Not only a good read, it's a useful thing to remember when you're trying to encourage people to give away their time. They have to get some reward (social, seeing things actually happening, whatever) and have their hassles reduced if you want to maximise how much they will give.

I suppose I should ask: how easy is it to get code into LJ these days?

It's not such a fair question, I know, because LJ is a large, high-availability prodution project many would consider critical (regardless of the merits of them having that view:))

A lot of people regularly get their code in, and avva and mart even have commit access, since I totally trust them. So they commit stuff, either their own, or from other people once it's been reviewed. It's not ideal, but it's safe. If others want to get commit access and have a good history, I'm happy granting it.

I really want to move to svk soon.

Let me know if you want a pugs committer bit :-)

In all my free time? :-)

P.S. You're crazy. I read all about your pugs test suite heroics in Autrijus' journal, Google groups, weekly P6 summaries.

We're rolling out 6.2.10 now! Join the channel and the fun, you know you want to.

Sounds as though you're doing fairly well.

That reads an awful lot like a "worse-is-better" case study. It's been proven the more effective approach, but it's not like it doesn't have it's drawbacks (not that I'm advocating Pugs be developed like Lisp).