[Numpy-discussion] consensus (was: NA masks in the next numpy release?)

Nathaniel Smith njs@pobox....
Fri Oct 28 21:49:05 CDT 2011

On Fri, Oct 28, 2011 at 3:14 PM, Charles R Harris
<charlesr.harris@gmail.com> wrote:
> Matthew, the problem I have is that it seems that you and Nathaniel won't be
> satisfied unless things are done *your* way.

Hi Charles,

I'm sorry if I've given this impression, and I know it's easy to feel
this way in a contentious discussion. I've even been tempted to
conclude the same about you, based on some of those emails in the last
discussion where you told us that we should only give feedback by
critiquing specific paragraphs of Mark's docs, even though our issues
were with the whole architecture he was suggesting. But I'd like to
believe that that isn't true of you, and in return, I can only point
out the following things:

1) I've actually made a number of different suggestions and attempts
to find ways to compromise (e.g., the "NA concepts" discussion, the
alter-NEP that folded in a design for "ignored" values to try and
satisfy that constituency even though I wouldn't use them myself, and
on the conference call trying to find a subset of features that we
could all agree on to implement first). I don't *want* my proposals
implemented unless everyone else finds them persuasive.

2) This is why in my message I'm *not* advocating that we implement
NAs according to my proposals; I'm advocating that you get just as
much of a veto power on my proposals as I do on yours. Let's be
honest: we both know all else being equal, we'd both rather not deal
with the other right now, and might prefer not to hang out socially.
But if we want this NA stuff to actually work and be used, then we
need to find a way to work together despite that.


-- Nathaniel

More information about the NumPy-Discussion mailing list