Brief article on benchmarks of Python repository with leading DVCSen

[Stephen J. Turnbull]

Matthieu Moy writes:

 > Try it: open a file in your text editor, add a line in the middle,
 > save, run "$vcs diff". It will show

I'm confused.  I don't understand what this example has to do with
"change detection" in the context of a VCS's capabilities.  In my
understanding "change *detection*" is what "$vcs status" does.  Ie, it
tells you that changes are present, not what they are.

 > And no, diff algorithms are not 100% reliable, it does happen that
 > the diff doesn't actually show your intention.

But Talden's point is that in the context of detecting name changes,
the diff is uninterpretable without substantial knowledge of language
semantics that $vcs does not have.  *But Eclipse's refactoring tools
do have that knowledge.*  So to me it makes great sense to have that
information recorded and used by the VCS, but most importantly made
available to other tools that understand it.  And I can easily accept
Talden's testimony that it's a deal-breaker for him if $vcs doesn't
support it.