On Wed, Jun 25, 2014 at 2:39 AM, Menno Smits <menno.sm...@canonical.com> wrote: > I completely agree with Ian's point about code needing to be > self-explanatory and stand on it's own. > > That said, the article mentions that the process of creating review > annotations encourages the author to review their own work in a way that > they may have not done otherwise, eliminating problems before anyone else > even looks at the code. Effectively, a phase of self review happens before > peer review driving the defect rate down and probably making the peer review > more efficient.
I've found this to be an effective part of my personal workflow, usually stepping away from the code (even overnight) before doing a self review. Doing it in the review tool sometimes makes sense but in most cases I'll just run through a prettified diff locally and make XXX comments in the respective files. However, review annotations are most useful when you notice something *after* you've published the review request. They're a good way to get focused feedback in that situation. -eric -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev