I was pleasantly surprised to see your discussion of REST in the Sample Chapter: Basic SOA Using REST of your book SOA Using Java� Web Services (amazon US). Do you think that there is any value in considering a REST based SOA, especially considering the prevalence of SOAP-oriented tools and the fact that most of the advanced web services specifications assume the presence of a non-RESTFUL SOAP infrastructure? [ May 29, 2007: Message edited by: Peer Reynders ]
My book is meant to be practical, and REST is practical. Many people find it much simpler and more intuitive than SOAP.
So, yes, I think that there is tremendous value to including REST support in your SOA.
Chapter 3 in my book describes one approach to REST using Java Web Services. In addition, there is a new API being developed - The Java API for RESTful Web Services (JSR-311) - that will make it even easier to deploy RESTful services using Java. I'm on the expert group for JSR-311 and hope to be blogging about it (with code examples) soon.
The reasons that most tools are SOAP-oriented, rather than REST-oriented, are: (1) SOAP came first. (2) There is no WSDL for REST.
Now that REST is clearly very popular, tools will start including more support for it. But (2) is a real problem. Without a standard interface definition language (like WSDL for SOAP), it is difficult for tools to process REST endpoints. A solution, the Web Application Description Language (WADL) is in the works to address this deficiency.
It's good to hear that there may be hope for REST yet and that it's architectural principles may be codified in some specifications that will make it easier to pin down for the tool vendors - giving it a fighting chance to become a viable alternative against the plethora of wizard driven SOAP tools.