I definitely appreciate the intent of it, and think measuring along that axis is incredibly important (and rare). But as someone who continually struggles more to dive into some codebases than my peers, I can say that this task is far too easy and wouldn't help you gauge my ability to do that at all. Something that requires implementing a new feature that isn't a near-mirror of an existing one (ideally one that requires a bit of thought at the design level) would probably be a much better measure, but I don't know what that might be for something like memcached.