[Some years ago, the sci.math.symbolic newsgroup at Drexel's Math Forum site stopped disseminating to the usenet group. In the meantime they have extended their group name to sci.math.symbolic.independent and have begun warning users about the separation, so their's nothing to complain about. Yet, all posts to the usenet group sci.math.symbolic continue to appear on their site, which is nice because they are archived. As most of the current activity is taking place on the usenet group, however, Math Forum users can apparently still be misled, as in the case of this May 14, 9:20 PM post from Hawaii.
> At long last the 4th version of Rubi (Rule-based integrator) has been > posted at > > http://www.apmaths.uwo.ca/~arich/ > > The entire system of rules Rubi 4 uses to symbolically integrate a > wide variety of expressions is freely available for viewing and/or > downloading. The rules in the files are hierarchically arranged in the > form of a decision tree. By answering true-or-false questions in the > tree, it is easy for a human or computer to determine exactly which of > the more than 5000 rules is the right one to apply to a given > integrand. Also available for those having access to Mathematica 7 or > better are the source files necessary to actually try Rubi on > integrands of your choice. Optionally, you can show the intermediate > steps and the rules Rubi uses to compute antiderivatives. > > As evidence of the merits of using a rule-based approach to automate > symbolic integration, the website has a table summarizing the result > of running Rubi 4 and the two major commercial computer algebra > systems on a grueling test suite of over 43,000 integration problems. > Both the test-suite problems and raw test results are available for > viewing or downloading from the website. There has been much > discussion recently on sci.math.symbolic about establishing an > "Independent Integration Test Suite". Perhaps some of the problems in > this test suite warrant inclusion... > > Aloha from Hawaii, > Albert D. Rich