[Numpy-discussion] NA masks in the next numpy release?
Sun Oct 23 17:34:27 CDT 2011
On Sun, Oct 23, 2011 at 2:29 PM, Eric Firing <firstname.lastname@example.org> wrote:
> Ultimately, though, the numpy core developers must decide what goes in
> and what does not. Consensus is desirable but may not always be
> possible or optimal, especially if "consensus" is interpreted as
> "unanimity". There is a risk in deciding to accept a major change, but
> it is mitigated by the ability to make future changes, and it is a risk
> that must be taken if progress is to be made. As a numpy user, I was
> pleased to see Travis make the decision that Mark should get on with the
> coding, and I was pleased to see Charles make the decision to merge the
> pull request.
Well, let's not jump to conclusions -- this is why I wrote an email
asking questions in the first place :-).
Consensus certainly does not mean unanimity, but yes, of course,
sometimes disagreements are irreconcilable. As a benevolent
dictator on other projects I've been stuck dealing with some of
these myself. But of the two core numpy developers who have been most
involved in this, Charles has just stated that he thought there had
been more discussion than had actually occurred, and Travis described
a "reasonable opposition", so it's not at all clear to me that the
core developers have decided that no consensus is possible and they
simply have to step in. (And in general, irreconcilable differences
are quite rare in FOSS projects... e.g., I remember the Subversion
folks set up a voting procedure to handle these cases, and then the
only time they used it in like a 5 year period was to settle an
argument about code formatting. Insisting on consensus really does
mostly work, even though it does often take longer than one would
like. And in this case I do think we can come up with an API that will
make everyone happy, but that Mark's current API probably can't be
incrementally evolved to become that API.)
So if there's been an executive decision then I can live with it, but
I'd like to see someone say that before I assume it's true. It's just
as likely that there was confusion, or Charles jumped the gun, or
whatever, and that consensus is still useful and desired in this case.
I hope so.
More information about the NumPy-Discussion