New community focus topic: Respectful Terminology

[Edwin Kempin]

Hi,

It is important that our community is respectful and inclusive for everyone. We realised that some terms in our code base are considered offensive and we think it's important to change them. To get this done the community managers have agreed to make this their new focus topic. Since there are many aspects that need to be considered and addressed we have started a new hotlist [1] and populated it with some initial issues. Please help us with identifying further issues by adding them to the hotlist [2]. We also welcome everyone who wants to help resolve these issues. You can simply pick up issues from the hotlist and assign them to yourself. Alternatively you can also get in touch with us [3] if you are unsure where to start.

Thanks, your community managers (Marco, Matthias & Edwin)

[1] https://bugs.chromium.org/p/gerrit/issues/list?q=label%3AHotlist-Respectful-Terminology&sort=priority
[2] https://bugs.chromium.org/p/gerrit/issues/entry?template=Respectful+Terminology+Issue

[3] gerritcodereview-...@googlegroups.com

[Oswald Buddenhagen]

On Fri, Jun 26, 2020 at 02:28:35PM +0200, 'Edwin Kempin' via Repo and Gerrit Discussion wrote:
> We realised that some terms in our code base are considered offensive
> and we think it's important to change them.
>

uhhh ... let's have a look into these words, shall we?

- slave. no question about that one - it's clearly offensive (unless
one's into bdsm, heh) and there is no technical need for using it
anyway.

- master. *way* less clear-cut already, because its oppposite is
"subordinate" in its many variations, of which slave is only one. it
certainly makes sense to replace it in contexts where it was
previously used opposite to slave, but when it comes to things like
the git master branch, the enterprise seems rather questionable. i
also have my doubts that many people will stop using their MSc, MA,
and MBA titles, and i'm certainly not going to rename my world
domination master plan. ;)

- blacklist/whitelist. the dualism that underlies that metaphor
(black/white, dark/light, night/day) was used long before pale-skinned
people even emerged, so from an etymological perspective these terms
are positively *not* racist, and trying to intentionally use them in a
racist way will earn one funny looks.
from a technical perspective one can argue that terms that don't use
metaphors are better, but allowlist, blocklist, etc., just sound bad
(because they aren't established words) and can be ambiguous, too
(denylist would be better than blocklist in this regard, but just as
awkward as allowlist).

- WAR. that's an acronym. i think the bar for avoiding overloaded ones
should be "a little" higher than people's pacifist leanings.

generally, i think the matter is a very real slippery slope - i'm sure i
can come up with "offensive" words (in different languages) in any
codebase if i just go looking for them. i don't doubt that some people
are triggered by various strings (to use the most generic term), but
there is a point where the correct response is "noted, but your
objection just isn't reasonable". so ... go slowly, maybe?

[John Doe]

Nothing in gerrit as it currently is offends me or anybody else in my team. I would strongly prefer if I did not have to retrain my muscle memory to new, respectful ways.

Are these changes going to break the existing workflow?

(for everything else, there's MasterCard)

[Luca Milanesio]

> On 26 Jun 2020, at 18:27, Oswald Buddenhagen <oswald.bu...@gmx.de> wrote:
>
> On Fri, Jun 26, 2020 at 02:28:35PM +0200, 'Edwin Kempin' via Repo and Gerrit Discussion wrote:
>> We realised that some terms in our code base are considered offensive and we think it's important to change them.
> uhhh ... let's have a look into these words, shall we?
>
> - slave. no question about that one - it's clearly offensive (unless one's into bdsm, heh) and there is no technical need for using it anyway.
>
> - master. *way* less clear-cut already, because its oppposite is "subordinate" in its many variations, of which slave is only one. it certainly makes sense to replace it in contexts where it was previously used opposite to slave, but when it comes to things like the git master branch, the enterprise seems rather questionable. i also have my doubts that many people will stop using their MSc, MA, and MBA titles, and i'm certainly not going to rename my world domination master plan. ;)

In Gerrit, the difference between the ‘master’ and the ’slave’ is something we discussed many times in the past and that we could also eliminate moving forward.

