Remove RMS from the GCC Steering Committee

[Joseph Myers]

On Sun, 28 Mar 2021, Mark Wielaard wrote:

> He does indeed show up randomly claiming authority even if the GNU
> community has told him no. And it is important to say upfront he has
> no authority and that his attempts to cancel the work of hardworking
> GNU contributors is unwelcome. IMHO for the GCC community this means
> to be explicit he doesn't have any authority and he shouldn't be on
> the GCC steering committee.

For example, consider the October 2019 discussion on libc-alpha of 
removing the abort "joke" from the glibc manual.  We rejected RMS's claims 
of authority to say that the joke should be kept, or kept indefinitely 
until various general points could be decided, and removed it from the 
manual anyway without waiting for conclusions on all those general points.

RMS only has authority over decisions taken about individual GNU packages 
where the people developing those packages let him have that authority and 
make or refrain from making changes based on what he says.  We should not 
give him such authority by treating his views as having some significance 
not given to such views expressed by other people; changes he suggests can 
be considered, and accepted or rejected, on their merits.  And the abort 
joke case illustrates that in fact he is not given such authority, when 
package developers are confident to stand up to claims he makes of 
authority, and provides an example that can speed up the rejection of any 
such assertion of authority to micromanage things that might be made in 
future.

I agree with the conclusion of Nathan's original message, that RMS behaves 
in a toxic way, it is harmful to have him listed as being in a leadership 
role that might suggest what he does is acceptable within the project, and 
he should not be on the SC.  This is based on the longstanding, 
well-documented patterns of how he has misbehaved towards women, *not* on 
the opinions he has expressed on other subjects, *not* on his choices 
regarding the use of language, *not* on his attempts to insist on language 
being used in particular ways, and *not* on where or when he has chosen to 
express such views.

For the same reasons, I think it is harmful for him to be Chief GNUisance 
(but as above, I think GNU packages should not give a Chief GNUisance 
authority to micromanage decisions, beyond ensuring GNU packages follow 
basic GNU free software principles and cooperate with each other and with 
their development communities), harmful for him to be on the FSF board, 
and harmful for him to be seen as leader of the free software movement.  
(For the last point, I don't think the free software movement needs a 
single leader; it needs many people advocating free software, and 
discussing issues related to free software, from diverse perspectives.  
RMS's ideas that form the foundation of the free software movement are 
still of fundamental importance today.  But other people can now build 
better on those ideas in today's context.)

RMS does not, in fact, contribute usefully to the SC.  Any time he 
suggests some feature for GCC, whether a good or a bad idea, that could be 
done just as well on the public mailing list (which would be a better 
place to find someone possibly interested in implementing a feature, and 
to discuss a feature's merits, in any case) without being an SC member.  
He's sufficiently far removed from toolchain development that he's not 
good at making reasonable suggestions for toolchain changes in any case.

We can consider individual proposals or patches from anyone on their 
merits.  We can have leaders who are accepted as leaders because 
contributors can see their relevant expertise that gives them legitimacy 
as leaders, and can see a good basis for decisions they make as leaders.  
But longstanding patterns of bad conduct by a leader, even when formally 
unrelated to the project, can reach the point where considering that 
person a leader is harmful to the project.  I think the ways RMS has 
behaved have long since reached the point where it is harmful for him to 
be considered a leader for GCC or GNU, and that's sufficient to stop 
considering him a leader (even if he were sufficiently involved to be able 
to contribute much more usefully on a technical level).

-- 
Joseph S. Myers
joseph@codesourcery.com