[Termtools] complexity analysis of integer (...) programs, preprocessors

[Johannes Waldmann]

Dear colleagues.


For 2016, the Termination and Complexity Competition has new categories

> "complexity analysis of integer programs (koat format)"
> "complexity analysis of integer C programs"

for which we welcome several new participants (and some old-hands).


As an organizer, I would appreciate very much your cooperation
in the following:

* provide benchmark sets
  (put them on starexec, tell me their space ids, cf.
http://www.termination-portal.org/wiki/Termination_Competition_2016_Registration#Benchmarks
)

* put the proposed specification for syntax and semantics
  in some canonical and public place, and announce it here
  (termtools list) Also, further discussion should happen here.


At least one participant asked me about how to use preprocessors,
to convert benchmarks between different formats.

Just to get a clear vision of what the proposed change is,
and the amount of extra work that it would bring:


1. What preprocessors are now needed, and by whom.

My current data model is this:
https://github.com/jwaldmann/star-exec-presenter/blob/master/Presenter/Registration/Code.hs#L113
(Each category has a postprocessor, there are no preprocessors).

Do you think of preprocessor per category, or per participant?

Can you fork my code and send me a pull request...
But even then, see next item.


2. Preprocessors have to be handled by starexec "community leaders"
(for Termination, that's me)
https://wiki.uiowa.edu/display/stardev/User+Guide#UserGuide-Pre-Processors
Is your preprocessor a stable piece of software than can be
installed as-is? I don't have time nor knowledge to test your
preprocessors, much less to  do re-installs repeatedly.


I would very much favour a solution where each participant
handles all his processing, including preprocessing, on their own.
That's just one extra line in your starexec_run_default,
so I really don't see the problem.


Best regards, Johannes.