Welcome

You may be interested in our completely free (license: BSD) cross-platform (Linux, Windows, Mac etc) software, written using Python + NumPy:

• OpenOpt framework - universal numerical optimization package with several own solvers (eg ralg) and connections to lots of other
• FuncDesigner - tool to rapidly build functions and get their derivatives via automatic differentiation. Also, you can solve systems of linear/nonlinear equations and numerical optimization problems coded in FuncDesigner by OpenOpt (see some examples in its doc)
• DerApproximator - tool to get (or check user-supplied) derivatives via finite-difference approximation

New FuncDesigner concepts: ooSystem & ooSystemState

• New LLSP solver for dense and
  sparse problems: lsqr
• Sad news: me & lots of our department workers were forced to furlough without pay
• Export OO and FD LP/MILP to MPS format files
• Fix latest compiler issue for LLSP solver bvls.f
• Add FuncDesigner operator ifThenElse
• Add scipy.optimize fmin
• Some changes for ralg
• OpenOpt 0.27, FuncDesigner 0.17, DerApproximator 0.17 release
• scipy.sparse: great speedup for csr matrices fancy indexation
• An issue with stacking FuncDesigner oofuns and numpy.arrays

• OpenOpt Foreword (українською мовою та чорнетка статті про OpenOpt російською), wikipedia.org entry, mloss.org entry, mac.softpedia.com entry, linux.softpedia.com entry
• Carnegie Mellon University started to teach students OpenOpt
• Some more Python scientific software ,
• "Using Python for scientific/engineering software development" (my point of view //Dmitrey)
• MIT has moved to Python from Scheme for its undergraduate CS program
• Python has won Linux Journal 2009 Readers' Choice Award for Favorite Programming Language
• Google intends to make Python speed 5 times faster
• Our former location at scipy.org