At the end of the day, they are the *same code-base* with just a different set of injectors.
By merging them into a single Gerrit Instance (I guess also “Server” is offensive? That suggest we are “servant” of someone, which is offensive as well) concept, which could be R/W or R/O.
No more “replica” (aka “slave”) mode and just a read-only flag.

If you take the current Gerrit instance and add the read-only plugin, you get exactly what I mean.
We could introduce this concept already in v3.3.

>
> - blacklist/whitelist. the dualism that underlies that metaphor (black/white, dark/light, night/day) was used long before pale-skinned people even emerged, so from an etymological perspective these terms are positively *not* racist, and trying to intentionally use them in a racist way will earn one funny looks.
> from a technical perspective one can argue that terms that don't use metaphors are better, but allowlist, blocklist, etc., just sound bad (because they aren't established words) and can be ambiguous, too (denylist would be better than blocklist in this regard, but just as awkward as allowlist).

That is less of a concern though: we have limited use in the Gerrit code-base: I don’t believe anyone would complain if we rename it to “allow” and “block” lists, as we use those terms already in the ACLs.

>
> - WAR. that's an acronym. i think the bar for avoiding overloaded ones should be "a little" higher than people's pacifist leanings.

That’s actually one of those is harder to eliminate, as the computer programming was invented in war-time for missiles and bombs: non a surprise that you have “military” terms:
- war
- execute
- deploy
- kill
- terminate

We *could* change those in Gerrit, but what about the operating system?
- kill $PID => request-to-end $PID ?
- java -jar gerrit.peace?

Of course we have to define a scope of this *exercise* and, unfortunately, we cannot fix decades of offensive terminology in IT overnight, it’s just not possible.

By unifying the concept of Gerrit master and Gerrit slave into a single “Gerrit instance”, we can achieve a first goal.

Luca.

[David Pursehouse]

Maybe I missed something, but where has it been suggested to stop using the term "WAR"?

 

generally, i think the matter is a very real slippery slope - i'm sure i
can come up with "offensive" words (in different languages) in any
codebase if i just go looking for them. i don't doubt that some people
are triggered by various strings (to use the most generic term), but
there is a point where the correct response is "noted, but your
objection just isn't reasonable". so ... go slowly, maybe?

--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/20200626172751.GB648447%40ugly.

[Oswald Buddenhagen]

On Fri, Jun 26, 2020 at 10:11:48PM +0100, Luca Milanesio wrote:
>That’s actually one of those is harder to eliminate, as the computer
>programming was invented in war-time for missiles and bombs: non a
>surprise that you have “military” terms:
>- war
>- execute
>- deploy
>- kill
>- terminate
>

i've read this argument while researching for my response, and i think
it's tenuous at best. i might buy it for "deploy", but "execute", "kill"
and "terminate" are just generic metaphors, and "war" is way too recent
to even think of an association, apart from it being obviously just
coincidence (which has certainly been noted and sniggered over even
before it was enshrined by public release).
also, programming was *not* invented for military use, it's just where
it found the first practical application.

[Luca Milanesio]

I believe the proposal is for eliminating the terms “master” and “slave” only.

Luca.

 

generally, i think the matter is a very real slippery slope - i'm sure i 
can come up with "offensive" words (in different languages) in any 
codebase if i just go looking for them. i don't doubt that some people 
are triggered by various strings (to use the most generic term), but 
there is a point where the correct response is "noted, but your 
objection just isn't reasonable". so ... go slowly, maybe?

-- 
-- 
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

--- 
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/20200626172751.GB648447%40ugly.

-- 
-- 
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

--- 
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/CAM5Va3%3DZLTAKv-aFtz4AJESi19sg99dCL6-Ca6vjxf9K7_aPRg%40mail.gmail.com.

[David Ostrovsky]

On Saturday, June 27, 2020 at 4:00:00 AM UTC+2, David Pursehouse wrote:

See original discussion on this issue:[1]. There were different suggestions

what terms in the code base should be replaced. Based on this discussion

separate issues were created and:[1] was set obsolete.

[1] https://crbug.com/gerrit/12945

[David Ostrovsky]

On Saturday, June 27, 2020 at 3:37:20 PM UTC+2, lucamilanesio wrote:

On 27 Jun 2020, at 02:59, David Pursehouse <david.p...@gmail.com> wrote:

What about this suggestion "[...] so we should start with 'war'." made by you: [2]?

