On 02/08/2012 11:56 AM, Dr. Alexander Klein wrote: > Dear list, > > for quite some time I've kept the software I wrote for my PhD spinning on my > hard drive. It's a library that provides two main functionalities: > > - Sampled continuous wavelet transforms, along with two flavours of > continuous inverses, currently with Morlet (real and complex), Paul, > Daubechies-Grossman-Meyer-, and Mexican Hat wavelets. > > - Sampled continuous matrix wavelet transforms of functions R->R^n - as > opposed to the R^n->R functions occuring with multi-wavelets; again with > several flavours of inverses. > > I'd really like to make it available under the GPL, but copyright issues > prevent me from publishing the software as is, so some kind of thorough > rewrite is due. > > I'm not sure, however, how to proceed. On the one hand, there already are > several wavelet toolboxes, but none of them does it all, and hardly any one > of them does arbitrary inverses. > > Moreover, I'd like to get rid of any procedural interfaces, because they > cause some headaches related to memory management that an object oriented > approach would simply make disappear, but then again, I haven't done a lot of > oo-programming in Octave, and I wonder if there are any pitfalls I'm not yet > aware of? > > Finally, I don't have any clear concept let alone a roadmap for pushing all > the heavy lifting to an external library. Since all the stuff has the chance > of being used in closed-source production code, too, I'd like to put it in a > library I can just re-licence under the GPL in the long run, but what would > be the best way to do it? I am admittedly not the biggest fan of C++, but > there is a Java package on Octave-Forge, and I already wrote some java > wrappers for FFTW. Can anyone perhaps comment on the maturity of the Java > package? > > Kind regards, > > Alex > Alex,
Unless I am missing something (as usual) I don't find wavelets in octave forge at all. Speaking for myself I would like to see a forge wavelets package. I don't use many forge packages, but if such a package is carefully constructed it would be a great addition. I agree that an oo approach would be much nicer than a procedural approach providing the objects are natural. The current octave object oriented capabilities are based on the old-style MATLAB approach. It is acceptable but really pretty clunky. If we could get someone fired up to implement the classdef approach it would be much nicer. Finally, the java interface is there and the little bit I have tried seems ok, but I would strongly recommend c++ just because it it more mature and is an integral part of octave. That is personal preference, of course. Bob ------------------------------------------------------------------------------ Virtualization & Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/ _______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
