ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Next steps in using ontologies as standards

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Pat Hayes <phayes@xxxxxxx>
Date: Tue, 20 Jan 2009 11:55:13 -0600
Message-id: <C2B7BBD0-44B0-455C-929B-944118B09192@xxxxxxx>

On Jan 20, 2009, at 8:51 AM, John F. Sowa wrote:    (01)

>
>> [ Ian Bailey wrote:] ....  I'm
>> not sure OWL and RDFS give you a proper foundation for ontology
>> development - there are some very strange things in the W3C spec
>> about how an individual in one ontology can be a class in another
>> (bizarre even in an intensional approach).
>
> I very strongly agree.  RDFS and OWL are horrible examples of how
> *not* to design an ontology language.  The designers started with
> two disastrous implementation-based assumptions:
>
>  1. They wanted to reuse their XML-based parsing tools by forcing
>     everything into the world's worst syntax.
>
>  2. They forced a weird semantics in which the only relations
>     are dyadic.  That means that you can't even say 2+2=4
>     because the "+" operator is triadic:  it takes two inputs
>     and generates one output.
>
> These two blunders are the source of those bizarre features you
> mention above.  You can't entirely ignore RDF and OWL because
> they were foisted on a large set of people who didn't know enough
> to see that they were dupes in a Ponzi scheme.  But you should
> always preserve your sanity by thinking in terms of something
> better, and Chris P's book is a good place to start.    (02)


---------    (03)

On Jan 20, 2009, at 9:17 AM, paola.dimaio@xxxxxxxxx wrote:    (04)

> when I first came to this forum I made it clear that
> (for me at least) 'semantic' and ' ontology' are not equal to RDF/
> OWL, despite a lot of misleading indoctrination perpetrated by
> semantic web institutions
>
> ( I recently began arguing with a tutor at ASWC summer school who said
> that an owl file is is an ontology in the w3c sense, and I wanted to
> cry)
>    (05)

------    (06)

Hey, guys, lighten up. (A Ponzi scheme?? Indoctrination? Foisted?  
Dupes? Wanting to cry because OWL is an ontology language?). Nobody,  
even the people who wrote their specs, claim that RDF and OWL are the  
ultimate solution or to all problems. But (1) they do represent a  
viable approach to the problem they were designed for, which was to be  
machine2machine communication notations for the semantic web, and (2)  
if we must criticize them, at least let the criticism be informed.  
Almost everything that John says here is just wrong.    (07)

Ian's complaint was the 'bizarre' property of RDF (not OWL-DL, in  
fact, though this will change in OWL2) which allows a class to be an  
individual. John says "I strongly agree", which is strange, as this  
particular property is foundational also to Common Logic, a project  
with which John is very publicly associated.    (08)

First, to Ian: In fact, its even more bizarre than you think: one  
thing can be both an individual and a class (and a property) in the  
very same ontology. But this only seems bizarre if you have been  
indoctrinated by an old-fashioned and limited perspective on  
representational formalisms. This particular indoctrination is  
widespread, and it was no more than common sense until about 15 years  
ago, when some basic advances in logic showed that the traditional  
'layering' of descriptions into individuals/classes/properties/ 
metaclasses/etc. was (a) not necessary and (b) expressively very  
restrictive. One can keep the categories but abandon the strict  
layering - in effect, allowing a given thing to be in many 'layers' at  
once - and no disasters arise, if one cleaves to a certain simple,  
natural syntactic discipline (which is built into both Common Logic  
and RDF). The result is greatly increased expressivity and a formalism  
which 'naive' users invariably find quite natural, and which makes  
perfect semantic sense. For example, in RDFS the class of all classes,  
rdfs:Class, contains itself. This "ought" to be impossible, a  
conceptual error. Naive users tend to find it simply obvious; it turns  
out, they are correct. The perception of bizarreness that you report  
is felt only by those who have received a training in one of the  
classical indoctrinations, typically an undergraduate course in  
mathematical logic. No-one has yet written a basic logic text which  
uses the newer insights, unfortunately. There is a lesson from the  
history of OWL. OWL-DL, the most widely used version, rejected this  
generalization and kept the traditional layered structure. The  
pressure from users, including consortia of professional ontology- 
writers, to relax this restriction was overwhelming, and the first  
revision of OWL, OWL2, under final review as we write, will move to a  
version of the newer approach.    (09)

To John: several points. (1) RDF syntax is in fact (that is,  
definitively according to the specs) defined abstractly, as a graph.  
An RDF graph is Pierce's graphical logic notation without negation.  
The XML serialization is only one possible serial syntax for RDF, and  
others are also in wide use. I personally refused to use the XML  
syntax in the RDF semantics document, as its relationship to the graph  
syntax, relative to which the semantics of RDF are defined, is so  
obscure. (2) Its true that RDF and OWL syntax is restricted to the  
(binary + unary) (= property + class) case; but this is hardly  
"weird". The fact that this is expressive enough for arbitrary logic  
has been well-known for almost a century, and is relied upon by a wide  
range of Krep notations and database conventions. It underlies the  
famous 'case-role' approach to giving more 'natural' expression of  
complex predications, as I am sure you know. So in this regard, RDF is  
simply following the herd. The binary reduction has both pros and  
cons. The cons are obvious and often noted, but the most important pro  
is that it permits the graph syntax to eliminate all notions of cyclic  
ordering on arcs, so that graphical syntax reduces simply to a  
collection of directed arcs. This was the primary reason for its use  
in RDF, in fact. (3) Neither XML nor the use of the unary/binary case  
have anything at all to do with the "bizarre"ness that Ian mentions.    (010)

To Paula: An OWL file *is* an ontology in the W3C sense. All this  
means is that (1) "an ontology" refers to a collection of formal  
sentences, which has been accepted terminology in this field since its  
inception, and (2) "in the W3C sense" means that the formalism being  
used is one of those sanctioned by a W3C-published specification  
document, so that existing Web processes which conform to the  
standards can emit, transmit, input, parse and process it  
appropriately. If this makes you want to weep, you are working in the  
wrong field. Notice that nothing here says or even hints that there  
cannot also be other formalisms or ontologies.    (011)

Pat Hayes    (012)


------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes    (013)






_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (014)

<Prev in Thread] Current Thread [Next in Thread>