[sat4j-dev] SAT4J 2.0 news

[Daniel Le Berre <leberre@xxxxxxxxxxxxxxxxxxx>]

Dear SAT4J users,

Here are some news regarding the status of the upcoming SAT4J 2.0  
release.

*SAT4J 2.0 will ship with Eclipse 3.4*

The Eclipse Foundation decided to include SAT4J in the next release of  
eclipse to power their
new generation Update Manager (codename Equinox P2).

There will be a few consequences for SAT4J users:
- SAT4J 2.0 will ship under the Eclipse Public License, and no longer  
under the GNU LGPL license.
Since the EPL is strictly freer than the GNU LGPL, it should not be a  
problem for most users.

- SAT4J 2.0 will have to run on 1.4 JVM, and if possible foundation VM  
(for embedded systems) : to achieve this, many dependencies to 1.5 or  
1.6 libraries are currently been removed:
  + removed use of enum: implemented the enum design pattern directly.
  + removed use of the Iterable interface.
  + removed calls to Integer.valueOf(int) method allowing to pool  
small integer values.
  + removed use of StringBuidler => used StringBuffer instead.
  + the use of the Scanner class will also be removed in the near  
future.

- SAT4J will ship as several bundles:
  + sat4j core for embedding sat solvers in Java apps
  + sat4j pb will add Pseudo Boolean capabilities to sat4j core
  + sat4j maxsat will add MAXSAT capabilities to sat4j pb
  + sat4j csp will add CSP solving capabilities to sat4j core
  + sat4j sat will be the frontend dedicated to experiments
  Since Eclipse is OSGi based, SAT4J will provide OSGi compatible  
bundles.
  Self content jar files (as the ones available in 2.0 RC1) will also  
be available for the users.

-More rigorous contribution management
One of the consequence of having SAT4J code distributed by the Eclipse  
foundation is that the process
to contribute code to SAT4J will be a bit more formal than it used to  
be in the past. Any addition from someone outside the CRIL
will have to be properly reported in the code by a comment with the  
name of the contributor. As a consequence, I cleaned up
the list of people that have commit rights to SAT4J code to the strict  
minimum (i.e. current regular contributors).

If all those changes appear to be constraints, one should consider  
that SAT4J is likely to receive much attention in the next few months.
I hope that by having a wider user base, the library will improve both  
in efficiency and scalability/stability over numerous platforms.

Merry Christmas and happy new year 2008 to all of you.

--Daniel