Both of these approaches are interesting.
Since this goes past what we want to do for just ourselves, I've renamed
the thread. I just want to remind everyone in this discussion that the
D&I group (in whatever official form it takes) can't set policy for the
ASF - we can only provide recommendations to be acted upon by leadership.
Even within that framework, I am a big fan of RFC 2119 terminology[1]
for our recommendations (that end up in front of the BoD/officers and
may eventually become policy, procedure, or general recommendations).
You'll recognize this as the "MUST," "SHOULD," "COULD," etc. approach to
standards and requirements.
I see the two options being presented (on-list, separate-list) both as
"COULD"s - I don't yet see a convincing argument for making either a
"SHOULD" over the other. But that doesn't mean we couldn't provide
guidance, such as:
"Big communities, or those with a substantial non-English membership
(such as our projects heavily active and originating in China like
CarbonData, Eagle, Griffin, HAWQ, Kylin, ROcketMQ, ServiceComb and
Skywalking) may opt for the Debian-like approach of a separate ML, where
critical discussion is diverted back to dev@ for English. Smaller
projects with less traffic on their dev@ lists, or who tend to use only
a single list for communication, may prefer to keep all threads, English
or not, on the same list."
Thoughts?
I'm finally excited about this initiative again. This discussion is
already bearing fruit that goes past the context of a FAQ and into what
could become part of the first preliminary report back to the Board.
-Joan "yay" Touzet
[1]: https://tools.ietf.org/html/rfc2119
On 2019-05-14 1:14 p.m., Jose Miguel Parrella Romero wrote:
In Debian, there are non-English mailing lists for user support, developer
discussions, i18n/l10n and news (think press)
The volume is rather small and threads that can potentially change the
direction of the project are invariably in English (whether in the bug tracking
system, mailing lists, IRC, etc.), but the infrastructure exists and at least
for i18n/l10n and news I can say there's an operational dependency on it. Of
course outside of Debian-hosted infrastructure, there are local Debian user
groups, meetups, conferences, etc.
I recently heard an anecdote from Chris Lamb (former Debian Project Leader) where he
mentioned that after a non-English speaker was repeatedly using the word
"abysmal" to describe the performance of a service (creating friction with the
developers and operators) they chose what I would describe as a probing/educational
communication style, where regardless of whether the performance of the service was
underwhelming or not, the nuances of the concept were written back to the original poster
to the point where either it was a personal attack (rare) or a different word emanated as
alternative.
PS: also sharing some resources from Debian's Anti-Harrassment team Wiki page
[1]
[1] https://wiki.debian.org/AntiHarassment
________________________________________
From: Alex Harui <aha...@adobe.com.INVALID>
Sent: Tuesday, May 14, 2019 9:46 AM
To: diversity@apache.org
Subject: Re: [FAQ] Proto-FAQ question 1
IMO, we can come up with a way to support non-English official communication
channels. We have to try to make it clear to as many people as possible that
it is ok to write an email in your native language on dev@ and users@. We
could decide that private@ is the air-traffic-controller channel and must be
English. Hopefully the kinds of discussions on private@ are limited enough
that people on that list can learn enough functional English to participate.
Most are coders using English keywords in programming languages.
When I see a post in non-English on the lists of the projects I participate in, I shove
it into Google Translate. It works well enough to give me an idea of what the post is
about. If I thought it was important, I would copy the English translation as a reply to
the post saying "Google Translate said you are saying: ". Sometimes, another
person on the lists who is fluent in the original poster's language will offer a
translation as well as discuss with the OP in that OP's language. And that's all good to
me.
I do not think we want to have language-specific channels any more than we
would tell some group they must use English when having a conversation in a
public space. That's one analogy I use for the ASF: that we are basically
working from a bench in the town square where everyone can listen in on our
conversations and I can listen in on others. We should not start from the fear
that some conversation that we cannot understand is going to have negative
consequences. Some will, for sure, but it should be ok to ask for a short
summary in English. We can make it a recommended or required practice for the
PMC members to try to provide those short summaries, even if it is via Google
Translate.
Again, a restaurant analogy. If some tourists come into your restaurant, use
their limited English to order food, then have a conversation in their native
language while eating, that should be allowed, and the astute restaurant worker
will still try to pick up on whether they are enjoying the food and the
restaurant in general. IMO, Google Translate has worked well enough for me to
pick up on non-English conversations on the lists and participate. I would not
want to have to monitor other language specific lists. I would encourage
everyone on our lists to use Google Translate to get a sense of the
conversations in other languages. I will be super happy if my project's users@
lists are a mix of languages.
