[snip] On 08/19/2013 05:06 PM, Joe Harrington wrote:
A reorg that would bring us to a very heirarchical structure would be disruptive to existing code. Yet, there are maintenance and startup speed arguments in favor of a heirarchy. However we resolve it, I don't know that singling out the financial routines is the right short-term approach, and it wouldn't make much of a dent. So, I'd suggest we take it as two issues to solve separately 1) keeping or tossing the fin routines and 2) moving toward a heirarchy of routines or not.
Talking about existing code, honestly I don't think we should worry about that much. I made a not-so-shallow search on GitHub and Google and I saw zero mention of numpy.financial in the NumPy User Guide, zero in the SciPy Cookbook, no mention on GitHub (people are pushing their virtualenvs to their repos, so NumPy source code is replicated a thousand times on the internets ;) including financial_test.py), and, well, I didn't go through all the ooooo's of Google but you get the picture. Making people used to IDL and MATLAB use namespaces is an issue completely unrelated to this in my opinion, or better put, much more general, because it affects the whole scientific Python ecosystem. Josef mentioned matplotlib.finance: https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/finance.... "This module is deprecated in 1.4 and will be moved to `mpl_toolkits`or it's own project in the future.". They might be in a similar situation than NumPy. I don't have an opinion on the usefulness of financial functions on NumPy and I have zero knowledge on finance myself, but if this discussion makes someone write a nice tutorial on using them (being on a subpackage or not) and attract people to actually use them (I would bet very few are doing so as of now), then great! My proposal is moving them to a separate subpackage as a first step to their disappearance. Maybe when someone sees the move and/or the deprecation note on the release notes that tutorial will be written :) Juan Luis Cano