Correct code/unit testing (was Re: Efficient python programming...)

[Kragen Sitaker]

Peter Hansen <peter at engcorp.com> writes:
> Kragen Sitaker wrote:
> > Of course!  People are fallible, it's much easier to do a bad code
> > review than a good one.
> 
> I have to confess I have never seen a successful code review, nor
> an environment in which they were used regularly and which had a
> high level of code quality.  Clearly as a result I'm fairly biased
> against them -- I should keep a more open mind on the issue perhaps.

You are experiencing code reviewing whenever you are writing
production code.  Do you find that it improves the code quality?

I've found that both pair programming and less-intense kinds of
reviewing are usually extremely helpful.

> I'm swayed also by the thought that programmers generally rebel
> against them, and that therefore they are likely not the best
> approach in most shops.

Yeah, that can be a problem.  That problem has made it hard to
introduce pair programming in some places I've worked; it largely
depends on the social dynamics of the place.

> Assuming the formal 'inspection' advocated by some heavy-weight
> processes, I think it's of little use on code pair-programmed.  (We
> could also be more general and just say that pair programming is on
> par with inspection and that either is effective in an appropriate
> context.)

I mostly agree that inspection during pair programming finds the bugs
any inspection is likely to find.