The SOAP envelope is as far as the API and protocol standards will get you, but it's what's inside the envelope that counts. Most of the world does not use ebXML or RosettaNet as their data dictionary...Whether one considers business-to-business integration, EDI, CORBA, EAI tools, or data integrationas the true integration technology, none of these solves the problem of incompatible business semantics...Interoperability is about maintaining the autonomy of the partisipating members of an exchange community...we should remain level headed about how far connectivity and message routing software can get us.[eAI Journal via Paul Kulchenko]Paul's take is that "documents should be the primary focus and they way you send bytes and bits is of secondary importance". I'm less certain. This article focuses on the work being done on the semantic web, and simultaneously bashes the idea of standard vocabularies. I have to agree that standards efforts have a pitiful track record with respect to ROI. But it seems to me that the call to focus on documents inevitably leads to a call to establish another standard. I don't like this, but I don't have a better answer. The semantic web may be the answer, but I don't think that it addresses the immediate need. I should probably be reading up on this, but I've always been scared away because the Semantic Web has always smelled too much like vapor, for my taste.
BTW, apologies on the atrocious look of the site. I was experimenting with new templates, and didn't realize that Radio had published them. Unfortunately, I can't fix them until I get home, so we'll just have to endure for now.
[Gordon Weakliem's Radio Weblog]
Interesting paper and comments from Paul and Gordon. I agree with both sides to some extent. Obviously the author of the EAI paper is going to tout Semantic Web technologies, working for a Semantic middleware tools company :-).
I agree with Gordon that the Semantic Web still appears to be a bit of vapourware. It seems like a technology solution looking for a problem or just another attempt to make AI take off yet again by riding the Web and Web Services hype.
For me the biggest value of the whole Web Services movement has little to do with SOAP encoding. Its more to do with providing connectivity between different systems across middleware, platform and corporate boundaries, using XML for message formats and providing support for both synchronous RPCs and asynchronous messaging. This fact alone is far more important than standard schemas or the Semantic Web. It provides a level playing field on which tools, software and companies can operate, using XML to communicate self-describing data.
The exact schema or encoding of the XML and whether you have a semantic model of your business is much less important IMHO.