[2] https://bugs.chromium.org/p/gerrit/issues/detail?id=12945#c3

[Luca Milanesio]

... apologies, it was a bit of "sarcasm" ;-)

I don't believe anyone is suggesting a general cleanup of the IT language, it will get too far and too difficult.

Luca.

[Balazs Toth]

My humble opinion respect for each other should be basic, replacing expressions like master/slave or whatever won't solve any issue in real world. This is a classic situation when people push something too far believe doing something grate. My question is this step will ever solve the inequality and racism in the world? We must know what those word means and we have to learn (as humanity) why racism or suppression, not a good thing. Remove a word from the dictionary is like when you put a newspaper on the coffee got on your carpet. 

Of course, do whatever you want ;)

[3] gerritcodereview-community-mana...@googlegroups.com

[Oswald Buddenhagen]

On Sat, Jun 27, 2020 at 06:46:07PM +0100, Luca Milanesio wrote:
>> [2] https://bugs.chromium.org/p/gerrit/issues/detail?id=12945#c3
>... apologies, it was a bit of "sarcasm" ;-)
>

what exactly is sarcasm in scare quotes?

i had the strong impression that you're pulling a poe, but thought that
playing along would be more useful than accusing you of trolling (an
outright coc violation).

apart from being really bad style, your stunt was also rather pointless
- you don't prove a slippery slope by jumping right off the cliff (to
stay in the realm of metaphors).

you also discredited yourself on that issue, because you showed that
you're not taking it seriously. the question isn't whether something
should be done, as there is ample scientific evidence for the harmful
effects of specific language. the question is only whether particular
proposed changes are actually based on evidence, or lack of
understanding and conjecture.

>I don't believe anyone is suggesting a general cleanup of the IT
>language, it will get too far and too difficult.
>

"it inconveniences the privileged" is a rather poor argument against
change in itself.

[Luca Milanesio]

> On 28 Jun 2020, at 13:16, Oswald Buddenhagen <oswald.bu...@gmx.de> wrote:
>
> On Sat, Jun 27, 2020 at 06:46:07PM +0100, Luca Milanesio wrote:
>>> [2] https://bugs.chromium.org/p/gerrit/issues/detail?id=12945#c3
>> ... apologies, it was a bit of "sarcasm" ;-)
>>
> what exactly is sarcasm in scare quotes?
>
> i had the strong impression that you're pulling a poe, but thought that playing along would be more useful than accusing you of trolling (an outright coc violation).
>
> apart from being really bad style, your stunt was also rather pointless - you don't prove a slippery slope by jumping right off the cliff (to stay in the realm of metaphors).

Let me rephrase: we should put some boundaries for the respectful terminology cleanup initiative, otherwise it would become unrealistic to implement in one go.

As I mentioned, many terms in the IT industry are inappropriate, because the technology was applied in the past (and also in the present, I’m afraid) to very unpleasant applications.
With regards to master/slave terminology, I believe we can do realistically do something.

>
> you also discredited yourself on that issue, because you showed that you're not taking it seriously.

I clarified above my position.

> the question isn't whether something should be done, as there is ample scientific evidence for the harmful effects of specific language. the question is only whether particular proposed changes are actually based on evidence, or lack of understanding and conjecture.
>
>> I don't believe anyone is suggesting a general cleanup of the IT language, it will get too far and too difficult.
>>
> "it inconveniences the privileged" is a rather poor argument against change in itself.

I mentioned the example of the underlying operating system: even if we would remove any offensive or inappropriate word from the Gerrit software, we won’t be able to do the same on the underlying operating system, for instance on the very inappropriate ‘kill’ word. Also the ‘war’ word was an example of very inappropriate word that would be unrealistic to eradicate for now.

Of course I would support a future where any inappropriate, violent or offensive word is eradicated in the IT industry: I just realistically think it should be done in steps and can’t be done all at once.
Starting from eliminating ‘master’, ’slave’, ‘blacklist’, ‘whitelist’ would be a very good step in the right direction.

