RE: FW: Cursors face defining moments on the Web

> I'm curious: which of "RDF's lessons" do you mean? What possible use
> of RDF could there be for Comet?

Good question.  Although, using RDF wouldn't necessarily mean
you were attempting to learn from its lessons.  The way I see
it there are two things you could assume:

1) People are gonna start doing semantic things with the web
(Comet is evidence).  If people stick to standards whenever
possible, we stay interoperable.  RDF is one attempt at a
standard for "semantic things", so perhaps people would use
RDF where appropriate in their semantic things simply as a way
to be respectful to future kids trying to make all this stuff
work together.

2) RDF represents a great deal of thought and some
debate from its creators about how the future of the
semantic web would develop.  Presumably, the standard is
developed with a "bigger picture" in mind, and embeds some
"lessons" about the difficulty of such that the creators
discovered while developing the standard.  So one could
hope to be aligned with the future and maybe  to avoid
pitfalls of a custom approach by adopting RDF.

Now to attempt an answer to "what possible use of RDF could
there be for Comet"?  In my earlier message [1], I suggested 
that there are only a few ways that a tool could arrive at 
semantic information about a particular resource.  In other
words, there are just a few ways that comet could be determining
that a particular phrase matches an entry in Britannica Online
or that it is a band documented at RollingStone.com.  Is comet
using RDF?  I doubt it.  Is comet using markup at all?  I doubt
it.

I completely respect the right of comet to pursue a business
decision of being "proprietary" so that they can make money
off of partnerships.  In other words, for them to support markup
that allowed the page designer to specify any generic music
site would obviously hurt their relationship with RollingStone.
However, imagine the case of a page with the word "Madonna".
Would comet not want to give the web page designer the power
to "hint" that "Madonna" in this case should point to the
"mother of Jesus", while another page would hint that it
should point to the former Ms. Ciccone, and yet another to
the RollingStone pages about her music?  In other words, if
cursor always attempts to "guess" what Madonna means, they will
sometimes get it wrong and offend people, and they can easily
solve this with a moderate use of markup.  If they use markup,
why not RDF?  This is a very silly example, but you should get
my point.

(Of course, the fear of any such company is that they will
encourage proliferation of the semantic info and then someone
else will come along and write a tool that "poaches" that
info in ways that were unintended.  This is, in fact, TimBL's
secret plan, I think :-)  If people are afraid that the unintended
use of their semantic info will not benefit them, things could
develop very balkanized.  I believe the example I show above
for comet is one sample of how baby steps could be taken without
jeopardizing the value of the business agreements that provide
incentive to step forward at all, but like I said I haven't 
thought much about it.)


P.S. Speaking of encyclopedias, it seems that FSF has
embarked on an "open encyclopedia" at www.nupedia.org.
In the W3C Architecture Note on Extensible languages [2]
is a reference to an interesting paper [3] that discussed
the idea that "Britannica" and other encyclopedias get
their value from how "authoritative" they are and that
the ideal example of a non-authoritative, open encyclopedia
is the WWW itself (presumably with search engines/directories).
So I am wondering exactly what space nupedia is hoping
to occupy.  Oh well...

--

[1]
http://lists.w3.org/Archives/Public/www-rdf-interest/2001Jan/0158.html
[2] 
[3] http://www.hf.ntnu.no/anv/WWWpages/Hyper/Hypermedia.html