[ndonnellan]

Caution though: this can lead to blind spots. I witnessed a critical bug become exacerbated by the first fix because the developer and reviewer had essentially co-wrote the solution, so the sign-off was essentially a rubber stamp.

If you want the least bias, you should find someone to review something that hasn't been deeply involved in designing it. And for pieces of code that are small enough, reading and getting up to speed shouldn't take that much time. I prefer sending reviews to engineers that are familiar with the area of the code, but we don't do pair-programming so the implementation hasn't been seen before.

We do, however, do a lot of design work before a single line of code is written.

*also this is at google.

[jeffbee]

Yep, "pocket reviewer" generally has a negative connotation. But, let's face it, most changes are minor. You have to be judicious about the level of review required for a given changelist.