean, and what any
> >computational
> > > system employing those terms (that is conformant to KIF), must be able to
> > > conclude based on their use. For example, if Ed is a Horse, then there
> > > must exists a head that is part of Ed.
> > > One could encode the statement above in many different ways in XML.
> >For
> > > example
> > >
> > > <implies>
> > > <antecedent>
> > > <clause>
> > > <predicate value="instance">
> > > <argument number=1 value="?X">
> > > <argument number=2 value="Horse">
> > > </clause>
> > > </antecedent>
> > > <consequent>
> > > <existential>
> > > <varlist><var name="?Y"></varlist>
> > > <and>
> > > <clause>
> > > <predicate value="instance">
> > > <argument number=1 value="?Y">
> > > <argument number=2 value="Head-PartOfBody">
> > > </clause>
> > > <clause>
> > > <predicate value="part">
> > > <argument number=1 value="?x">
> > > <argument number=2 value="?Y">
> > > </clause>
> > > </and>
> > > </existential>
> > > </consequent>
> > > </implies>
> > >
> > > This is the sort of thing that's being done in the RuleML effort. The
> > > problem though is getting all this "right" with respect to the semantics
> >of
> > > first order logic, and then creating tools that support those semantics,
> >as
> > > well as the XML-encoded input format.
> > >
> > > This seems unnecessary to me. I'd suggest using an ontology tool that
> > > reads and understands KIF, instead of wrapping another layer of syntax
> > > around it.
> > >
> > > Adam
> > >
> > > At 03:33 PM 11/13/2002 -0800, Monica Martin wrote:
> > > >Why don't we consider using ebXML Reg / Rep to store these ontology
> > > >artifacts, or am I missing something?
> > > >
> > > >-----Original Message-----
> > > >From: Bob Smith [mailto:robsmith5@xxxxxxxxxxxxxx]
> > > >Sent: Wednesday, November 13, 2002 10:52 AM
> > > >To: ontolog@xxxxxxxxxxxxxxx
> > > >Cc: Christian Fillies
> > > >Subject: RE: [ontolog] ontology tools and an ontology repository?
> > > >
> > > >
> > > >Hi Leo,
> > > >
> > > >Thanks for the XML.com survey of 53 tools...what a range of options
> > > >today. What will the market look like in 18 months?
> > > >
> > > >A web-based ontology repository hosted on the site could help shape this
> > > >evolving market by illustrating which features of various tools are more
> > > >in demand than others (at least for this select audience...)
> > > >
> > > >I am using Semtalk (www.semtalk.com ) to support a few client's
> > > >requirements.
> > > >
> > > >Since Peter is developing a survey format, I will just wait for his
> > > >email.
> > > >
> > > >Thanks !!
> > > >
> > > >Bob Smith, Ph.D.
> > > >Tall Tree Labs
> > > >
> > > >
> > > >
> > > >-----Original Message-----
> > > >From: owner-ontology_site22@xxxxxxxxxxxxx
> > > >[mailto:owner-ontology_site22@xxxxxxxxxxxxx] On Behalf Of Leo Obrst
> > > >Sent: Wednesday, November 13, 2002 5:20 AM
> > > >To: Ontolog-forums-cim3-net
> > > >Subject: [ontolog] ontology tools and an ontology repository?
> > > >
> > > >
> > > >All,
> > > >
> > > >We are considering one possibility for our site: an ontology repository,
> > > >wherein folks can register ontologies and/or build ontologies using
> > > >Web-enabled tools, possibly hosted at our site.
> > > >
> > > >So some questions:
> > > >1) How do you feel about:
> > > > a) our site supporting an ontology repository?
> > > > b) our site supporting a Web-enabled ontology tool (for developing
> > > >ontologies)?
> > > > c) none of the above.
> > > >
> > > >2) Which ontology tools do you use?
> > > > a) Can you characterize the tools: i.e., Web-enabled, ontology
> > > >languages supported, cost/licensing, POCs, experience, etc.
> > > > b) Do you know of a tool provider who might support such a public
> > > >effort, hosted on our site?
> > > >
> > > >3) Which ontology languages (knowledge representation languages) should
> > > >be the standard(s) for the ontologies?
> > > > (Some examples: Ontolingua/KIF, Common Logic, OKBC, CycL, RDF/S,
> > > >DAML+OIL, OWL, etc.)
> > > >
> > > >4) Additional Comments?
> > > >
> > > >Thanks!
> > > >Leo
> > > >
> > > >--
> > > >_____________________________________________
> > > >Dr. Leo Obrst The MITRE Corporation
> > > >mailto:lobrst@xxxxxxxxx Intelligent Information Management/Exploitation
> > > >Voice: 703-883-6770 7515 Colshire Drive, M/S W640
> > > >Fax: 703-883-1379 McLean, VA 22102-7508, USA
> > > >
> > > >
> > > >
> > > >--
> > > >To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> > > >An archive of the [ontolog] forum can be found
> > > >at http://ontolog.cim3.org/forums/ontolog
> > > >
> > > >
> > > >--
> > > >To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> > > >An archive of the [ontolog] forum can be found
> > > >at http://ontolog.cim3.org/forums/ontolog
> > > >--
> > > >To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> > > >An archive of the [ontolog] forum can be found
> > > >at http://ontolog.cim3.org/forums/ontolog
> > >
> > > --
> > > To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> > > An archive of the [ontolog] forum can be found
> > > at http://ontolog.cim3.org/forums/ontolog
> > > --
> > > To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> > > An archive of the [ontolog] forum can be found
> > > at http://ontolog.cim3.org/forums/ontolog
> >
> >--
> >_____________________________________________
> >Dr. Leo Obrst The MITRE Corporation
> >mailto:lobrst@xxxxxxxxx Intelligent Information Management/Exploitation
> >Voice: 703-883-6770 7515 Colshire Drive, M/S W640
> >Fax: 703-883-1379 McLean, VA 22102-7508, USA
> >
> >
> >
> >--
> >To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> >An archive of the [ontolog] forum can be found
> >at http://ontolog.cim3.org/forums/ontolog
> >--
> >To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> >An archive of the [ontolog] forum can be found
> >at http://ontolog.cim3.org/forums/ontolog
>
> --
> To post messages mailto:ontolog@xxxxxxxxxxxxxxx
> An archive of the [ontolog] forum can be found
> at http://ontolog.cim3.org/forums/ontolog (06)
--
_____________________________________________
Dr. Leo Obrst The MITRE Corporation
mailto:lobrst@xxxxxxxxx Intelligent Information Management/Exploitation
Voice: 703-883-6770 7515 Colshire Drive, M/S W640
Fax: 703-883-1379 McLean, VA 22102-7508, USA (07)
--
To post messages mailto:ontolog@xxxxxxxxxxxxxxx
An archive of the [ontolog] forum can be found
at http://ontolog.cim3.org/forums/ontolog (08)
|