October 2, 2003

Manifest Destiny?

Fall is my favorite time of year, but I'm amazed how quickly it has stormed into the DC area. Tonight it's going to be in the mid-30s. I had to turn on my heat today! Wasn't it just 90 a couple of weeks ago? I'm waiting for the leaves to start turning. I'll post some pictures when they do.

I have a plan to kickstart Camino development. The main bottleneck is yours truly. For that, I apologize. I know there are a lot of people who want to contribute, and even have contributed, but the culture that smfr and I developed while we were working on it full time doesn't translate at all to the world we're in now. Development has ground to a standstill and people are frustrated. Regardless of why we are in that situation, it's time to move forward and put that lull behind us.

To that end, my main goal is two-fold: to get more people involved and familiar with the codebase, and free me from having to look at every line of every patch. Here's how it will work:


  • For every patch that's submitted for mozilla/camino, two (2) reviews are required. These reviews may come from anyone in the community, but they need to be someone who knows Cocoa. Each reviewer must apply and test the patch and (most importantly) look it over to make sure nothing it wrong with it.
  • Once it gets two reviews, request a super-review from me through bugzilla and I will give it a quick once-over and land it.

The primary benefit is that this frees me from having to commit large chunks of time to each patch. Having two reviewers means twice the opportunity to catch mistakes, and not allow someone to just go get their best friend to rubber-stamp their huge patch. It will also help people get up to speed with the code. As people become more familiar with the code we will relax the rule of two reviews, and some might even be able to get mozilla.org cvs access. That process is cumbersome, so I'm hoping to avoid it in the short term.

Additionally, I still retain the power to veto patches with which I disagree. Camino has reached its current level of success through strong ownership and decisive decision-making. Not every patch is right for this product and I think it would be a shame to lose that oversight.

Patches to other areas of the mozilla tree fall under the mozilla.org rules. I can't control that at all. I think that as we reduce the number of outstanding patches for mozilla/camino, we'll have a better handle on what needs to be driven into mozilla-proper. Also, Mozilla.org is driving towards 1.5 and the tree is locked down tight. Waiting might be easier on all fronts.

I realize that I am still a bottleneck in this new process. However, I'm already an ever bigger one. It is not possible to get worse than it is now. I believe this will get the gears turning again. If two people have tested a patch and agree that the code is good, I will feel much more comfortable and not feel like I have to invest a lot of time with a painstaking review. Instead I can invest it checking in patches, and maybe even fixing some small bugs!

I need the help of the community more than ever right now. It's time to start growing the circle of people who are intimate with the code and its design. The project will fail if we don't all work together.

Who's with me?

Posted by pinkerton at October 2, 2003 10:44 PM