How to reproduce the bench for "Expressing Polymorphic Types in a Many-Sorted Language" of FROCOS'11

The result of this bench can be reproduce using Why3 in the following way:

- You need to come back to version 69dd88e5c34808de3905aa3e257288d19e3b643c with
  git checkout 69dd88e5c34808de3905aa3e257288d19e3b643c, and
- You need to configure with --enable-local and compile why3.
- After that you should set in ./why.conf in the
  section [main] the variable "running_provers_max" to your own
  number of core. You need CVC3, Z3 and Yices.
- Run ./gen_allbench.sh generates one file, bench, by possible set of parameters
- Run ./run_bench.sh  (it takes 10 hours with 7 core). It runs the benchs used in the paper
- Run ./compute_result.sh (it takes 10seconds) to compute the comparison tables. It needs csvtool.

nb: bool_inf and unit_inf are used in order to be able to use explicit