Re: mechanisms for reviews needed

[Jacob]

Dave Korn wrote:

> Why not just _tell_ your engineers that they mustn't check-in unreviewed
> code?  

Sure.

And how will the reviewer(s) get to know about the changes
and where will the changes be found? Inside the VCS would
be quite natural, to me at least.

> If they're so out-of-control that they won't follow a very simple and
> direct request from their employer regarding their working practices, you
> already have a very serious disciplinary problem that I don't think you're
> going to be able to solve with CVS ...

This is not about programmers out-of-control. Everyone who have ever
managed an enterprise software project (or participated in one) will
understand what I mean.

I just don't like 20 (say) programmers to contribute to a repository
and each one of them being themself solely responsible for the
quality of what goes in there. That is simply not good enough.

The prerequisites, guidelines, process, tools and diciplince must be
there in advance of course so that the review process becomes smooth,
but the review _must_ be there. At least in certain phases of a project.