Maggie Xiong wrote: > I like the idea of breaking up PDL into a lean core and separate > graphics etc components as well. I use PDL for statistical analysis on > large datasets and my pdl mantra goes like this > > use PDL::LiteF; > use PDL::NiceSlice; > use PDL::Graphics::PGPLOT; > > On systems with no Fortran compiler I have to let go of PGPLOT and use > PDL::MatrixOps. I would like to move away from Fortran but there is so > far not a comparable module to PDL::Slatec and PLPLOT had problem > closing windows on xwin. I don't know if it's fixed in the new version > or not.
I would like to see the base PDL buildable with perl and C. For example, we could use an f2c version of the SLATEC library rather than fortan sources in the PDL distribution. Another alternative is to use C based external librarys for the basic functionality. For example, GSL has *lots* of functionality that could be a baseline default. If you needed more performance, there would be additional modules you could install. > PDL is a great general numerical package and it's not just > astrophysicists' package any more. Not everyone needs TriD. If it has to > be a bundle, maybe it's appropriate to learn from MS and put out PDL > Home and PDL Professional and PDL Ultimate versions?? OK, maybe more > like PDL Core, PDL Astro, and PDL Ultimate? Maybe we need an interactive PDL installer module? That could then run cpan, ppm, whatever to get the required modules, test build, and then install. Cheers, Chris _______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
