Teaching materials for scientists

[Matthieu Moy]

Greg Ward <greg at gerg.ca> writes:

> I seem to have volunteered to teach a bunch of physics grad students
> and postdocs the basics of version control with Mercurial.

I teach the basics of version control (with Git, but at this level it
doesn't really make a real difference).

One difficulty is to convince students that version control is a good
thing. They normally work by sending files by email or using USB keys,
and do not necessarily realize how wrong this is :-(. Then, a common
mistake for them is:

1) encounter conflict, and decide that conflicts are scary
2) tell the teacher "look, your tool doesn't work"
3) exchange files by email or USB key for a while
4) teacher tells the student to commit/push (e.g. "I'm grading you based
   on what is in your repository")
5) student tries to merge, nothing works
6) student tells the teacher "see, I told you 2), it doesn't work".

So, it's very important to insist in fixing problems/missunderstandings
as soon as they occur, to reduce or eliminate step 3).

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/