I’ve spent some time looking over the WS-Reliability and WS-ReliableMessaging specs recently and reading what people have to say about them. Dave Chappell seems to be the most outspoken person on the Oasis TC and seems to genuinely care about pulling all the different folks together on the topic. See his article about the current state of affairs which I think is a decent summary.
I think it is interesting how WS-Reliability was rushed to committee as opposed to addressing technical shortcomings in the initial spec. I do think WS-ReliableMessaging was generally more thought through (and hey, I’ll admit my biases right now so take it for what it is worth). I think most people would agree with me.
So was the quick creation of a committee for an incomplete spec a power-play? I doubt it was a non-issue, but I might also expect that the authors found themselves in a naturally-occurring bureaucracy during authorship that may have mimicked what the TC committee would have looked like. Therefore I assume they decided to just publish and go to committee. I also think there was a political statement being made about MS et al working on specs without the involvement of a bigger community.
At issue is whether to create a fleshed out spec first, then take it to committee or whether you do the fleshing out in committee.
There is a lot to be said for doing the majority of spec-creation work in non-committee so that bureaucratic problems are minimized and things can progress. Getting 95% of the work done, and then going to committee to work out the problems of specific scenarios seems like a pretty good way for the process to work. I would say that the WS-Reliability spec is only about at 80%. In contrast I think WS-ReliableMessaging is closer to 90%. Of course it means you have to have the right people writing the initial specs—and let me say right here that Microsoft and the others involved have a lot of really smart people spending a lot of time on these things. I think its safe to say that some non-MS folks are feeling left out. They will get their say when things go to standards bodies (and earlier if they want to), but I’m not sure if that is considered to be enough or soon enough by those that.
It should be very interesting to see how Web service reliable messaging will flush out, but I’m confident that it will and that the industry will have a single spec to deal with. It will also be interesting to see if reliable messaging will flush out faster or slower than similar areas of the WS spec world.