Hope this clarifies.

Luca.

[Patrick Georgi]

Sonntag, 28. Juni 2020 um 10:16:11 UTC+2:

My humble opinion respect for each other should be basic, replacing expressions like master/slave or whatever won't solve any issue in real world. This is a classic situation when people push something too far believe doing something grate. My question is this step will ever solve the inequality and racism in the world? We must know what those word means and we have to learn (as humanity) why racism or suppression, not a good thing. Remove a word from the dictionary is like when you put a newspaper on the coffee got on your carpet. 

It's not about removing a word from the dictionary, it's about not using problematic words where they're not required to talk about their worrisome aspect. Talking about slaves when discussing slavery (past, present or future) is different from talking about slaves when it's about computers syncing data. One is necessary, the other is not.

It's also not intended to solve all the world's problems, it's about improving one tiny aspect in a place that realistically speaking hurts nobody when fixing but hurts people when not doing so.

Patrick

[Edwin Kempin]

Thanks for all the input on this.

I agree that not all terms that have been mentioned so far are equally offensive and there should be a priority-based order in which we address them. "Slave" is probably the term which is first on the list, but we also want to seriously evaluate all other terms which are brought up as offensive. That's best done by having an individual issue for each term and then discussing it in that issue.

I do find the discussion about renaming the "master" branch important as it's not only about our project, but about us enabling projects that are hosted on Gerrit to make this change. Since renaming the default branch is very much considered in the Git community, we should support this and allow projects to use whatever default branch they want to use. This includes continuing using "master" as default branch for those projects that don't want to change this.

On Sun, Jun 28, 2020 at 8:56 PM 'Patrick Georgi' via Repo and Gerrit Discussion <repo-d...@googlegroups.com> wrote:

Sonntag, 28. Juni 2020 um 10:16:11 UTC+2:

My humble opinion respect for each other should be basic, replacing expressions like master/slave or whatever won't solve any issue in real world. This is a classic situation when people push something too far believe doing something grate. My question is this step will ever solve the inequality and racism in the world? We must know what those word means and we have to learn (as humanity) why racism or suppression, not a good thing. Remove a word from the dictionary is like when you put a newspaper on the coffee got on your carpet. 

It's not about removing a word from the dictionary, it's about not using problematic words where they're not required to talk about their worrisome aspect. Talking about slaves when discussing slavery (past, present or future) is different from talking about slaves when it's about computers syncing data. One is necessary, the other is not.

+1

 

It's also not intended to solve all the world's problems, it's about improving one tiny aspect in a place that realistically speaking hurts nobody when fixing but hurts people when not doing so.

Thanks, I agree.

 

Patrick

--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/7b4876b9-310f-4da2-bcad-ac4ca4880825n%40googlegroups.com.

[Saša Živkov]

On Fri, Jun 26, 2020 at 11:11 PM Luca Milanesio <luca.mi...@gmail.com> wrote:

> On 26 Jun 2020, at 18:27, Oswald Buddenhagen <oswald.bu...@gmx.de> wrote:
>
> On Fri, Jun 26, 2020 at 02:28:35PM +0200, 'Edwin Kempin' via Repo and Gerrit Discussion wrote:
>> We realised that some terms in our code base are considered offensive and we think it's important to change them.
> uhhh ... let's have a look into these words, shall we?
>
> - slave. no question about that one - it's clearly offensive (unless   one's into bdsm, heh) and there is no technical need for using it   anyway.
>
> - master. *way* less clear-cut already, because its oppposite is   "subordinate" in its many variations, of which slave is only one. it   certainly makes sense to replace it in contexts where it was   previously used opposite to slave, but when it comes to things like   the git master branch, the enterprise seems rather questionable. i   also have my doubts that many people will stop using their MSc, MA,   and MBA titles, and i'm certainly not going to rename my world   domination master plan. ;)

In Gerrit, the difference between the ‘master’ and the ’slave’ is something we discussed many times in the past and that we could also eliminate moving forward.

At the end of the day, they are the *same code-base* with just a different set of injectors.
By merging them into a single Gerrit Instance (I guess also “Server” is offensive? That suggest we are “servant” of someone, which is offensive as well) concept, which could be R/W or R/O.
No more “replica” (aka “slave”) mode and just a read-only flag.

