[Numpy-discussion] On responding to dubious ideas (was: Re: Advanced indexing: "fancy" vs. orthogonal)

Fernando Perez fperez.net at gmail.com
Wed Apr 8 16:32:27 EDT 2015

Trying to bring the meta back into this thread (sorry for Robert's PB :)...

The only thing I'd like to add, is that it's perhaps worth messaging that:
a PR is just (as the Github folks like to say) "a conversation based on
code". It is NOT necessarily something intended explicitly for merging.

In IPython, we do sometimes create PRs that we mark explicitly as *Not for
merging*, just so that we can play with a concrete implementation of an
idea, even when we know in advance it's not going to go in. But it may be a
useful way to explore the problem with code everyone can easily grab and
run, to have a thread of discussion right there next to the code, to evolve
the code together with the discussions as insights arise, and to finally
document anything learned, all in one place.

So, with a bit of messaging, "encouraging PRs" doesn't need to be seen as
"the numpy core devs would like to see every last crazy idea you have in
mind to see what we can merge in our next drunken stupor". But rather,
"some ideas, even crazy ones, are easier to understand when accompanied by
code, could you send a PR for what you have in mind, knowing we're nearly
certain we won't merge it, but it will make it easier for us to have a
fruitful discussion with you".

Back to lurking ;)


Fernando Perez (@fperez_org; http://fperez.org)
fperez.net-at-gmail: mailing lists only (I ignore this when swamped!)
fernando.perez-at-berkeley: contact me here for any direct mail
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20150408/416db098/attachment.html>

More information about the NumPy-Discussion mailing list