[Baypiggies] A talk on testing frameworks

[Minesh B. Amin]

Hi Seth,

> So, i'm curious if anyone has experience doing stuff like this, sort
> of applying the "fit" model of automated testing, to big scale testing
> challenges?  

I am not sure about the "fit" model of automated testing.

However, if you accept the premise that the DAG for a suite of tests has
a lot of attributes in common with the DAG for a typical build system,
check out an approach first described by Prof. D. J. Bernstein
( http://cr.yp.to/redo.html ). 

Avery Pennarun recently implemented this build system in Python in under
a month ( http://apenwarr.ca/log/?m=201012#14 ) !

I would classify solutions in this space as:
  make             -> generation 1.0, 
  CMake, SCons     -> generation 2.0,
  Avery's approach -> generation 3.0

Notice that with each generation, less and less is expected from the
developer ... while more and more is expected from the "system" :)

Cheers!

Minesh