David,
I don't have write or comment permission on that document, so I'll give you my feedback here.
From the Servlet Spec point of view, the process felt very chaotic and out of control. There are too many committers on the project, some of whom appear to either have more experience with the process, inner workings and/or out of band information. The two spec leads felt very out of control as PRs were often submitted by one committee, reviewed by another and then merged in short order without any opportunity for wider discussion or review. The same happened with releases candidates that were built and promoted also without any input from the spec leads!
All this was done in good faith by well meaning individuals, who for the most part did the right thing.... but that approach does not encourage community, communication, consensus. I know everything was a bit rushed and having busy spec leads being swamped by process probably was a bit of a hand brake. But if the spec leads don't know what should be done, educating them is better than bypassing them. I'm not saying that the spec leads opinions are the only one that count, but they should be responsible for communicating to the expert group and getting a rough consensus.
For this release, it didn't really matter that much because we were not changing anything technical. But I fear that if the same is repeated for the next iteration, then the chaos is likely to be damaging to quality and consensus. I would very much like to see a change in process where only spec leads can merge PRs or promote releases.
cheers
