On 29 Jan., 22:22, Bruno Postle <[email protected]> wrote: > Python is already a compile-time option, so presumably a non-python > build wouldn't need numpy either.
correct. I just want to propose that either numpy becomes a standard dependency for hsi builds or a mechanism is proposed to enable the vectorized transform and it's dependencies with an additional CMake option, when hsi is built. I'd vote for the first option, because I think more hsi code can benefit from numpy integration. > If and when Hugin switches to a current vigra library, there will be > new dependencies here anyway - vigra-1.8.0 has a boost-python and > numpy extension, so a future Hugin that links to this is likely to > indirectly require numpy anyway. The integration between vigra and Numpy is exemplary. The python module integrating vigra into python is actually called vigranumpy, because they are so tightly interwoven. Numpy's multidimenional arrays are just the sort of datatype you need when you're wrapping a generic library, and vigranumpy has very intelligent concepts to transport as much of vigra's genericness into python. I'd say numpy's way of dealing with structured homogeneous N-dimensional data is about as good as it gets, and even C++ programmers should consider using it, even though it's nominally python - after all everything that is nominally python is a C program in disguise, apart from a few exotic python implementations which use java instead. Kay -- You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/hugin-ptx