If you take the current Gerrit instance and add the read-only plugin, you get exactly what I mean.
We could introduce this concept already in v3.3.

Why would renaming of slave to replica require using a plugin or, in general, require any other change in the code except renaming?

 

>
> - blacklist/whitelist. the dualism that underlies that metaphor   (black/white, dark/light, night/day) was used long before pale-skinned   people even emerged, so from an etymological perspective these terms   are positively *not* racist, and trying to intentionally use them in a   racist way will earn one funny looks.
>  from a technical perspective one can argue that terms that don't use   metaphors are better, but allowlist, blocklist, etc., just sound bad   (because they aren't established words) and can be ambiguous, too   (denylist would be better than blocklist in this regard, but just as   awkward as allowlist).

That is less of a concern though: we have limited use in the Gerrit code-base: I don’t believe anyone would complain if we rename it to “allow” and “block” lists, as we use those terms already in the ACLs.

>
> - WAR. that's an acronym. i think the bar for avoiding overloaded ones   should be "a little" higher than people's pacifist leanings.

That’s actually one of those is harder to eliminate, as the computer programming was invented in war-time for missiles and bombs: non a surprise that you have “military” terms:
- war
- execute
- deploy
- kill
- terminate

We *could* change those in Gerrit, but what about the operating system?
- kill $PID => request-to-end $PID ?
- java -jar gerrit.peace?

Of course we have to define a scope of this *exercise* and, unfortunately, we cannot fix decades of offensive terminology in IT overnight, it’s just not possible.

By unifying the concept of Gerrit master and Gerrit slave into a single “Gerrit instance”, we can achieve a first goal.

Luca.

--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/B8125A40-FADD-4ECB-BBF1-2BEFEBA1B5C9%40gmail.com.

[Luca Milanesio]

On 29 Jun 2020, at 08:34, Saša Živkov <ziv...@gmail.com> wrote:

On Fri, Jun 26, 2020 at 11:11 PM Luca Milanesio <luca.mi...@gmail.com> wrote:

> On 26 Jun 2020, at 18:27, Oswald Buddenhagen <oswald.bu...@gmx.de> wrote:
> 
> On Fri, Jun 26, 2020 at 02:28:35PM +0200, 'Edwin Kempin' via Repo and Gerrit Discussion wrote:
>> We realised that some terms in our code base are considered offensive and we think it's important to change them.
> uhhh ... let's have a look into these words, shall we?
> 
> - slave. no question about that one - it's clearly offensive (unless   one's into bdsm, heh) and there is no technical need for using it   anyway.
> 
> - master. *way* less clear-cut already, because its oppposite is   "subordinate" in its many variations, of which slave is only one. it   certainly makes sense to replace it in contexts where it was   previously used opposite to slave, but when it comes to things like   the git master branch, the enterprise seems rather questionable. i   also have my doubts that many people will stop using their MSc, MA,   and MBA titles, and i'm certainly not going to rename my world   domination master plan. ;)

In Gerrit, the difference between the ‘master’ and the ’slave’ is something we discussed many times in the past and that we could also eliminate moving forward.

At the end of the day, they are the *same code-base* with just a different set of injectors.
By merging them into a single Gerrit Instance (I guess also “Server” is offensive? That suggest we are “servant” of someone, which is offensive as well) concept, which could be R/W or R/O.
No more “replica” (aka “slave”) mode and just a read-only flag.

If you take the current Gerrit instance and add the read-only plugin, you get exactly what I mean.
We could introduce this concept already in v3.3.

Why would renaming of slave to replica require using a plugin or, in general, require any other change in the code except renaming?

You are right in terms of the rename: it has happened already and is not directly related to the RO/RW topic.

However, ‘replica’ becomes ambiguous when you move into a clustering scenario: is the ‘replica’ just one of the Gerrit RW instances? Or is the ‘replica’ one of the ones used to be called ’slave’?

Hence, my suggestion to just eliminate the difference between the two and bring into core the concept of RW and RO.

