[Termtools] complexity competition

[Georg]

Dear All,

as  it was  brought  to  your attention  by  Johannes, the  complexity
category has  finished.  As  this category was  present for  the first
time, perhaps a very short report  is in order (a more detailed report
will be part of the usual report on the competition).

In  the  subcategories  "derivational  complexity"  and  "derivational
complexity - innermost"  CaT and TCT ran a  competition, which was won
by CaT.   I suppose I can  speak for the  whole TCT team, when  I take
this  opportunity  to congratulate  the  CaT  developers Martin  Korp,
Christian  Sternagel  and  Harald  Zankl  to  this  victory.   In  the
subcategories   "runtime  complexity"   and   "runtime  complexity   -
innermost" TCT ran as a demonstration.

A perhaps remarkable point of this new category is that the contesting
tools  published their  source  under  the lesser  GPL  and that  this
category  was  the first  to  use  a  scoring function,  which  worked
surprisingly well. These ideas may be of interest  in other categories
as well.

Unfortunately the analysis of  the results revealed some problems with
the chosen  testbeds. In all  four subcategories the testbeds  used in
the competition are not conform to the description on the wiki page:

http://termination-portal.org/wiki/Complexity

This  was due to  a combination  miss-communication between  Simon and
me and some  technical mishap. The SC was  informed about this and
agreed  that the  competition results  should be  changed accordingly.
Simon agreed to do so, as soon as the competition is finished.

Note  that this  will  only  change the  absolute  numbers of  points,
yes-instances,  etc,  but  will  not  influence  the  outcome  of  the
competition.

best wishes,
Georg.