On Jul 7, 2012, at 8:03 AM, John F Sowa wrote: (01)
> Dear Matthew and Pat,
>
> MW
>> Oh dear! What a tangled web we weave :-)
>
> PH
>> True, but not by practicing to deceive. The tangle arose from a conflict
>> between what one might call rival philosophies of the Web, and the need
>> for a committee to produce a single product, which turned out to be the
>> proverbial camel rather than horse. But this was in 2004, and the camel
>> has turned out to be quite useful in practice.
>
> I agree -- and I put the emphasis on conflict, not practice.
>
> Short summary: Too many cooks spoil the pot.
>
> Longer point: Fred Brooks in _The Mythical Man-Month_ wrote the
> classic analysis of how *not* to design a large system. A small,
> tightly-knit design group is essential for designing the core.
> After the core has been designed, then you can expand the number
> of people who work on the applications above the core.
>
> Fundamental error: Any project with the word 'semantics' in the
> title must be based on logic. (02)
You and I know that. Guha knows that. Tim B-L knows that. But there are 10|4
people out there who not only don't know that, but are passionately opposed to
it and will work tirelessly for years to undermine your efforts and oppose you
are every turn, if you try to base any standard on logic. When Guha and I wrote
the Lbase proposal, the RDF WG kindly, with the air of a patient and benevolent
auntie, allowed us to publish it under their auspices. It had as much influence
as a fart on Sunday. (03)
> Instead, the W3C thought of the SW
> an extension of the WWW. That is why they mistakenly thought that
> the WWW core (Unicode, *ML markup, documents, and URIs) would
> suffice for the SW core. (04)
Nobody thought that it would *suffice*; but they did assume, I think correctly,
that the SW must be based on the universal use of URIs, and its constituents
would include marked-up documents. Which immediately introduces a host of
game-changing issues that had not, AFAIK, arisen previously. (05)
>
> More general view: The SW is a problem of distributed knowledge
> representation and reasoning among an open-ended collection of
> heterogeneous agents that must build on, adapt to, and interoperate
> with trillions of dollars of legacy software while preserving
> sufficient generality and flexibility to support anything and
> everything that anyone might imagine in the future.
>
> Hubris: In the mid 1990s, there was 40 years of R & D on the above
> issues, but the full problem is still a research area today. Instead
> of looking at what had worked, what had failed, and why, the W3C
> committee claimed that their problem was so totally different from
> anything else on the planet that they could ignore all the rest and
> start from scratch (06)
The OWL development was largely driven by the state of the art in description
logics, which had a 25 year pedigree in 2004. And there were some new issues,
mostly arising from the use of URIs, which had not been tackled previously, and
which made quite a lot of that older work irrelevant (eg vocabulary alignment
issues.) The RDF design wasnt due to hubris, but if anything to a combination
of wanting to keep things as simple as possible, having to accommodate a host
of existing and deployed ideas and systems, and keep the whole thing having a
coherent model theory. (07)
> -- with a cast of thousands voting on fundamental
> design issues. (08)
It is easy to emit wise criticisms, but somewhat harder to actually achieve
anything positive. Nobody is forcing anyone to use the W3C standards. If you or
anyone else can come up with something better, and persuade the world to use
it, I'm sure they all will. But until you (or someone else) actually makes a
proposal, and even better, gets some people to implement some supporting
software, nothing will happen. I can attest, there is a long and painful road
between griping about how terrible the world is, to actually doing anything to
make it better. (And by the way, even if you build it, they will *not* come,
unless you persuade them that there is some payback to making the switch, and
maybe not even then.) (09)
>
> Summary: I agree with Pat that some useful technology has come from
> the SW. But I keep pointing out that 18 years have passed since
> the founding lecture by Tim B-L in 1994. The useful combination
> of RDF + OWL + SPARQL is far less advanced *and* usable than some
> widely available AI software in the mid 1990s. (010)
That is completely irrelevant. What IS relevant is, how widely any of this
software is actually used. IMO, Common Logic is vastly superior to the entire
RDF/OWL/SPARQL suite, which in turn is superior to deployed RDBM technology.
However, as I have been forced to accept, my opinion is not worth a damn when
it comes to actually getting people to use anything. Usability without actual
use is oxymoronic. (011)
>
> Future directions: It's time to rethink the foundations. Google,
> for example, is doing that. They hired R. V. Guha, the original
> designer (with Tim Bray) of RDF. But their vision treats RDF,
> OWL, and SPARQL as legacy software. I believe that we need an
> open design competition to determine how to proceed from here. (012)
Go ahead and organize it, John. Or else shut up about it. (013)
Pat (014)
>
> John
>
> _________________________________________________________________
> 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
>
> (015)
------------------------------------------------------------
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 (016)
_________________________________________________________________
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 (017)
|