Lists Home |
Date Index |
Bill de hÓra wrote:
> One of the things that helped kill the expert system industry was
> the (multi-dimensional) cost of extracting knowledge from experts.
> Another thing was that an expert group is not usually an internally
> consistent body of knowledge. If part of the work building reusable
> XML vocabularies for commerce is extracting domain specific
> knowledge, then it's going to be tough to realise a return.
The problems are real, but there are some healthy forces at work too,
compared to previous generations of application interconnection:
the lower the cost of implementing an interface for any transaction, the
more rapidly it can evolve solutions to its weaknesses. I'd argue that
WSDL and XML Schema significently reduce the cost of several
traditionally expensive project components (the hype, of course, will
overlook all the other expensive components).
Impress *or* Imprimatur - UD/EDI standards are being developed for Web
Services, but if these prove less viable than emergent alternatives, Web
Services as a technology will still be gaining critical mass.
technical flexibility - an EDI guy told me one of the problems with EDI
is that it was so highly specified that in order to adapt it to any
particular use you would have to (ab)use existing but un-required fields
for totally different purposes, by silent bi-lateral agreement. Maybe
for this reason, XML Schema has pretty good support for clean schema
customisation - for instance you can permit an element to contain
attributes and/or elements from specific or any other namespaces.
I suspect that the eventual winners will include a mix of surprise
evolvers (possibly dominating emerging activities) with the
consortia-blessed (probably for the more mature verticals).
"Weeds are not supposed to grow / yet by degrees
some achieve a flower although / no one sees." - Larkin.