In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (John M. Gamble) wrote:
> >> The > >> original source for the algorithm used in the module is > >> from Hiroshi Murakami's Fortran source, and it shouldn't > >> be too difficult to repeat the translation process to python. > > > >Ah ok, I'll try to locate that (following the instruction in Solve.pm > >didn't work for me :( ). > > > > Ouch. I just did a quick search and found that that site has undergone > a few changes, and the code that i reference is missing. A few other > links in the docs are stale too. I need to update the documentation. > > Anyway, doing a search for 'hqr' and Eispack got me a lot of sites. > In particular, this one is pretty friendly: > > <http://netlib.enseeiht.fr/eispack/> > > Look at the source for balanc.f (does the prep-work) and hqr.f > (does the solving). Minor annoyance: the real and imaginary > parts of the roots are in separate arrays. I combined them into > complex types in my perl source, in case you want to make a > comparison. Thanks! I'll check that out. > Of course, all this may be moot if the other suggestions > work out. SciPy indeed appear to contain a solver, but I'm currently stuck in trying to _get_ it for my platform (OSX). I'm definitely not going to install a Fortran compiler just to evaluate it (even though my name is not "Ilias" ;-). Also, SciPy is _huge_, so maybe a Python translation of that Fortran code or your Perl code will turn out to be more attractive after all... Just -- http://mail.python.org/mailman/listinfo/python-list
