Chris, (01)
On that point, we certainly agree: (02)
CM> I agree with this, although it appeared I was praising the
> choice of XML per se. To the contrary, the target of my
> praise was the combination of a *standardized* language + URIs,
> which simultaneously gives us an effective namespace mechanism
> (which alone solves a host of problems) and enables us to take
> advantage of the web's infrastructure (03)
However, the important part to standardize, as we and many other
people agreed decades ago, is semantics, not syntax. The problem
with XML is that it does nothing to standardize the logic. (04)
What XML contributes to syntax was designed to support document
markup, not logic or programming languages. The RDF and OWL
designers had to borrow a page from Procrustes to force a rather
mundane subset of logic into the XML syntax bed. (05)
JS>> An HTML/XML tag that specifies LANG=xxx has proved to be
>> much more versatile and successful; e.g., Javascript and PHP.
>
CM> Sure, I agree -- though note it will take standardization
> on the order of RDF/OWL for every desired LANG. (06)
Designing a new language certainly requires a great deal
experience and good taste. But there have been 50 years of
role models for logic and programming languages, and zero
prior art for mapping any of those things into the extremely
weird formats of XML. Tim Bray, the original RDF designer,
admitted that his first try, which is still the standard,
is far more user-hostile than it should have been. (07)
Just look at PHP and Javascript. The designers copied
familiar syntax from popular languages and adapted it
to the task. A parser for those languages is trivial and
takes a trivial percentage of the entire project effort. (08)
As I said, I like the *ML languages for document markup,
and I believe that some XML-based languages, such as XUL,
have been brilliant successes. But XUL succeeded in an
area where there was no prior art, and the XML formats
were suited to the task. (09)
When designing a new language, semantics comes first.
Second are familiar user conventions and practices.
The edict of forcing all languages into the XML bed
violates both of those requirements. (010)
John (011)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit (012)
|