Once the RO instance instance concept is brought into core, we don’t need any plugin.

Today, the Gerrit ‘replica’ is a read-only headless instance.

Hope this clarifies.

Luca.

[Saša Živkov]

On Mon, Jun 29, 2020 at 9:40 AM Luca Milanesio <luca.mi...@gmail.com> wrote:

On 29 Jun 2020, at 08:34, Saša Živkov <ziv...@gmail.com> wrote:

On Fri, Jun 26, 2020 at 11:11 PM Luca Milanesio <luca.mi...@gmail.com> wrote:

> On 26 Jun 2020, at 18:27, Oswald Buddenhagen <oswald.bu...@gmx.de> wrote:
> 
> On Fri, Jun 26, 2020 at 02:28:35PM +0200, 'Edwin Kempin' via Repo and Gerrit Discussion wrote:
>> We realised that some terms in our code base are considered offensive and we think it's important to change them.
> uhhh ... let's have a look into these words, shall we?
> 
> - slave. no question about that one - it's clearly offensive (unless   one's into bdsm, heh) and there is no technical need for using it   anyway.
> 
> - master. *way* less clear-cut already, because its oppposite is   "subordinate" in its many variations, of which slave is only one. it   certainly makes sense to replace it in contexts where it was   previously used opposite to slave, but when it comes to things like   the git master branch, the enterprise seems rather questionable. i   also have my doubts that many people will stop using their MSc, MA,   and MBA titles, and i'm certainly not going to rename my world   domination master plan. ;)

In Gerrit, the difference between the ‘master’ and the ’slave’ is something we discussed many times in the past and that we could also eliminate moving forward.

At the end of the day, they are the *same code-base* with just a different set of injectors.
By merging them into a single Gerrit Instance (I guess also “Server” is offensive? That suggest we are “servant” of someone, which is offensive as well) concept, which could be R/W or R/O.
No more “replica” (aka “slave”) mode and just a read-only flag.

If you take the current Gerrit instance and add the read-only plugin, you get exactly what I mean.
We could introduce this concept already in v3.3.

Why would renaming of slave to replica require using a plugin or, in general, require any other change in the code except renaming?

You are right in terms of the rename: it has happened already and is not directly related to the RO/RW topic.

However, ‘replica’ becomes ambiguous when you move into a clustering scenario: is the ‘replica’ just one of the Gerrit RW instances? Or is the ‘replica’ one of the ones used to be called ’slave’?

OK, maybe we need a better name than "replica" but still it is only about a rename.

 

Hence, my suggestion to just eliminate the difference between the two and bring into core the concept of RW and RO.

The RO mode in the current slave mode is achieved by not even starting or enabling some functionalities like: receive-pack, REST API, etc...

I guess that the read-only plugin just rejects all write requests but the Gerrit core would still serve them if the read-only plugin would unload or fail to recognize a read-only request and reject it.

The former approach is a more robust way of achieving read-only-ness than the latter. But this discussion (except for the terminology) is off-topic here.

[Luca Milanesio]

On 29 Jun 2020, at 09:01, Saša Živkov <ziv...@gmail.com> wrote:

On Mon, Jun 29, 2020 at 9:40 AM Luca Milanesio <luca.mi...@gmail.com> wrote:

On 29 Jun 2020, at 08:34, Saša Živkov <ziv...@gmail.com> wrote:

On Fri, Jun 26, 2020 at 11:11 PM Luca Milanesio <luca.mi...@gmail.com> wrote:

> On 26 Jun 2020, at 18:27, Oswald Buddenhagen <oswald.bu...@gmx.de> wrote:
> 
> On Fri, Jun 26, 2020 at 02:28:35PM +0200, 'Edwin Kempin' via Repo and Gerrit Discussion wrote:
>> We realised that some terms in our code base are considered offensive and we think it's important to change them.
> uhhh ... let's have a look into these words, shall we?
> 
> - slave. no question about that one - it's clearly offensive (unless   one's into bdsm, heh) and there is no technical need for using it   anyway.
> 
> - master. *way* less clear-cut already, because its oppposite is   "subordinate" in its many variations, of which slave is only one. it   certainly makes sense to replace it in contexts where it was   previously used opposite to slave, but when it comes to things like   the git master branch, the enterprise seems rather questionable. i   also have my doubts that many people will stop using their MSc, MA,   and MBA titles, and i'm certainly not going to rename my world   domination master plan. ;)

In Gerrit, the difference between the ‘master’ and the ’slave’ is something we discussed many times in the past and that we could also eliminate moving forward.

