[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: The Roadkilll of Ximplifications (Was RE: ZDNet Schema article,and hiding complexity within user-friendlyproducts)
- From: Tony Presti <firstname.lastname@example.org>
- To: "Bullard, Claude L (Len)" <email@example.com>,xml-dev <firstname.lastname@example.org>
- Date: Mon, 23 Apr 2001 22:04:28 -0500
Let me clarify:
What businesses need are the tools and approaches developed in the attempts to
apply semantics to the web. The reality of real world applications
development requires the deployment of constraining mechanisms at one or more
levels to accomplish specific solutions to specific problems with information
that businesses will pay for. This of course includes approaches and tools.
Already, as I am sure you know, we have several semantic approaches and tools
that are applicable. These include DAML, RDF, Description Logic, Decision
Logic, and Ontologies, to name a few - even XML Schema, within limited domain
application architectures. There are even experiments into handling hunderds
of thousands of intelligent agents for deterministic responses in battlespace
scenarios by Jim Hendler's group at DARPA. This can border on chaotic, but we
will learn from it.
So, the search for semantic approaches will continue, spawning various
standards and tools along the way - some of which will even be useful in the
long run - some complex, some less so. This is why we as developers need to
stay informed to make proper architectural decisions for our applications.
Unbound knowledge assembly using some form of semantics over the web may or
may not happen, but we will learn from it. I certainly use some of these
approaches now, regardless of standards, unless specified by application
requirements. As you pointed out, the end user is concerned with
functionality, not software development. I will add the customer is concerned
with Return On Investment, not approaches.
You may have determined by now my comment concerning unbound knowledge
assembly using the semantic web was factitious.
Regards, Tony Presti
"Bullard, Claude L (Len)" wrote:
> Ok, but some curmudgeonly thoughts ...
> Keiretsu requires locally stable vocabularies,
> but only that of necessity. XML Schemas work
> fine at the level of exchanging data objects.
> What do you need to exchange and can you afford
> Do businesses need the SemanticWeb? Do we?
> Standards don't always incubate; sometimes they block.
> In other cases, without a core data standard,
> there is no market not because the data is complex,
> but the business rules are and they vary by locale
> and contract. No one can afford to build the app
> unless they can customize it and sell it to different
> customers in the same market for dollar costs that
> exceed what the perceptions for web-based apps sustain.
> Complexity isn't the enemy; the perception of cost
> and its impact on market is. When the Dogs of Marketing
> sell a customer on the WebAppsAreCheaper bit, the
> DevelopmentSlaves groan at their oars. It ain't so.
> I would hope XSD and RDF work together at least as
> far as intertransformability confers identity or similarity.
> As for the semantic web, I reserve some doubts. Grand visions
> are good for rallying effort, creating orgs to direct, etc...
> o MSXML 4.0 is on the street with XSD support. XSD wins
> the first round of practical ubiquitous support.
> o XML Schemas are already backed up with practice in XDR,
> publicly available documents explaining best practices, etc.
> Books are already in the hoppers if not on the shelves.
> o XSLT is a heckuva harder to apply than XML Schemas and
> so far, people are figuring it out.
> o Feedback from serious minds on the notion of the semantic
> web suggests the infrastructure won't be there for ten years, maybe 20.
> We'd have to debate infrastructure but if that is a given, by
> the time that passes, the situation with the semantic
> web will be as it was for SGML and XML: when the requirements
> and the technology converge, it is time for new faces, new
> names, freshly scrubbed ideas, and so forth. So it might
> happen but not for the current inventors/investors.
> o Metadata may be more expensive to create and maintain than
> content. Hired a semiologist or ontologist lately?
> o Content has to last long enough to recoup costs and profit.
> This means a lifecycle that includes maintenance and
> upgrade. Sustainment: we have to be able to afford to own
> content. That is why the kudzu of HTML will never go away.
> o Vocabularies imposed from above are not as stable as
> those emergent from contract requirements. Money talks
> but better than that, it schedules. Companies don't
> join the United Nations. They sell systems to other
> companies. Sometimes, they sell them to the UN members,
> but that market isn't as big as it used to be. The local
> surfer doesn't buy SWs: he subscribes to services. What
> will be the cost of maintaining semantically-aligned services
> and who will pay it? So far, the ISPs, the telcos, etc. and
> they have to pass that on to the customer who benefits by having
> a cell phone that can tell a microwave to ask the refrigerator
> about the best temperature to cook a pizza?
> So the basic semantic web application is a well-done pizza?
> Huh? They used to be $10 and were delivered hot in thirty
> minutes or less. Now because a cellphone doesn't have the
> power to validate a form entry and I can't hit the chiclets,
> I get pizza that's soggy and the refrigerator won't take it back
> or even give me a frikkin' coupon!! And for this, all of the
> best minds in computer science slaved to build the SW!!!!! Some deal...
> Don't Bogart the future. What may appear to be one system lasts
> as long as it takes the weather to change the temperature
> of the road surface. Then the illusion fades in the rear
> view mirror among the road kill of never-ending Ximplifications.
> Ekam sat.h, Vipraah bahudhaa vadanti.
> Daamyata. Datta. Dayadhvam.h
> -----Original Message-----
> From: Tony Presti [mailto:email@example.com]
> I remember those G3 and G4 spec issues when attempting imagery integration
> Xray film....
> And you are correct, Len - the issue IS XML approach with RDF schema. XML
> associated schemata approaches can only work within limited domain
> for semantic applications. The addition of RDF schemata to XML apps seems
> hold much more promise for reduction of computational intensity in
> or (dare I say it?) unbounded semantic web knowledge assembly.