[Termtools] [color] Re: Results for certified categories.

[Frederic Blanqui]

Dear all, I think that this is urgent now to update the competition 
webpages or, at least, add comments at the beginning of it.

- for SRSs, on 
http://colo5-c703.uibk.ac.at:8080/termcomp/competition/competitionResults.seam?category=10236&competitionId=15991&cid=16800,
cime3 still appear with the best score...

- for TRSs, on
http://colo5-c703.uibk.ac.at:8080/termcomp/competition/competitionResults.seam?category=10235&competitionId=15991&cid=16800,
computed scores are wrong as noted by Adam.

Then, what to think about papers referencing to the results of the 
competition webpage if those results are wrong?...

Frederic.

Adam Koprowski a écrit :
>       To begin with the score of AProVE-cert in the TRS category says
>     594 (it's just an example, I believe this problem concerns all the
>     tools). However after downloading all the results and grepping
>     through them I could only find 520 successful proofs. I counted 74
>     certification timeout in the table, so it seems that the systems
>     where certification timed-out are still counted as successful
>     proofs. This seems to be a problem on the part of the termination
>     competition platform.
> 
>   The interesting consequence is that, contrary to what the results page 
> says, actually AProVE+CoLoR, and not AProVE-cert, is the winner in the 
> certified TRS category with 558 verified certificates out of 580 
> certificate candidates. Those figures are 520 and 594, respectively for 
> AProVE-cert (=AProVE+A3PAT+CoLoR). But perhaps again it's news only to 
> me :).
>   Best wishes,
>    Adam
> 
> -- 
> =====================================================
> Adam.Koprowski at gmail.com <mailto:Adam.Koprowski at gmail.com>, 
> http://www.cs.ru.nl/~Adam.Koprowski
> The difference between impossible and possible
> lies in determination (Tommy Lasorda)
> =====================================================