At the end of the day, they are the *same code-base* with just a different set of injectors.
By merging them into a single Gerrit Instance (I guess also “Server” is offensive? That suggest we are “servant” of someone, which is offensive as well) concept, which could be R/W or R/O.
No more “replica” (aka “slave”) mode and just a read-only flag.

If you take the current Gerrit instance and add the read-only plugin, you get exactly what I mean.
We could introduce this concept already in v3.3.

Why would renaming of slave to replica require using a plugin or, in general, require any other change in the code except renaming?

You are right in terms of the rename: it has happened already and is not directly related to the RO/RW topic.

However, ‘replica’ becomes ambiguous when you move into a clustering scenario: is the ‘replica’ just one of the Gerrit RW instances? Or is the ‘replica’ one of the ones used to be called ’slave’?

OK, maybe we need a better name than "replica" but still it is only about a rename.

 

Hence, my suggestion to just eliminate the difference between the two and bring into core the concept of RW and RO.

The RO mode in the current slave mode is achieved by not even starting or enabling some functionalities like: receive-pack, REST API, etc...

I guess that the read-only plugin just rejects all write requests but the Gerrit core would still serve them if the read-only plugin would unload or fail to recognize a read-only request and reject it.

The former approach is a more robust way of achieving read-only-ness than the latter. But this discussion (except for the terminology) is off-topic here.

Indeed, will move the discussion to the relevant issue.

Luca.

[jonas....@meraki.net]

Here's a small improvement in this area: https://gerrit-review.googlesource.com/c/gerrit/+/276192/1

While I'm certainly not an active member of the gerrit community, I would like to chime in on the point which has been raised by several folks in this thread.

Retraining muscle memory is hard. As a white male, I also find it hard to see the world from the perspective of those in groups which haven't historically had power.

That doesn't mean I shouldn't try.

-- Jonas

[Oswald Buddenhagen]

On Thu, Jul 23, 2020 at 05:38:02PM -0700, 'jonas....@meraki.net' via Repo and Gerrit Discussion wrote:
>As a white male, I also find it hard to see the world from the
>perspective of those in groups which haven't historically had power.
>

conversely, that doesn't mean that every proposed "fix" actually makes
sense.

when language changes are asked for based on clearly bogus (ahistorical)
perceptions of reality (like that 'blacklist' has racist roots), that
undermines the credibility of the whole endeavor. one can make a
stronger argument against 'blacklist' by pointing to flawed human
psychology (i.e., facts don't matter - perception does). however,
ultimately, the exercise will remain pointless: the bad-is-black effect
will certainly remain even if the language is purged of related
metaphors, because it is primordial.

what remains in these vague cases is a *technical* argument against
using metaphors in the first place - and providing an actually sensible
replacement and a migration path on a case-by-case basis (i've already
seen a case where an outside activist clearly just didn't care about the
practical side of it).

[anna.fr...@gmail.com]

But it is not bogus. People are offended by using these words. If the word slave is neutral to you, it shouldn't matter that we switch to a word neutral for everyone?

For myself I don't connect the word slave to only BLM, but to the thousands of slaves we currently have in the world. To not only use the word when we are speaking about humans that are owned by other people weakens the word. To let the word slave only mean what it originally meant empathizes the word, makes the word matter again.

BR, Anna 

[Oswald Buddenhagen]

On Fri, Jul 24, 2020 at 06:38:07AM -0700, anna.fr...@gmail.com wrote:
> If the word slave [...]
>
i suggest you re-read what i actually wrote, preferably also the whole
thread from the beginning.

[Ben Rohlfs]

Thank you Anna, very well said. -Ben

--

--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/bd132ffa-f925-4284-97a6-6c4eb16543bdn%40googlegroups.com.

