>Maybe discuss with that person the possibility of continuing their work but having a small group of other people between them and the general public as an interface layer that handles the less important work.
So you mean you want to unseat me from leading the project I'm in charge of? What kind of punishment is this for violating some silly guidelines? [1]
>If you think that setting up such a compromise might either be too expensive, might risk burning out the interface people or has some other downside, maybe have a frank discussion with them about the benefits and costs of their contributions to the process.
Wait, but I thought that this was a guideline. Are you saying that there's actually an implicit threat within this guideline that I might be removed from the project against my will? Why wasn't that stated explicitly? How am I, a valued contributor, able to make informed decisions if rules are arbitrary and decided on behind closed doors without any ability for public input, comment, or even knowledge?
In case my point isn't obvious here, an explicit CoC has a number of advantages over a document like this one when it comes to actually resolving conflict, instead of trying to prevent it. It makes value judgements, but only because at some point in the conflict resolution process, leadership will be forced to make value judgements.
Making those value judgements explicit, and public, is a commitment from project leadership to not favor the in-group over the outgroup either explicitly or implicitly, when resolving conflict. It allows people to cite a prior commitment from the leadership to act in a certain manner. You still must trust the leadership to resolve conflict in the manner they stated they would, but you can verify that their resolution was in-line with the process and commitments they provided.
This doesn't mean that a CoC prevents leadership from attempting to mediate conflict when it comes up, and come to mutually agreeable solutions in cases when that is possible. But it does mean that in those rare cases when it is no longer possible to assume good intent that everyone is aware of the values that leadership and the decision-makers opted in to, and can hold them accountable to making decision in line with those values.
With this, Stallman hasn't made a commitment. He's written some nice platitudes, but when it comes down to it, how am I supposed to know if (again, hypothetically speaking) I'd had bad interactions with GNU or its members previously, Stallman would really value my experience strongly enough to take action instead of just letting me suffer to maintain the status quo? How can I try to hold him accountable when I don't know his values, and it would be completely possible for him to write off my complaint as simply not assuming good intent on the part of other contributors?
I can't. That's the problem.
[1]: Note that in this little hypothetical, you're my "boss", so the suggesting to put someone between me and the people I lead can be reasonably interpreted as a demotion or attack on my authority.