PseudoQ is currently "good enough" to release. It has some rough edges, and there are a number of things I still plan to improve (see the tracker pages on Sourceforge for the latest "to do" and bug lists). The puzzle generation, while better than in earlier versions, could still be better. There should now be at least one solution to all generated puzzles, although (depending on the selected initial values) PseudoQ may not be able to calculate the solution itself. Whether this is a limitation of the generation or solver is debatable. However, I plan to make further improvements to the initial value selection before the 1.0 release. Asides from the automatic puzzle generation, however, PseudoQ can handle entering puzzles by hand and playing them, and it is able to solve almost all of the puzzles I have tried entering from various newspapers and magazines.
Ranking: 18567, Activity percentile: 90.84%, as of 3 Dec 2007
2039 total downloads to date
20 Nov 2007 : Another 1000 downloads for PseudoQ
According to Sourceforge's stats, PseudoQ recently reached its 2000th download. Nowhere near as many as most of the projects on Sourceforge that I have used, but I suppose it's not too bad considering the lack of time I've had to work on it recently.<br />However, despite these additional downloads, I still have not had a single email from anyone to tell me what they think of it, good or bad. Please. contact me directly or via the mailing list, and let me know.
09 Sep 2007 : PseudoQ is Kibitzed
The SourceKibitzer web site is tracking various code metrics of the PseudoQ project. The results of this analysis can be found at <a href="http://www.sourcekibitzer.org/ProjectPage.ext?sp=Spseudoq">http://www.sourcekibitzer.org/ProjectPage.ext?sp=Spseudoq</a>
20 Aug 2007 : PseudoQ switches from CVS to Subversion
PseudoQ recently switched version control system, from CVS to Subversion. Detail for accessing the latest version from the repository can be found at <a href="http://sourceforge.net/svn/?group_id=150500">http://sourceforge.net/svn/?group_id=150500</a><br />Future changes to the code will only be made in Subversion, so anyone currently using CVS to access the latest sources should switch as soon as possible.
Suggestions for improvements to the program (new features, usability issues, new strategies for solving puzzles, etc.) should be directed to the project's mailing list at firstname.lastname@example.org. Alternatively, you may contact me via Sourceforge.
If you find a published puzzle which PseudoQ can't solve but you know to be solveable, please let me know the details, especially if you can demonstrate a way to deduce a further value or values beyond the point where PseudoQ gets stuck (in the Puzzle GUI windows, switch to Play mode and keep clicking "Hint" until it can't find another value).
You can access the source code to PseudoQ via CVS; details are available here. Patches (in the form of a CVS diff) should be uploaded to the SourceForge trackers, and should be accompanied by a statement granting the copyright to the project. This is for convenience in case I ever need to, for example, switch licenses for some reason (say, if it turns out that any of the other libraries' licenses is incompatible with LGPL). Otherwise, I'd have to chase around emailing everyone who has ever contributed a patch to get their permission.
The source in CVS includes the project files for Netbeans 5.0. If you don't use Netbeans, you should still be able to use the project files as since version 4.0 Netbeans projects are Apache Ant (1.6+) scripts. Every Java IDE I know of has some form of Ant integration. However, I'd prefer that you use Netbeans to create any patches for the GUI components, so that they include the relevant changes to Netbeans' .form files.
The source code includes extensive (JUnit based) unit tests; the test (and dist) target of the build script runs these and collects code coverage metrics. The test results and coverage reports for the latest release may be viewed online. Any patches should include additional unit tests as appropriate, and patches which improve the code coverage of the test suite are more than welcome.