Punf 8.51 parallel

  - The changes from the previous release are too many to
  list. Only the major new features are listed below.

  - The -M option allows to construct the merged process
  (experimental feature). It is saved in the *.mp format
  (description is available).

  - A new total adequate order is available with the -@4
  option. It works best for merged processes.

  - The implementation of the concurrency relation has been
  improved, and so is the performance, especially in the
  parallel mode.

  - The implementation of the preset trees computation has
  been improved.

  - The -# option allows to use non-local configuration as
  cut-off correspondents; this usually reduces the size of
  the prefix but requires more computation time.

  - The unfolding prefix generation for STGs has changed so
  that the signals enabled (perhaps, via a chain of dummies)
  by any configuration can be correctly computed, even if
  some of the cut-offs are dummies. This increases the size
  of the prefix. This behaviour can be disabled by the -d
  option.

  - The -r option allows to replicate Petri net places with
  multiple self-loops. This transformation can significantly
  reduce the prefix size.

Punf 6.03 sequential
  -  A small bug in the concurrency relation has been fixed.
  Strangely  enough,  it didn't show up on 100+ benchmarks I
  used  to  test  version 6.01. Thanks go to Claus Schroeter
  for sending a net which revealed this bug.

  - A lot of code shuffling.

  -  The  -c  option allowing not to store cut-off events is
  added.

  -  The  -@  option  allowing to switch between McMillan's,
  size-lexicographical and Esparza/Roemer/Vogler's orders on
  configurations is added.

  -  Punf  now tries to identify (using a simple sufficiency
  test)  the  nodes in the net which are obviously dead, and
  removes  them before building preset trees and starting to
  unfold.

  -  Punf  now  uses  a  better structure for storing preset
  trees.

Punf 6.01 sequential
  - A lot of minor improvements and fixes.

  -  The  -! option switching off the optimization of preset
  trees   is   added   (cheap   optimisation   is  performed
  anyway).

  -  Support  for  concurrency  relation  is added. Punf can
  delete  it  and  switch  to  a  slower  algorithm  after a
  specified  number of events has been generated (controlled
  by the -n option).

  -  Support  for  STGs  is  added  with a cut-off criterion
  taking  into account  not just final markings but also the
  final codes.

  -  Support  for  high-level  Petri  nets (M-nets) is added
  (experimental feature).

Punf 5.10 sequential
  - The first official release.