Notes from the Don Box, Tim Ewald, and Martin Gudgin session, “XML Web Services Perspective“:
The first part of the presentation flowed around the System.Xml stack. Mostly the basics… XmlReader, XmlWriter, XmlDocument, some XPath, namespaces, XSLT, and schemas. Nothing new, but entertaining. Gudge spoke most of the time, with interjections from Box and Ewald. Ewald did the coding, and used a KVM switch (orr some other switch) to swap between the presentation and the code on 2 separate physical machines. Good tip for future presentations: this worked much smoother than trying to minimze PowerPoint, then maximize it again once done showing code.
After lunch, as expected, we started into SOAP, and Box presented while Gudge ran the keyboard.
“SOAP 1.2 is the last version the industry can handle… If the extensibility of SOAP 1.2 was not sufficient, then what makes us think that a SOAP 1.3 would be any better?“
The next version of the platform will speak 1.2 natively.
SOAP is not a serialized object graph. If you want to treat it as such, it is a lifestyle choice… It has no bearing on the architecture…. it is all just a text file, and that is what we all agree on. That’s the beauty… both can work together just fine.
The portion on header processing for SOAP messages was pretty well done. He explained the 1.1 SOAP actor feature, and the 1.2 SOAP relay feature. If you do not understand the header content, then you ignore it. If you understand the header content and you are the actor, you process the header and remove it before passing on down the wire. If you understand the header content but a SOAP 1.2 message indicates relay, then you simply relay the message intact.
The next section is WSDL.