Hi,

Michał Górny <mgorny@gentoo.org> writes:

> Hello,
>
> We have been opposing the "industry trend" to use more inclusive
> language for a long time.  I don't disagree with our pragmatic
> rationale -- we are simply using the original branch name as used by
> git(1) back when we created the repository.
>
> I honestly don't think any of us ever meant to use the word "master"
> in any derogatory sense, neither I do think the git author or the
> Linux kernel authors meant to.  In fact, given how many of us are not
> native English speakers, and may have learned much of English
> vocabulary in IT, I can imagine that we may have learned such words as
> "master" and "slave" primarily as IT terms, without even realizing
> their derogatory meaning, or suspecting anything sinister about it.
> It does sound like a major hassle to do a change that can't be fully
> transparently handled everywhere (add a lot of muscle memory on top of
> that) for something that wasn't our (or our users, who also will be
> affected) malice.
>
> I've only recently learned of a possible link, from David Rooney's
> book "About Time".  To cite:
>
>> The clock that was originally installed at the Greenwich observatory
>> in 1927 to transmit the time to Rugby for broadcast to the British
>> Empire was a type called a "free-pendulum," which actually comprised
>> two devices, one called the “master” and the other termed
>> the "slave."  The master clock was free and did little work;
>> the slave clock was forced to march to the master’s beat and carried
>> out all the labor of time distribution to the radio transmitters.
>> The terminology of slave clocks had been coined in 1904, by a British
>> government astronomer, but not at Greenwich.  It was in Africa,
>> at the Cape of Good Hope observatory, at the height of the Western
>> imperial scramble to carve up the continent and its people, that
>> a white British official, working at an institution whose very walls
>> had been built by enslaved people, chose to enslave clocks themselves.
>> Over a century later, people still routinely use the racist
>> terminology of master-and-slave systems in engineering and horology,
>> yet it carries a violent weight of the imperial past, born in Africa.
>
> That said, I don't know whether the Linux terminology is directly
> linked to the timekeeping terminology.  Perhaps it's a just a
> coincidence.  I definitely hold no blame for whoever copied the terms
> that were probably industry-standard by the time.
>
> Still, as they say, now that I've seen it, I can't unsee it.
>
> So I'd like to propose that we finally consider changing it.  For
> a start, we could try with gentoo.git, as our flagship repository.
>
> Of course, there's also the matter of choosing the name.  The common
> choice at this point seems to be "main".  Other options include
> "trunk" (for Subversion fans, I suppose), "default".  However, given
> that we've been making major progress, I think we could take two steps
> at a time and choose something that actually embraces diversity.  My
> favorite, one that rhymes with "main", would be "nyan".

IMO, this is fine.  I'm personally partial to "trunk" but won't object
to any name.  The GCC project does this by having 'master' and 'trunk'
always point to the same thing and enabling pushing to both, IIRC.

That should preserve compatibility.

If it is not possible to enable push to both, having master alias trunk
but pushes be forced to trunk would also work.

Have a lovely day!
-- 
Arsen Arsenović