On Thu, May 26, 2011 at 6:36 PM, Craig Yoshioka <crai...@me.com> wrote: > Thanks, I will. I was just seeing if there was any intention of adding this > to type of support to numpy directly. It would be faster, and I'm sure it > would make projects like dataarray much simpler to implement (dataarray does > a lot more than my suggestion).
Datarray was deliberately implemented as a subclass of the numpy array so that it would be as easy as possible to eventually merge things into numpy proper. We simply wanted to experiment with the api and code in a separate project for simplicity of development, but once things are sufficiently baked out and validated by others, there's every intention of pushing this into numpy itself. And yes, we are indeed trying to make some progress on it, we spent some time at Berkeley today on it, and a new grad student from the stats department, Jonathan Terhorst, is putting some solid work into it over the next few days. I really hope it won't be much longer before this is finally ready for production use, I know we need it badly in multiple places. Best, f _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
