I just finished reviewing a chapter in an upcoming computer science textbook on Web Services. The authors made a heroic effort to give the reader a solid grounding in Web Services including HTTP, SOAP, WSDL, BPEL, WS-TX, WS-CO, UDDI, etc. all in 60 or so pages. In terms of information density, the result was the book equivalent of depleted uranium. To make matters worse many of the specifications they were describing had already changed since the time they wrote the chapter and will surely change even more before they publish. Which got me to thinking about the book I would want to read about Web Services Protocols.
It would focus on theory and only use implementation for examples. The languages/protocols/etc. may change but the theory stays the same. From a theory perspective there is little new in Web Services, perhaps some of the more interesting bits of REST, and from a practical perspective about the only two things that are truly better then what went before are URIs and XML.
So I think a book written about the theory of Web Services Protocols could have a very long shelf life. Such a book should focus both on what Web Services want to be (e.g. Service Oriented Architecture (SOA)/REST as well as what they currently are, RPC.
Here's the outline I would like to see.
-
The Challenges of Internet Messaging
-
Latency, Latency & more Latency
-
Asynchronous Communication
-
Message Routing & Intermediaries
-
Name Resolution
- Distributed Authentication
-
-
A Theory of Resources
-
URIs – The Power of Names
-
Different Strokes for Different Folks – RPCs, Procedures & REST
-
Principles of Web Services Protocol Design
-
The pressing need for loose coupling
-
Standardizing Message Structure
-
- mustUnderstand, schemas, etc.
-
Transport Neutral Messaging
-
-
Configuration and the Tower of Bable
-
Reliable Messaging
-
Signaling
-
Distributed coordination
-
e.g. distributed 2PC compensation based frameworks
-
-
The Age of Exploration – Resource Discovery
-
Bringing it all together – Choreography