I happen to have worked on a nice framework to build WS-* services.  I happen to have worked on a nice framework to build REST services.  It turns out the same framework can be used for NET-* services too.  It also turns out the same framework can be used for MSMQ services.

Actually now that I think about it, the framework I worked on can be used to build just about any kind of service you can dream up -- even whatever REST++ is called (trust me, there will be a REST++).  That wasn't an accident, it was the plan. 

We can debate the relative value of all of the above approaches, but the only thing I care about is the ability to communicate data between potentially heterogeneous parties in whatever way possible.  WCF does that, so I am happy. 

That said and at the risk continuing to be labeled a WS-* fanboy, there is a lot of value in what was done in WS-*.

There is the business value.  Think about the world in 1998-2000 (when all this got started).  The conversation used to be about what Java application server you were going to use.  Now it is about how you are going to build your REST or WS-* (SOA) services (implementation details are somewhat interesting, but not the key point anymore).  The Web did a lot of this, but WS-* helped change that whole conversation, especially in the enterprise.

There is the technical value.  Yes, there is some.  The key thing for me is writing down the 'model' for the service.  What are the messages, what are the transmission requirements, etc.  With these 'models' in place, you can do some really interesting things (WCF does somethings here, but we are looking at taking this to a whole new level in some of the other work we are doing). 

Sure, there are things I wish I could have changed.  I do wish we had seen the importance of GET earlier.  I do wish we had seen that pure binary payloads without the Infoset are just fine.  I do wish we had had simpler formats to write down the 'models' (XSD & WSDL (and maybe even XML) just don't cut it).  We have already addressed the first two in WCF and we may have something in mind around the last.

That all said, I think we have delivered significant value to both the industry and our enterprise customers in the form of WS-*.  More importantly, we have well served all of our customers, including those using REST, WS-*, and even customers still using DCOM, with WCF.

I think this may be my last post on this whole WS-*/REST thing.  If you just use WCF, everything is fine.  If you can't use WCF and you are in a enterprise, you should consider using some product that supports WS-*, especially if federated security, transactions, reliable messaging are important.  If you can't use WCF and you are in the Internet, you should consider using some product that has a nice HTTP stack (nee REST).

As for me, I am clearly working on something else.

WS-* is to REST as Theory is to Practice