ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Extensional vs Intensional semantics for RDF/RDFS

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Sun, 24 Nov 2013 13:48:44 -0500
Message-id: <52924A0C.3070201@xxxxxxxxxxx>
Pat, Enrico, Michael, and Simon,    (01)

I agree with Pat that it's too late to discuss a definition that is
already specified as an official standard.    (02)

EF
>> But I’m not willing to argue about what is obvious and what is not,
>> since there may be different categories of 'obvious', so let’s not go
>> on with this potentially infinite discussion    (03)

PH
> We are not having a discussion. I am telling you why the RDF(S) data
> model was designed in the way that it was, which involves not treating
> classes as identical to their set extensions    (04)

But as somebody who has spent many years teaching students something
about logic and ontology, I am always concerned about how to present
the material as simply, accurately, and understandably as possible.    (05)

Goal:  the KISS slogan.    (06)

MB
> The entailment that a property has a certain domain or range
> actually makes a difference to this application: Facts are copied
> into less tables if the corresponding property has known
> domains/ranges and the entailment may produce those domains/ranges.    (07)

I agree with this observation.  But the following comments raise
serious issues that do require discussion:    (08)

MB
> So if this stuff is missing in RDF 1.1, maybe some prominent warning
> about what is not entailed should be there.    (09)

EF
> In all my discussions with people around the world, I always had
> a surprise reaction; people never realised these inferences were
> missing and they wonder why they don’t have them.    (010)

PH
> But most RDF users have not even read the OWL specs.    (011)

SS
> There are a few warnings ⚠ of what is not entailed ⚠ - e.g. section
> 10 on Datasets, and appendix D.1 on reified triples. These are cases
> where there are obvious intuitions that do not apply in RDF(S).    (012)

Guha was the associate director of Cyc, which was decidedly not easy
to learn.  So he designed RDF as a "simple" subset of logic that he
hoped would be easy to teach, learn, and use.    (013)

RDFS was a simple application of RDF for specifying some simple
constraints on an ontology.  OWL was designed as subset of FOL that
was more expressive than RDFS, but it was harder to learn and use.
So they gave us SKOS as Yet Another Simplified Logic (YASL).    (014)

But these email notes and the voluminous W3C clarifications have
raised serious concerns:  The YASLs don't meet the KISS test.    (015)

Steve Jobs had a better idea:  Don't write bigger manuals to explain
the difficulties.  Design systems that eliminate the difficulties.    (016)

Question:  What should we do about (a) the current logics and
(b) future directions for new ones and revisions to the old ones?    (017)

John    (018)

_________________________________________________________________
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    (019)

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