Voting for PEP 308

[Paul Moore]

Erik Max Francis <max at alcyone.com> writes:

> Roy Smith wrote:
>
>> Fascinating.
>> 
>> Erik and myself read exactly the same thing and came to two completely
>> different conclusions.  Erik is afraid Aahz is stacking the vote in
>> favor of defeat.  I'm afraid Aahz is stacking the vote in favor of it
>> being passed.
>
> Could you explain why you think that the suggested vote process is
> stacked in favor of the vote passing?  I explained (or at least tried
> to) why I thought it seemed stacked against it passing; could you do the
> same?

I didn't follow your explanation in the slightest.

> I'm surprised to hear someone suggest that it seems stacked in favor of
> the passing, since the voting proposal starts with a no-confidence vote
> that would affirmatively reject the proposition,

Ah. I see. You want conditional expressions (pretty much) regardless
of syntax, so you see the first vote as a way for the nay-sayers to
scupper the process before we get to the real meat - discussing the
best syntax.

On the other hand, as someone who wants the PEP rejected (but who is
willing to admit the possibility that there just may be a syntax that
works, even if no-one has found it yet), I see the 2 stage process as
a trap. People like me may be reluctant to be quite so absolute as to
say an unconditional "no". Then we get left with a syntax we'd reject
immediately, because we hate all the proposed syntaxes.

> and then and only then does it consider possible implementations.
> (The failure of the no-confidence vote to pass would then, in the
> wording of the original proposal, constitute a validation of a
> "threat to forever withhold ternary conditionals."  That seems
> pretty loaded to me.)

Not really. Guido just wants to know if the community can reach
consensus. He'll go with that consensus.

It seems self-evident to me that there *is* no consensus. That's a
shame for the people who like the PEP, but it's not my problem as
such. Guido has said that if the community can't agree, the PEP gets
rejected.

So basically, Guido has put the onus on those who want conditional
expressions to persuade the rest of us. You may not like this, but
that's Guido's right as BDFL. The burden of proof is on you.

> He was even talking about a 2/3 or a 3/4 (!) supermajority on the
> no-confidence vote right out of the gate!  How can that possibly be
> considered stacking in favorite of _passing_?  I haven't been tallying
> up, but from the jist responses so far, a three-quarters supermajority
> requirement would almost guarantee a failure (certainly more than one
> out of four people who have piped up have expressed displeasure with the
> idea).  Two-thirds, maybe; simple majority probably.

Neither side can get a 3/4 supermajority. So it's undecided. That
seems an accurate assessment to me.

What you don't like is what Guido proposes to do on an undecided
vote.

Of course, that's easy for me to say as Guido appears to be coming
down on the "no" side in the case of an undecided vote. But don't try
to misrepresent the community position just because you don't like
what will happen if you're honest...

> But this all misses the point.  In my opinion, a vote tally should only
> serve as a simple record of 1. whether people are in favor of a
> conditional operator being added to Python, and 2. if so, which one. 
> Such a vote should only serve as a guideline to Guido as to what the
> community feels about it, and then (obviously) he will make the final
> decision, as in good BDFL style, taking into account the results of the
> votes (or not, should he deem them misguided).
>
> This nonsense about required supermajority no-confidence votes seems to
> miss what should be the real goal of such a vote, which is simply to
> provide Guide with data about what the Python community feels about the
> PEP.

To some extent I agree. The community is undecided. That strikes me as
a self-evident fact. Trying to make a consensus where there isn't one
is the nonsense.

What Guido does with an undecided result isn't for us to say.

Paul.
-- 
This signature intentionally left blank