Re: On the "Words Matter" issue (was Re: The request of words matter updated; quotes deleted)

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

Thoughts and questions.

I see a number of people have capitalized "Words Matter".  I also find
it interesting that this thread was initiated by IBM employees in
China who are apparently not regular participants in zsh-workers or
zsh-users.  Is there an IBM corporate initiative called "Words Matter"
that led to this discussion being opened at this particular time?  As
has been mentioned, this topic has been around a long time (2007 at
least; https://www.snopes.com/fact-check/masterslave/).

Obviously the term "master" has definitions and connotations that
reference skill level, origin of concept or data, etc.  Those
connotations are typically clear from context.  On the other hand, the
word "slave" always refers to subservience and captivity (whether or
not accepted by "consenting adults"), and therefore carries the
emotional baggage that also attaches to "master" when the words are
paired.  (The Snopes article is some evidence that this is not just a
"white knight" issue.)

Arguments focused on those other connotations of "master" are missing
the point.  Whether or not one thinks emotional baggage is being
self-righteously exaggerated or is a valid basis to make a technical
change, it can't just be waved away with "but that shouldn't matter in
this abstraction."  If they weren't evocative, those words wouldn't
have been chosen to begin with.

(In grad school I worked on a project to compute fractal graphics with
a multiprocessor computer.  I called the nodes that did the
computation "students" and the thread that collected the results and
rendered an image the "faculty".  When I explained during a
presentation that the former did all the work and the latter got all
the credit, the reaction was ... mixed.)

In any case this mailing list is not the place to speculate about
motivations or debate the evolution of language.  We have a request
before us from an interested third party and a proposed partial
response to that request.  (List of files mentioning "slave" follows,
to explain why I say "partial".)

Questions I think it is reasonable to discuss are:

Assuming the original request is part of an IBM initiative, does
rejecting it have an impact on the availability/adoption of zsh?

Is it likely that other influential companies are going to follow
suit?  E.g., Apple appears to have adopted zsh as a default shell in
MacOS.  Is a similar request eventually to come from that quarter?  If
we don't act, are they likely to fork the code and do it themselves?

Is a change in terminology going to cause confusion with upstream
source or with packages we don't control?  E.g., several of the files
below are completions, and arguably it would not make sense to remove
strings still in use by the corresponding commands/contexts.  (We've
already answered this for zpty.c in particular.)

Is a change going to have a detrimental user-visible effect?  (This is
a more general version of the previous question.)

If the foregoing are all "no", what's the degree of effort and who is
prepared to work on it (rather than just spend time debating it)?

Here's the list of files mentioning "slave"; answers to some of the
above may vary per file:

./Functions/TCP/tcp_spam
./Src/Modules/zpty.c
./Etc/ChangeLog-3.1
./ChangeLog
./Completion/Linux/Command/_ethtool
./Completion/Linux/Command/_lsblk
./Completion/Linux/Command/_unshare
./Completion/Linux/Command/_sshfs
./Completion/Linux/Command/_networkmanager
./Completion/X/Command/_xinput
./Completion/X/Command/_mplayer
./Completion/Debian/Command/_update-alternatives
./Completion/Unix/Command/_mysql_utils
./Completion/Unix/Command/_mount

There are a lot more mentioning "master" but with the exception of
mod_zpty.yo I think they may be considered to fall among the
inoffensive connotations of that term.  (If someone wants to dig
deeper, see "who is prepared to work on it?" above.)  For what it's
worth, I think the "master" git branch belongs in that category too,
and am not advocating to rename that in our project.  I am also not
advocating for a rewrite of history, so the ChangeLog files are IMO
out of bounds.