Simon Fell > Its just code > May 2003
Geez, this is starting to sounds like the Monty Python skit. Why is it so hard to make it clear? I've heard two different answers from five different WG members.The discussion contines to go on, with more different answers. One thing (about the only thing so far) that is clear is that the membership of the WSDL-WG don't fully understand the decision, its implications, and how rpc/encoded is to be handled in the new world, I really hope they can get together and resolve exactly what they are doing any why, because if they can't clearly explain it to a bunch of people who write web service toolkits, they stand no chance trying to explain to web services users at large.
In related news Jeffrey Schlimmer responds to my devils advocate question. I don't know if that'll be enough. As I work through adding SOAP 1.2 to PocketSOAP I find myself asking the same question about SOAP 1.2
update IBuySpy is still labeled as demonstrating best practices, ouch don't think so. An interesting demo yes, best practices, don't see any. YMMV.
- Doug Kaye's Loosely Coupled. I thought Doug did a fantastic job of explaining why you should care about web services.
- Revelation Space. Started off well, but seemed to come off the rails towards the end.
- Altered Carbon. I loved this, couldn't put it down.
- The Stone Canal. I seem to be reading Ken Macleod's books in the wrong order, but they've all been great.
As many people have pointed out in the past, XML is a lousy wrapper for XML documents (conflicting encodings, DTD's, ID clashes etc). Given the continued trend to push for SOAP to be a transport of XML messages (aka doc/lit), and its inability to transport arbitrary XML documents (see above, plus various restrictions SOAP places on the XML structure) I've been thinking whether trying to map the whole shooting match into a single xml document is the right approach. In the end this thread on xml-dist-apps pushes me over the edge, and it looks to me that managing the headers and body as two independent xml docs would be a much better solution. Obviously I'm about 4 years too late to this realization for SOAP, but I'd be interested to know if at any time in SOAP's history this was considered.