Re: IESG Statement On Oppressive or Exclusionary Language

[Nico Williams <[e-mail]>]

On Tue, Jul 28, 2020 at 12:59:36AM -0400, John C Klensin wrote:
> (2) Creation of a dictionary of bad words (even to the extent
> that the I-D does it) is unlikely to be satisfactory, especially
> if, in practice, it devolves into "anything not forbidden by the
> list is allowed".  Encouraging some group of people to
> participate in the IETF by playing enforcers of cultural norms
> (whether as "language police" or otherwise) or in the form of a
> WG is unlikely to work out well for either the IETF or the
> portions of the Internet that we are supposedly trying to make
> better... and the people who depend on our work.

Yes.  The way to enforce reasonable language is to review for it as
empathetic adults rather than with a list of verboten words, and to
raise and discuss any issues with document authors.  This would be (is)
much broader and effective than a list of verboten words.  And the group
involved in enforcing this rule is the entire community (rather than a
special group of "enforcers"), starting with editors/co-authors, WG/RG
participants, shepherds, responsible ADs, the IETF/IRTF (e.g., during
IETF LC), the I* leadership, the RPC, and the RFC-Editor.

BTW, regarding {color}lists, we generally refer to "local policy" in
RFCs, not to specific implementation designs for local policy.  I've
checked my cache of RFCs I frequently check and there are many many more
hits for 'local policy' than '(white|black)(-| |)?list', and many more
hits for 'master' than 'slave' (only NTP and DNS in my cache refer to
'slave').  There are also more instances of 'primary' than 'master', and
'replica' than 'slave'.  I can happily say that none of my RFCs (unless
I've somehow failed to cache some of them) use offensive language.  My
cache is fairly small though.  A more complete analysis of Internet
RFCs' use of problematic language would be useful.

Speaking of which, I don't see in draft-knodel-terminology anything like
a survey or performed a survey of RFCs, current I-Ds, and maybe even
expired I-Ds, of how many instances of offensive language appears in
them, and even the frequency of offensive language as a function of
publication/submission year.

It would seem that evidence in this matter is important.  Evidence of
our having a problem would help drive faster adoption of better
solutions.  Conversely, lack of evidence, or even evidence that we don't
have a problem, would help us quickly dispose of solutions to non-
problems.

> (3) We are an international community with aspirations to be
> even more so.   That may imply that a term or acronym that is
> neutral or otherwise acceptable in English may be offensive,
> oppressive, or exclusionary when translated or transliterated
> into another language.  We should probably be aware of that too.

There are limits to how sensitive we can be to issues we're not aware
of.  I.e., we depend on reviewers to tell us about the issues they are
aware of.  Which brings us back to your point about banned word lists
not possibly being sufficient.

Nico
--