We will most likely not distribute the source code to keep the COSI-Corr development under our supervision and quality assessment. If we were to distribute the source code, it's rewriting in another language with multi-processor capabilities will not be trivial matter. The COSI-Corr source code is now quite large and we're still extending it. I believe that if an external team were to understand our codes and rewrite it, it would take a significant amount of resources (I'm thinking about 4-5 computer scientists already familiar with remote sensing design, working full time for at least 6-8 months). If you are indeed willing to invest that much resources, it certainly means that your earlier tests with COSI-Corr were conclusive. If so, I would be happy to have your feedback on the results you obtained from COSI-Corr, how well you were able to measure what you are interested in, what level of accuracy reached, what kind of images you used, etc...
If I remember well, you are interested in measuring sand dune migration rates. Here's a paper from Pieter Vermeesch you might be interested in if you haven't read it yet: http://pvermees.googlepages.com/VermeeschGRL2008.pdf