Code reviews are overrated

So, bad methods of code review are bad.

To be fair, it helps junior developers much more than senior developers, but I remember one team of good mid-level developers that did peer review for every change, and QA had to work hard to find bugs because developers caught more corner cases because multiple developers were looking at the code. Further, at least two people knew all portions of code, which was a lifesaver when someone left. Finally, it helped developers get better because we learned what others were looking for in their own code.

As the most experience with shell scripting on my team, I'm asked to review most changes right now. I do them on down time, and I've noticed that their scripts are getting better because they know what to look for. Further, mine are getting better because they're using different methods than I learned.

Meetings are overrated, but code reviews are quite useful. If you use a tool like Code Collaborator to asynchronously do code review (rather than pulling everyone into a synchronous meeting) you get something close to an optimal solution.

They're pretty nice using Guido van Rossum's rietveld: http://code.google.com/p/rietveld/