This package gives you a combined math library, vector and matrix library, and statistics
library, all packaged for .NET 2.0. There are way too many functions and algorithms
to list here, but a sampling will give you the flavor. On the math side, you get
curve-fitting, numerical integration and differentiation, and optimization algorithms,
as well as prime factorization. Vector and matrix math in both the real and complex
domains is supported, along with special matrices and numerous factorizations. Statistical
functions range from regression and ANOVA analyses to a variety of hypotheses tests
and a wide variety of random number generators as well as non-random distributions.
All of this is designed to play well with .NET 2.0. You'll find generics used where
they make sense, and a debugger visualizer for matrices. On some of the core routines,
you have a choice between processor-specific hand-tuned implementations and 100%
managed code for portability. The vendor understands .NET metaphors and so you'll
also find good use of things like delegates where necessary and desireable. The
object model appears to be extremely well organized and consistent with the .NET
Framework design guidelines as well.
There's also very good documentation here, in the form of an extensive help file
chock-full of examples, a batch of quickstart sample code, and a few larger samples
(the Mandelbrot generator may suck you in if you're of a certain age). This is good,
as the math itself is tough enough. Clearly, not everyone will need this stuff,
but if you do, this looks like a solid and serious implementation. You can download
a 60-day trial if you want to judge the performance and other details for yourself.