xml-dev - On frank discussions of RDF, AI, "The Semantic Web", OLAP, etc.

[uche.ogbuji@fourthought.com]

> Thanks for entering the thread.  Much appreciated. 

I don't mean to disappoint, and if it looks as if I can't defend my POV
there's not much I can do about it, but I really don't have the time right
now for lengthy debate on these topics on this list.

I got through XML-DEV by mass-deleting posts on W3C policy and skimmed the
RDF thread out of interest.  Of course I came to attention when I saw my
name.

My main purpose for posting was to point out that my article is probably a
pretty poor stand-in for W3C explanation of "semantic web".  I took great
pains in my article to mention that we use RDF in closed space alone,
where we control URIs, data-streams, and process.  At the end of the
article I explicitly state that we have not tested 4RDF in anything that
could be called an open Web situation (i.e. widely-distributed resources
by arbitrary authors in ad hoc location and with ad hoc content
structure).  Maybe I should have said that RDF itself has not really been
tested in that space, but I'm not sure that is true, and it wasn't the aim
of the article to give a general judgment of RDF.

You ask what RDF lends us that existing relational processing doesn't.  I
already mentioned that it is the extensibility and the flexibility
accorded by less rigorous structure.  In the sort of applications we
build, these things do save the clients time and money.  Of course, you
talk about how "proven" OLAP and other technologies are.  This is a red
flag indicating a nasty time sink to my reading.

No one has the objective tools to debunk the assertion that relational
tools work very well.  This is because they have been so successful in the
marketplace that there is no control case (that I know of) to indicate
that, as I believe, other technologies could have had the same benefit at
hundreds of times less the cost.

But I'm drifting into the discussion I can't afford.  I do hope that I'll
be better able to contribute in the near future, and I have a feeling
that some of our up-coming announcements will result in some imperative to
do so.

The point I should take care to make is this: there is rarely magic in
technology itself.  The magic comes from developers.  You may like Java
and I like Python, and we can beat each other up over it, but in the end,
good developers will write good apps regardless of the tools, and bad
developers will continue to sink dot-coms.  The same goes for RDF.
Fourthought has built decision-support and content integration systems
that cost far less than the relational equivalents that preceded
them.  This is all our clients care about.  Possibly RDBMS would have
acquitted itself better in your capable hands, and that's fine.  In the
end, we like RDF and we have good developers, so we use RDF to write good
applications.  That's really all the justification of RDF that's needed
for me.

Thanks.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python