My 2 cents,
-Alex
On 5/14/19, 9:19 AM, "Joan Touzet" <woh...@apache.org> wrote:
This is something that's been on my mind a lot recently.
We can't come up with a way to support non-English official
communication channels for the Foundation as Bertrand mentions, but if
we don't allow non-official channels in other languages, as the saying
goes, "the Internet will route around us" and just create them outside
of the ASF.
I, for one, would rather see those discussions happening on ASF lists if
at all possible. Maybe we can't have a dev-chinese@project.a.o, but
couldn't we at least support a users-chinese@project.a.o?
I think it's equally important to consider the i18n and l10n of the
software projects that we shepherd. This is something I want to ask the
assembled minds in the board/officers at the face-to-face meeting this
week.
I know there are some woefully under-recognized solutions within the ASF
(such as our Pootle instance,
https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Ftranslate.apache.org%2F&data=02%7C01%7C%7Cda3f0dd6e5d34047c12408d6d88bbd79%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636934492017566418&sdata=Y1N%2FdjHRKBSqSrGZns9QxLSJenscUnv4uUhvW4IS1H0%3D&reserved=0).
Look how
few projects even try to use it. This is tragic. And doing what we can
to improve things here I feel does fall under D&I. It even meshes with
the US non-profit mission of software for/in the public good, since the
US does not have an official national language, and has a large
percentage of citizens and residents who speak other languages.
On the intake side, we should consider translation of the FAQ and other
documents into a few languages, with reverse translations back to
English to ensure nuance is not lost (especially for critical documents
like the CoC, general project bylaws, and maybe some of "The Apache Way"
stuff.) These documents can clearly be marked as "not official" so as to
not overburden our thin legal resources to date.
-Joan
On 2019-05-14 12:03 p.m., Alex Harui wrote:
> It has been interesting to me how important words are in these discussions. In this case, the word
"fix". We may not be able to "fix" as in "eliminate" the requirement to communicate in
English, but I believe we can and should encourage communities to choose their words more carefully so that
"fluency" is not perceived as a requirement. I already do filter my choice of words to try to avoid word
patterns that non-US citizens may not understand since all but one of the most active committers on my project live
outside the US.
>
> We might come up with other social suggestions like allowing posts in
other languages on users@ lists, allowing posts in other languages on dev@ but
recommending translation of the original post to English and requiring posts in
English on private@. That's sort of how Flex and Royale operate today without any
official documentation of this approach.
>
> My observation is that restaurant workers in the US in popular tourist
areas are generally trained to be patient with non-English-speaking customers and
also choose their words carefully. Why shouldn't we all take that approach at
Apache? Another observation is that airline pilots fly to the US from all over
the world and communicate in English to flight controllers in the US (and other
countries too, iIRC). I'm fairly certain the US-based flight controllers choose
their words carefully. I do not think all of those pilots from other countries
are fluent in English. They have learned enough functional English to communicate
on how to navigate and land an airplane. Does/should software development at the
ASF require an even larger English vocabulary than navigating air-space?
>
> My 2 cents,
> -Alex
>
> On 5/14/19, 6:27 AM, "Bertrand Delacretaz" <bdelacre...@apache.org>
wrote:
>
> On Tue, May 14, 2019 at 3:09 PM Daniel Gruno <humbed...@apache.org>
wrote:
> > ...It's like democracy - while not perfect, it is (I would say)
the least
> > sucky way of ensuring fairness, openness and proper governance,
as it is
> > the most universal (and de facto standard) language we have..
>
> +1
>
> > I do not see the need for or lack of English skills as a thing
*we* can fix...
>
> Same here, I was mentioning as one invariant that does affect who
> joins our communities - or more precisely the way people contribute.
>
> -Bertrand
>
>
---------------------------------------------------------------------
> To unsubscribe, e-mail: diversity-unsubscr...@apache.org
> For additional commands, e-mail: diversity-h...@apache.org
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: diversity-unsubscr...@apache.org
> For additional commands, e-mail: diversity-h...@apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: diversity-unsubscr...@apache.org
For additional commands, e-mail: diversity-h...@apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: diversity-unsubscr...@apache.org
For additional commands, e-mail: diversity-h...@apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: diversity-unsubscr...@apache.org
For additional commands, e-mail: diversity-h...@apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: diversity-unsubscr...@apache.org
For additional commands, e-mail: diversity-h...@apache.org
