RE: [OEP] Scope - (was philosophy of SWBPD (was Re: [OPEN] and/or [PORT] : a practical question))


At 12:52 -0700 4/5/04, Uschold, Michael F wrote:
>  -----Original Message-----
>From:	Jim Hendler []

>Ahhh - the root of a lot of our disagreements becomes clear -- in my
>world view, reasoners are not likley to be a major force on the Sem
>Web, and those that are are not likely to last very long if they
>cannot handle inconsistency -- I won't make the argument here, I've
>done it lots of times in many other forums and in print (see my talk
>at ISWC last year [1] for one example) -- thus, I think that the
>claim "being able to prove anything at all" is clearly just plain
>ridiculous for any tool that has any chance of working on the Web.

><MFU: are you sayng that the the logicians claim that inconsistency 
>enables a reasoner to prove anything is not relevant in the  context 
>of the Web?  If so, that might be true, and argues for not worrying 
>about consistency. If not, then I don't know what you are saying. I 
>cannot parse it into English that I understand, because "being able 
>to prove anything at all" is not a claim, per se it is a just a 
>sentence fragment.

I'm claiming that pragmatically we'd better learn to deal with it --- 
consider, if I publish a document that says:

I am male.
I am female.
Male and female are disjoint classes.
You owe me a thousand dollars

and your reasoner sends me the money, then I don't think you'll be in 
business very long.   More seriously, the problem is that the 
inconsistency can (and I argue will) arise from many sources -- 
accidental (you type 5 and I type 55 when we mean to say the same 
number), conflicting views (see if you can get right to lifers and 
abortion proponents to agree on what "birth" is), and from 
maliciousness - as in the case above, if I can fool your reasoner 
into making wrong conclusions, I have a whole new kind of net attack 
to explore.  There's many ways to deal with these kinds of 
inconsistencies, but note that the key is that we will have to deal 
with them somehow...

>So in that light, I have no objection to consistent stuff, I just
>don't see it as the single most important thing to always worry
>about...   but notice Mike, we' ve just turned this into a trade-off
>discussion - so where you said we sometimes have to take sides, I
>just basically think this is harder than you think....
>I doubt I could agree to this - I have no idea what "all other things
>being equal" means in a dynamic changing system like the Web.
><MFU: I mean something very simple. You are building an ontology for 
>a specific application. You have a very simlilar notion come up in 
>several circumstances and there are several ways to represent this 
>notion. If you represent the same notion in several different ways 
>for no reason, then it is likely to confuse people trying to 
>understand and use the ontology. They will keep wondering: did he do 
>it all these different ways for a reason? It would be simpler and 
>more perspicous to do them all the same way.  

ok, I'll grant you that within a single ontology DOCUMENT (I won't 
get into a fight now about how you define what an ontology is when 
they are linked to others and include some imports) it probably makes 
sense to define things in a straightforward way and to document 
disparity from that ...

>In case anyone hasn't figured it out by now - I THINK IT SHOULD BE
>Working Group.  If you'd like me to state it clearer, let me know
>what to addd
><MFU For starters, it will be necessary to give clear criteria for 
>determining whether something is or is not "related to the Semantic 
>Web". Next it would be very helpful to apply those criteria for 
>several examples where we can all see that aha, this ontology 
>engineering guideline is clearly not related to the Semantic Web, 
>but those are. 
>I would be surprised if it was possible to do this. As it is, for 
>example, some of the papers in the first issue of the new Journal of 
>Web Semantics don't have much if any specific relevance to the Web, 
>semantic or otherwise. Yet, they address important ideas that CAN 
>are are LIKELY to be applied to the Semantic Web.  And so, they were 
>deemed appropriate for the journal.  Analogously, many/most ontology 
>engineering guidelines that we may come up with that are illustrated 
>with OWL are likely to be used by people building ontologies for use 
>in the Semantic Web. The fact that they are illustrated with OWL 
>although  key in one sense, is quite irrelevant with respect to the 
>underlying OE guideline, which could be expressed in many different 
>ontology/KR languages.

Now you finally get to the crux, although backwards in a way -- Web 
Semantic is a RESEARCH journal -- but read the W3C process documents 
and our charter and you'll see that WGs do NOT so research -- this 
was always a challenge in the OWL WG, and will be one for this WG as 

>Yes, but it wouldn't be a SEMANTIC WEB langauge, and thus I would
>argue (see above) this WG should not spend time discussing it
>They ARE central to the design of OWL, in the sense that OWL is
>specifically FOR the Web, and thus had to have a few things that
>typical KR/O languages lack.
>Yes, and strangely, this makes them SIGNIFICANTLY new and different,
>and this WG, being part of the W3C SEMANTIC WEB activity should
>primarily be concerned with the new aspects.
><MFU: Following on from above, perhaps you can clearly state exactly 
>what the {Semantic Web}-specific parts are and give some examples of 
>some guidelines that are useful and specifically focus on these new 

I can name that tune in one line:
  It is the use of RDF as the basis for the language, allowing linking 
between separately created ontologies.

In fact, I don't even have to make this up -- here's what it says in 
the OWL FAQ [1]:

Q. How is OWL different from earlier ontology languages?

A. OWL is a Web Ontology language. Where earlier languages have been 
used to develop tools and ontologies for specific user communities 
(particularly in the sciences and in company-specific e-commerce 
applications), they were not defined to be compatible with the 
architecture of the World Wide Web in general, and the Semantic Web 
in particular.

OWL rectifies this by providing a language which uses the linking 
provided by RDF to add the following capabilities to ontologies:

     * Ability to be distributed across many systems
     * Scalable to Web needs
     * Compatible with Web standards for accessibility and internationalization.
     * Open and extensible

so I think if we stick to best practices that draw from these four 
key capabilities (arrived at with great work, and signed off on by 
the members of the W3C)



Professor James Hendler 
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  240-277-3388 (Cell)

Received on Monday, 5 April 2004 17:09:43 UTC