[python-committers] Transfer of power

[Larry Hastings]

(separate reply to discuss the "what do we do now" topic)


On 07/12/2018 07:57 AM, Guido van Rossum wrote:
> I would like to remove myself entirely from the decision process. [...]
>
> I am not going to appoint a successor.
>
> So what are you all going to do? Create a democracy? Anarchy? A 
> dictatorship? A federation?

Although the timing is a surprise, the idea of Guido retiring isn't.  I 
remember shooting the breeze with Guido about it as far back as the 
Santa Clara PyCons--and I'm sure the topic goes even further back.  So 
we've had quite a while to think about it.  Here's my opinion, which I 
reached years ago and haven't appreciably changed since.

First, there's no single person in the community who can take over as 
BDFL.  It's simply impossible.  The Guido we have today is who he is 
because he's been doing the BDFL job for more than twenty years. The job 
has shaped and taught him as he did it; as Python grew, so did Guido.  
Literally anybody else we might appoint as BDFL would have to start 
fresh and grow into the job--and I don't think we can afford the growing 
pains.

On the other hand, I also think that deciding PEPs by popular vote would 
be folly.  Python is mature enough to be simultaneously robust and 
fragile, and leaving its design up to popular vote seems like a recipe 
for chaos.  In my opinion, the final arbiters of Python's evolution 
should be experts, not the masses.  (Cue "Twitch Plays Pokemon" here.)

I think the happy medium is a Council Of Elders.  Summarizing this approach:

  * The number of Elders on the Council should be an odd number greater
    than two.  It can't be one person, as that'd just be a BDFL.  And we
    want an odd number to prevent tie votes.  My instinct is that three
    would be fine.
  * For most PEPs the Elders should delegate, just as Guido has
    generally done in the last few years.  Although I expect the Elders
    to be seasoned Python core developers, they probably won't have
    domain-specific knowledge necessary to rule on most PEPs.
  * I'm not sure how to appoint the initial round of Elders. Maybe a
    popular vote?
  * However, once appointed, Elders are appointed is "for life". The
    only way to remove one would be for them to voluntarily step
    down--there would be no mechanism to remove one from office.
    (Perhaps this is too strong--perhaps one could be removed by a
    unanimous vote from all other Elders?)  I want the Council to be
    immune to popular opinion, to be empowered to do what they think is
    right without fear of anything beyond negative public opinion.
  * I'm not sure how we'd replace Elders.  Maybe they'd hold an
    internal-only election?  ("Jo has decided to step down, and we have
    elected Sam as Jo's replacement.")

Your reaction to this might be "but running Python's evolution by 
committee will slow it down!"  I suspect that's right.  Not being Guido, 
I think the Council would be more cautious in approving changes to the 
language.  But I think that'd be appropriate anyway. Python's already a 
fine language, and I can live with it evolving more slowly in the future.


Personally I'd nominate the following three Elders.  In alphabetical order:

  * Barry Warsaw (knows where the bodies are buried)
  * Benjamin Peterson (young enough that we'll get many years of service
    out of him)
  * Brett Cannon (the only candidate tall enough to be worth considering
    as BDFL replacement)

I like this union of experience and personality.  My intuition is that 
they'd find the mixture of caution and let's-go-for-it spirit that we 
need to keep Python moving forward without making too many mistakes.  
And we could call them "the B's" for short!


Finally, I agree with Raymond's call for slow deliberation. Python's not 
going anywhere, there are no burning needs for changing the language 
that need to be addressed immediately.  We can all collectively sit and 
stew on this for a while.


//arry

/p.s. I know nobody is suggesting this, but I'll preemptively say it 
anyway: let's not simply appoint all the Release Managers as our initial 
Council Of Elders.  While that'd net us some very fine Elders indeed, 
you'd also wind up with me on the Council--an obvious mistake!
//
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-committers/attachments/20180712/ba8f0b26/attachment.html>