Bottom-up vs. top-down taxonomy is an old, ongoing KM struggle. But
the emerging architecture of business process automation may help
us cut that Gordian knot. XML documents, produced and consumed by
Web services but also by people running a new generation of
XML-savvy applications, will be the currency of the information
economy. Richly structured, easily captured, and embedded in
well-defined business contexts, they'll be a godsend for tools that
mine knowledge from documents.
Full story at InfoWorld.com
Here's Edwin Khodabakchian's take on InfoPath, an example of the kind of "XML-savvy application" I had in mind:
Infopath is a kind of Blog++: the manipulated data is rich and structured (expense report, travel request, hotel reservation, employee review), meaning that when the data is published back to the server can be processed by an array of services, processes, agents. [ Collaxa's Take]Exactly. Collaboration tools have to move heaven and earth to mine knowledge and infer social networks from email traffic. While it is notionally private, many email exchanges -- "here's the revised version with the changes we discussed" -- are really semi-public in scope. The same holds true for many voice interactions.
Edwin goes on to say:
Tools like Infopath offer developers a great incentive to design their applications as document-centric message exchanges (not RPC) and may be even dis-intermediate WSDL and SOAP!Yes, but we should clarify how and why. Document-centric message exchange is, indeed, the lifeblood of business. The backbone architecture of business process automation will be a SOAP pipeline. But people can -- and in fact must -- intermediate. A goal of InfoPath is to ensure that when they discuss or annotate a document (i.e., a "business token"), the full fidelity of its core schema is preserved. The growing social acceptance of the blogging mode of discourse starts to make it seem more natural to "post" your contributions rather than "send" them. And although we're not there yet, I think this defines a compelling use case for audioblogging as well. A voice conference about a business event is semi-public, subject to the scope of that business event. As Jeremy Allaire has intuited for some time now, we're going to want to be able to initiate, capture and replay (some of) those conversations (and I would add search them) in business-relevant contexts.
Former URL: http://weblog.infoworld.com/udell/2003/03/15.html#a639