Re: State of the GNUnion 2020

[Dmitry Gutov]

On 20.02.2020 15:45, Samuel Thibault wrote:

> > > The activity by itself, yes, but the choice of where to start a new
> > > project, or starting contributing an existing project, leadership does
> > > have a lot of importance.
> >
> > What kind of choice? Contributors come and go, largely depending on their
> > own needs and interests.
>
> Yes, but also, and I believe most likely, depending on their knowledge
> of project places (github vs gitlabs vs savannah) and the contact they
> get with the people there. The GNU project is less and less known
> compared to other free software platforms, so it'll get less and less
> newcomers.

That is a problem. But one that wouldn't be solved simply by the 
leadership's say-so. GNU is usually all volunteers, and if existing 
developers don't accept the new project management platform, they won't 
use it. And vice versa, if they like it, the project can migrate to it 
unilaterally. With the exception of strictly proprietary stuff like 
Github, but it's out of the question for GNU anyway.

Depending on how flexible the core developers are, I guess RMS's 
recommendation on questions like that will have some influence. But so 
would enthusiastic, targeted advocacy toward the use of new tools. One 
doesn't really have to be RMS to work with individual projects and their 
contributors and discuss their choices and needs.

> > > > And it's a more difficult endeavor (think Mozilla-type initiatives) than
> > > > just releasing a document saying "hi all we don't discriminate and accept
> > > > everyone", which is basically stating the already obvious.
> > >
> > >   From seeing the discussions here, it doesn't seem so obvious :/
> >
> > Really? For all the shouting and stomping of feet, I haven't seen here any
> > one email stating or even implying that the gender or the race of a
> > contributor is somehow important, or that we'd turn somebody away because of
> > it.
>
> Sure, the contrary was explicitly said indeed. But anything one can
> bring about not only acknowledging it, but also making efforts on
> inclusiveness is mostly rejected with arguments like "it's too hard to
> take care when writing something on a mailing list".

The argument was, in a typical programmer fashion, that a lot of things 
could be taken as "harassment", and so a subjective thing like that 
can't be explicitly disallowed. And that our existing "kind 
communication guidelines" cover a lot of the same ground already.

Regarding punishing repeat offenders anyway, as we've seen just 
recently, you can't censor a determined individual on a public mailing 
list anyway. Limit their audience, sure, but banning them outright seems 
impossible. And I can hardly see the whole GNU project migrating off 
mailing lists.

> > On the flip side, an argument is made that your initiative might make GNU
> > more exclusionary because of the extra conditions on what it takes to be a
> > part of it.
>
> At some point you have to exclude some people in order to include other
> people, yes.  We can see that in various communities: when somebody is
> having a toxic behavior and does not changes behavior even after strong
> warnings, one has to exclude that person, because otherwise that person
> will make a lot other people fly away.  Not taking the steps to exclude
> the toxic person does mean excluding people that can not stand the toxic
> behavior, even if that latter exclusion is not explicit.
>
> That seems to be the ground of what some people do not understand here:
> full inclusiveness can not work, there will always be some people you
> will be excluding one way or the other, voluntarily or not.  Making sure
> that the choice of who you exclude gets written down seems important to
> me.

This is a very common argument about CoC's. I was talking about 
something different. Forget harassment and antagonistic behavior.

If we declare that all of GNU should share a certain set of values, and 
especially that maintainers must share the free software values, 
whatever it really means in practice, *that* sounds exclusionary already.

For better or worse, a lot of my colleagues, and a lot of users and 
Emacs contributors (the main GNU project I contribute to) use 
proprietary OSes. Even the maintainers do (though not exclusively). I am 
not fond of that, but I started using Emacs in a similar position years 
ago, and I wouldn't want to exclude any of them from being a part of our 
project because their stance is more lax, or that their end goals are 
more utilitarian (at least for the time being).

And that is where your argument stops working.