On Wed, Oct 13, 2010 at 6:02 AM, Lluís <xscr...@gmx.net> wrote: > Yup, I know it brings havok to the tests, as the constructor interface > is essentially different. > > I'll try to find some time at the beginning of next month to fix all > missing bits on my pull requests. > > An yes, I saw you comment. On the naming issue... I thought that > datarray module would be inside numpy, so the construction interface > would be: > numpy.datarray.array(...) > User-friendly constructor with "inlined" contents. > numpy.datarray.DataArray(...) > Low-level contructor with uninitialized contents.
I see your point, but I think overriding the 'array' name is not very safe in the long run. Imagine reading a fragment of code using datarrays, you'd have to constantly look at the top for the import line, to see 'is this the 'array' from datarray or the numpy one? And if you need to use both (likely a very common scenario) you'd be aliasing one of them all the time. So let's avoid the naming collision from the beginning, rather than having everyone work around it always. Cheers, f _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
