[python-committers] Anatoly has been warned about his behaviour potentially leading to his loss of tracker privileges

[martin at v.loewis.de]

Quoting Ned Deily <nad at acm.org>:

>> You can't fix people, but you can prevent them from actually being
>> harmful.
>
> The thing is it's a technical solution to a social problem.

No, that's not true. The ban itself is a social reaction to a social
problem. The technical reaction is only to actually enforce the ban.

I have personally banned two people so far from "python-dev", and
at least in one case, the ban wasn't actually enforced, but honored
nevertheless.

It *would* be a technical solution if the ban wasn't actually communicated,
but only implemented (something which is quite common in RL, e.g. when
people change the locks on their doors to lock out their former partners)

> the former tend to be all that effective for the latter.  And I  
> think reasonable people can disagree about the degree of  
> harmfulness.  I personally don't see his behavior, in and of itself,  
> as all that harmful.  I *do* see the negative reaction it provokes  
> as being harmful.  Clearly, it bothers people and that is  
> disruptive.  But it would be a whole lot less disruptive if we  
> didn't let it be, e.g. by just letting it go and ignoring it.

Since nobody mentioned it this time (or since I missed if somebody did),
I'll mention the "poisonous people" talk from Collins-Sussman/Fitzpatrick):

http://www.youtube.com/watch?v=Q52kFL8zVoM

I said this several years ago, and I still believe that anatoly is a
poisonous person, in the sense of this talk.

Several strategies just don't work here, e.g. trying to win an argument
with anatoly. A strategy that I believe that *also* doesn't work is
to let "the community" ignore him. In a free software project, fluctuation
is just too high to make this work.

It takes several years (for some of us) to recognize that ignoring
him entirely is the only reasonable personal reaction. If we wanted to
effectively make it work, we would have to educate every single contributor
"don't talk to anatoly, and don't respond if he is talking to you".
This can't work in the large scale.

>> If python-list is a troll magnet, that's a pity, but how is that
>> relevant to the *development community*?
>
> It's relevant because python-list is yet another forum hosted by the  
> PSF via python.org mailing lists and is viewed as part of the  
> broader Python community as a whole.  If we propose to ban someone  
> from python-list, along with other lists, that raises the question  
> of what standards are being used.

I don't think anybody should be banned from python-list; I think talk
is just about "python-dev" (including all core cpython infrastructure).

> It is a problem.  And choosing to not participate is a perfectly  
> rational and legitimate response.  But it doesn't necessarily follow  
> that banning someone is a better response.

I think it is. Based on past experience, it would be temporarily anyway,
and it may buy us a year or so of mental peace.

Regards,
Martin