I just finished doing an immersion in the ebXML specifications for an article I've been working on, and I've got just one thing to ask: what the hell were they smoking?
This is supposed to be a set of standards that will allow anyone on the planet to conduct electronic business with anyone else, using everything from batch FTP transfers to web services to push electronic documents over the wire to each other.
But I don't see anybody ever implementing all of the parts of ebXML–and there are a lot of parts. And implementing even some of the parts in their current form requires grafting on pieces of other platforms to get something that comes close to working.
If I was a cynic (and I am, by most accounts), I'd say something like, “What do you expect from a technical standard first defined by the United Nations?”
Then on Thursday, it came out that IBM was asserting its intellectual property rights to one of the major pieces of ebXML, the Collaboration Protocol Profiling and Agreement Specification–the part that is supposed to negotiate how two parties connect.
Rumor has it that IBM has since said they won't enforce the patent. But who's to say they won't change their mind when an opportunity to exhert competitive leverage presents itself?
There's only one real solution to this situation–keeping it simple, and sticking to a standards-based point-to-point approach. Rather than working toward totally automating the process of lining up with another busines online, it would be much easier to connect to processes as they were needed, using a simply-defined interface? Like maybe WSDL or something? I mean, CPP/A and ebXML's registry services specs make WSDL look like a walk in the park.