Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

SUO: Re: Enhancing Data Interoperability with Ontologies...




John F. Sowa wrote:

> Danny,
>
> That depends on how you define the semantic web:
>
> DA> It's adequacy depends on the purposes to which
> > it's applied - so far it looks adequate for
> > Semantic Web purposes.
>
> If you limit the semantics of the semantic web to
> what can be done with OWL, then that's a tautology.

How about : "the existing web with improved machine-readability of 
information".

> But as that paper explained, OWL does not support
> units and measures.  And other people are for more.
>
> As Scott McNealy has been saying for years, the
> network is the computer.  The reverse is also
> true:  there should be no distinction between web
> applications and computer applications of any kind.

Sounds reasonable.

>
>
> In other words, the same semantics that supports
> the execution must also support the presentation.

I don't see how that follows.

> That means that the semantics must support everything
> that can be specified in SQL, UML, Petri nets, and
> other declarative languages.  

Why? Why not support what you need in a given context - surely that will 
be simpler.

> You can express all
> that semantics in logic, and you can express all
> that logic in controlled natural languages.

Ok.

>
> DA> ... I await your examples which show how CGs
> > can be used to carry out syntactic translations
> > on XML formats better than XSLT.
>
> I agree that the web must be supported:  that means
> there must be support for URIs and HTML and XML
> formats.  Languages such as Java, Perl, Python, PHP,
> and Javascript (or ECMAscript) do that very nicely.
>
> For declarative languages, I recommend controlled
> natural languages, which would interact with the
> XML files in a way that is similar to the procedural
> languages above.   For all declarative information,
> systems analysts, web designers, and application
> programmers should be able to use controlled NLs,
> with diagrams such as UML for visual supplements,
> and with logic as the internal representation (in
> notations such as CGs, KIF, predicate calculus,
> or even SQL).

I'll be interested in where this leads.

>
> DA> btw, I've seen plenty of argument from the other
> > end of scale, web developers and XML people saying
> > that the RDF model is far too complex, and ontologies
> > absolutely unnecessary.
>
> I certainly agree that RDF is "far too complex".
> I would recommend controlled English as a replacement.
>
> As for "ontology", I would prefer to get rid of that
> word altogether because it has caused far more confusion
> than progress.  The information that is contained in UML
> diagrams and SQL constraints is a superset of the kind
> of information that people are encoding in OWL.
>
> I would propose controlled NLs for all of it.  And I
> wouldn't even tell people that what they were saying
> is either logic or ontology.  I'd just tell them it's
> English -- or French, German, Italian, Japanese, etc.

That's certainly an interesting approach, though there's presumably 
quite a lot of work needed. In the meantime, I suspect the web of 
semantics is more likely to resemble TimBL's layer cake:
http://www.w3.org/2001/09/06-ecdl/slide17-0.html

Cheers,
Danny.

-- 
----
Raw
http://dannyayers.com