On 9/4/09, Sheldon Newhouse <sen1 at math.msu.edu> wrote:
> So, this is a suggestion that people take some time away from
> development to document the files which already exist. The
> documentation could even be in the source files themselves.
> It would be much healthier for developers so inclined to add descriptive
> comments on routines "as they do the development."
Well, I don't know what we would do differently.
New code written in the current development cycle (since 2001
let's say) generally has more comments, document written at the
same time as the code, and test cases. So we can continue that.
On the other hand, the code from Project MAC in many cases
is quite obscure, and it would benefit from documentation of some
kind, but one cannot write the documentation until one understands
the code. Slowly we are adding comments (pretty much only when
some old code needs to be revised). That seems like the best
we can do.
best
Robert Dodier