Re: State of the GNUnion 2020

[Alexandre François Garreau]

Le samedi 22 février 2020, 22:50:03 CET Samuel Thibault a écrit :
> Alexandre François Garreau, le sam. 22 févr. 2020 22:43:36 +0100, a 
ecrit:
> > Le samedi 22 février 2020, 21:10:33 CET Samuel Thibault a écrit :
> > > Alexandre François Garreau, le sam. 22 févr. 2020 19:21:18 +0100, a 
ecrit:
> > > > 
> > > > No, you can’t know.
> > > 
> > > Strictly speaking, sure, but at some point when things only get
> > > worse
> > > and worse, it's way better to resort to exclusion than continuing to
> > > see a project get stuck due to only one person, even if with
> > > several dozen years you might find a solution.
> > 
> > However I’m still dubitative about the fact a single person could get
> > ALL A PROJECT project stuck alone, without exception.
> 
> I was indeed taking an extreme example. But a much simpler case would be
> somebody calling everything names, thus driving away newcomers, who
> will wonder that the heck this project is about.

That seems pretty simple to issue a warning about a such issue, to explain 
a such person why it’s not good (because it looks really repetitive and 
simple), and possibly to filter the way I propose later. 

> > > How different is this from exclusion?
> > 
> > It is different in that that anybody have the free choice of
> > moderation, or not.  You could have your own moderation filters
> > overriding those of your trusted friends, you could change the amount
> > of trust you give to each, etc.
> 
> Ok, so newcomers would have to know about it and make it work to avoid
> the nasty messages?

No, it has to be instantiated either as an explicit choice at mailing-list 
subscribing, either as a default, but one that can be tweaked, either as a 
standard protocol to be complied with (be it based on something already 
existing such as MIME and/or rfc822 headers, but if it’s backward-
compatible we have to make sure the non-complying clients can get a 
server-side “default“ that would comply to desired behavior according 
historical implementation (that is classical moderation that is possible 
to tweak/disable)).