A D-ish wrapper around PLPlot's low-level D-to-C bindings sounds great to me too. I frequently use the D -> data file -> Python matplotlib route myself. Something more direct would be great.
--bb On Sun, May 10, 2009 at 5:51 AM, Fawzi Mohamed <fmoha...@mac.com> wrote: > On 2009-05-10 05:19:53 +0200, dsimcha <dsim...@yahoo.com> said: > >> As the scientific computing libraries for D improve, I find myself wanting >> more and more to be able to plot stuff straight from D without having to >> rely >> on kludges like writing data out to a file and then calling Python or >> Matlab >> or something. I've noticed that PLPlot has D bindings. Its license is >> also >> reasonably permissive (LGPL). This is certainly an improvement over >> nothing, >> but the API kind of sucks because it was written in C. For example, >> instead >> of ranges or D arrays of arbitrary type, you pass data in as a double* and >> a >> number of data points. >> >> On the other hand, all the nitty-gritty, low-level, probably >> platform-specific, stuff needed for a plotting library is (I guess) pretty >> good. This led me to the following idea for how to get a good D plotting >> lib >> for relatively few man-hours: Take the low-level stuff from PLPlot, and >> reimplement the higher level stuff on top of it in pure D, using the full >> power of templates, ranges, builtin arrays, etc. I'm considering making >> this >> my next hobby project, and I'm interested in some suggestions on how it >> should >> be done (what a good API would be, etc.), as well as getting an idea of >> how >> many people are interested in something like this. > > This is definitely very interesting, having an integrated plot would be very > nice > > Fawzi > >