[Numpy-discussion] A numpy based Entity-Component-System

[Benjamin Root]

As a bit of a real-life example where things can go wrong with naming. The
"pylab" name was accidentally hijacked a couple years ago on pypi, and
caused several bug reports to be filed against matplotlib for failing
scripts.  Some people thought that one should do "pip install pylab" to do
"from pylab import *" -- crazy, I know, right? ;-)

That was at least two years ago, and we are just now getting the person who
uploaded that errant package to fix the problem. We also have not yet been
able to find out the owner of the long defunct matplotlib twitter handle
(we started the process to reclaim it, though). There are a couple other
situations that have caused confusion among users, but these have been
mostly an issue of trademarks (which I believe NumFOCUS holds for
matplotlib?).

What you really have to watch out for is when someone creates a package
that uses some open-sourced library, and then claims that it is "supported"
by it. This can cause an undue burden on the original maintainers in
fielding bug reports and other issues. It also creates a false sense of
association/coordination -- which gets me to the issue at hand. I highly
suggest coming up with a unique name. It benefits you as it becomes
something distinct from numpy (so search results are more relevant), and it
benefits the numpy developers because support requests go exactly where
they are supposed to go.


On Fri, May 20, 2016 at 9:40 PM, Nathaniel Smith <njs at pobox.com> wrote:

> Like naming your child "Human-legsarms"?



You owe me a new monitor!

Cheers!
Ben Root
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160523/44fd8b8e/attachment.html>