[llvm-dev] Resuming the discussion of establishing an LLVM code of conduct

Charles Davis via llvm-dev llvm-dev at lists.llvm.org
Thu May 5 04:14:22 PDT 2016

In the interests of individual liberty and individual justice, I feel I
must speak now.

The last sentence of the third paragraph bothers me:

In addition, violations of this code outside these spaces may affect
> a person's ability to participate within them.

​This essentially gives the committee *carte blanche* to police our
thoughts no matter where we are or what we're doing. I don't like the idea
of having my thoughts policed. There are people out there who *will*
abuse ​this
for their own ends! I can't let those people do that.

I'm afraid if this sentence goes in, I go out--and fork the LLVM family.
Yes, I feel *that* strongly about freedom of thought.

In the sample list of unacceptable behaviors, I'd consider adding the

   - Demanding special treatment for being a particular race, sex, sexual
   orientation, gender identity, etc. *Nobody* gets this privilege.
   - Kafkatrapping (e.g. denying X proves you are part of problem X)
   - False accusations
   - Dog-piling (inviting a bunch of people, many outside the community, to
   join the conversation and attack the target)

I'd also consider, in the "Personal Attacks" item, that the emphasis on
racist and sexist terms be removed. Yes, they're bad. The individual is not
the mass, after all. Perhaps in addition to the "Personal Attacks" item we
should also have an item for treating people as parts of groups instead of
as individuals. There's no need to deny their lived experience by jamming
them under some worthless label.

​I also would like​​ a clause demonstrating our commitment to technical
excellence. In particular, I would like it to emphasize that the
*only* standard
by which a contribution is judged is its merit. No other factors may be
used--and certainly no irrelevant factors such as race or sex or sexual
orientation or gender identity or... (And I don't just mean technically
excellent code here. Artistic contributions should be judged on artistic
merits. Documentation and other writing should be judged on correctness,
readability, and usefulness.)

Finally, I fear that the reporting process will be abused by less savory
people to destroy their enemies. For this reason, I suggest that there also
be consequences for the *accuser* if:

   - The accused is punished, and
   - The accused is later found to have been innocent.

In this case, the accuser would also suffer the punishment. (Of course,
this can be abused, too. We'll have to strike the right balance between the
rights of the accuser and the rights of the accused. This is hard to get

Just my two cents. I actually don't expect you to act on any of these. In
fact, I expect you all to write me off from this point forward, just for
that last proposal. ;)  But if you act on only one, please make it the
first one. I'm willing to compromise on the others, but no controls on my
speech outside of LLVM's spaces is non-negotiable.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160505/c0512627/attachment.html>

More information about the llvm-dev mailing list