Hi, I'm glad to see that everyone agrees to try to see what are the merging/collaboration possibilities between all the different libraries currently being developed. Indeed I think it would be a good thing as each library have its strengths and weaknesses, this being principally due to the fact that we all have different priorities. But for a geometry library meant to be widely used, everything is a priority. So having several people each bringing his own part would surely be more productive. As Luke said, the problems are not really about algorithms but about how genericity will be handle. Once this is defined, adding algorithms only raises implementation issues for which consensuses should be easily found. A lot of things have changed since the second preview of Barend's library. Especially for the data types adaptation techniques, we have first implemented a traits approach, then a more fine-grained approach with separate metafunctions. We will try to release a preview as soon as possible to discuss those choices and see what are the points of agreement and disagreement between us. Regards Bruno