Lists Home |
Date Index |
Mike Champion wrote:
> Seriously, and speaking of that post, it's intriguing to me
> how so many people (independently, AFAIK)
> latched onto the basic idea of combining XML and HTTP to move serialized
> objects around and/or do RPC calls.
For a technology to end up a huge interoperability success like email or
the Web, you need two things. First, it has to have all of the
evolvability characteristics. Second, the parts have to integrate
together into some whole that is greater than their sum. Everything I
hear and observe about the SW technologies is that they consider both
halves better even than XML (which sort of stumbled upon its popularity)
and certainly better than the web services technologies which seem to
have essentially no story about integrating independently designed
services except "just write a program that does it" and/or "we can sell
you a product that does that."
> The semantic web seems much the opposite: an idea that COULD work
> and MAY work someday (perhaps when they "boil the ocean" and persuade
> lots of people to take metadata and ontologies seriously), but
> doesn't seem to have nearly as much success in the wild today.
The growth curve (and # of reported success stories) of the SW
technologies is much faster than SGML at a similar stage in its development.
>... I saw SGML *really work* for real people with hard problems
> that Word, TeX, etc. choked on when I started work for Arbortext
> in 1996.
Don't you think that if you worked at intellidimension or semaview or
FourThought you would see the same thing around RDF today?
In fact, the customers are many of the same big SGML consumers. I'd
better not divulge what I know, but to use public information, think
about the the employers of the two people who wrote the article that
kicked off the thread!
> There's some Semantic Web stuff I find very intriguing, e.g. applying
> WebOnt to SNOMED, or the Canegie-Mellon stuff on scheduling
> (http://www.sys-con.com/xml/article.cfm?id=529). Then again, they're
> applying RDF to well-understood, localized domains.
No doctor would say that medicine is well-understood.
> Sortof like SOAP-RPC -- a reasonable solution
> to a limited domain problems, but my hackles go up go up when
> people start talking about it as the Next Big Thing. I guess it really
> *is* the hype that bothers me ... and there's something about hype
> that just gets under my skin and makes me want to debunk it rather
> than doing something more productive with my time ...
There is ten times as much hype around SOAP RPC. And based on my
scratching of what's under the covers of both technologies, ...