The module Math::TrulyRandom has a number of flaws in algorithm and implementation, and currently will not work on Windows or Linux. I believe the best plan is to remove the module (and its references in Perl core documentation).
If this is undesirable, I would like to request COMAINT on the module so I can improve it. The author and only current maintainer, Gary Howland (GARY), passed away in 2003. It has had no updates since its initial release in 1996. The module is referenced in core documentation (functions rand), which should be addressed but that's another issue. The module will not compile on Windows and the tests hang on Linux. It has 7 passes of 27 reports (most systems will not report due to hanging tests, and the last success was with 5.10.0 running Solaris). It has no reverse dependencies, and with its current status I don't think anyone would be relying on it. A different CPAN module exists with similar functionality (but I am the author). I wrote a patch for Math::TrulyRandom in Jan 2013 (in thread for RT 34693) but there is nobody to apply it. This patch fixes the hang issue but does not fix Windows nor does it fix the known flaws in the underlying algorithm. My PAUSE ID is DANAJ. If people have opinions on directions for this module, please let me know. Some possibilities: - Keep flawed TrueRand 1.1 algorithm. Make this module as portable and robust as possible. - Move to TrueRand 2.1 algorithm which solves some of the worst issues. Make it as portable and robust as possible. - Make new module, e.g. Crypt::Random::TrueRand, and start process to deprecate and delete this module. - Just deprecate / delete this module, regardless of any plans for new modules. Thanks, Dana Jacobsen (DANAJ).