[Luca Milanesio]

On 28 Jul 2020, at 11:18, 'Ben Rohlfs' via Repo and Gerrit Discussion <repo-d...@googlegroups.com> wrote:

Thank you Anna, very well said. -Ben

On Tue, Jul 28, 2020 at 10:05 AM anna.fr...@gmail.com <anna.fr...@gmail.com> wrote:

But it is not bogus. People are offended by using these words. If the word slave is neutral to you, it shouldn't matter that we switch to a word neutral for everyone?

As a matter of fact, we have already abandoned that term in Gerrit.

I don’t believe anyone wants to reintroduce it.

Luca.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/CAEWS%2BaPp9wo_D5zbxSmMHcgQ6SvPkJqG-r%3DcDRthobTd-%2BUShA%40mail.gmail.com.

[David Ostrovsky]

Am Dienstag, 28. Juli 2020 11:24:05 UTC+2 schrieb lucamilanesio:

On 28 Jul 2020, at 11:18, 'Ben Rohlfs' via Repo and Gerrit Discussion <repo-d...@googlegroups.com> wrote:

Thank you Anna, very well said. -Ben

On Tue, Jul 28, 2020 at 10:05 AM anna.fr...@gmail.com <anna.f...@gmail.com> wrote:

But it is not bogus. People are offended by using these words. If the word slave is neutral to you, it shouldn't matter that we switch to a word neutral for everyone?

As a matter of fact, we have already abandoned that term in Gerrit.

Looking at: [1] it is still open. Code search still returns a lot of places: [2].

And primary documentation is still using it, e.g.: [3].

[1] https://crbug.com/gerrit/13013

[2] https://cs.bazel.build/search?q=r%3Agerrit+slave

[3] https://gerrit-review.googlesource.com/Documentation/config-gerrit.html#container.slave

[Luca Milanesio]

On 28 Jul 2020, at 12:50, David Ostrovsky <david.o...@gmail.com> wrote:

Am Dienstag, 28. Juli 2020 11:24:05 UTC+2 schrieb lucamilanesio:

On 28 Jul 2020, at 11:18, 'Ben Rohlfs' via Repo and Gerrit Discussion <repo-d...@googlegroups.com> wrote:

Thank you Anna, very well said. -Ben

On Tue, Jul 28, 2020 at 10:05 AM anna.fr...@gmail.com <anna.f...@gmail.com> wrote:

But it is not bogus. People are offended by using these words. If the word slave is neutral to you, it shouldn't matter that we switch to a word neutral for everyone?

As a matter of fact, we have already abandoned that term in Gerrit.

Looking at: [1] it is still open. Code search still returns a lot of places: [2].

With regards to [2], many places where it is used are NOT in Gerrit core but in plugins and other associated projects.

Yes, it’s still a WIP on the eco-system. Every project owner (including me for my projects) has still to act on it.

Thomas has introduced a new annotated injection that should return the fact of running in a Gerrit replica, which should allow to remove the banned keyword in many more places, whilst keeping backward compatibility.

We will still need to keep it in 1 or 2 places in Gerrit, for backward compatibility.

And primary documentation is still using it, e.g.: [3].

Yes, for backward compatibility, we need to mention it.

We need to keep the terms respectful, but we need also to respect the integrity of the current Gerrit setup.

Breaking eveyone’s production system would be disrespectful for our users also.

We need to allow them to migrate to the new term gradually, without interruption.

Luca.

[1] https://crbug.com/gerrit/13013

[2] https://cs.bazel.build/search?q=r%3Agerrit+slave

[3] https://gerrit-review.googlesource.com/Documentation/config-gerrit.html#container.slave

--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/b5266685-60e5-4899-b05e-c132f2da7791o%40googlegroups.com.

[David Ostrovsky]

Am Dienstag, 28. Juli 2020 12:56:41 UTC+2 schrieb lucamilanesio:

On 28 Jul 2020, at 12:50, David Ostrovsky <david.o...@gmail.com> wrote:

Am Dienstag, 28. Juli 2020 11:24:05 UTC+2 schrieb lucamilanesio:

On 28 Jul 2020, at 11:18, 'Ben Rohlfs' via Repo and Gerrit Discussion <repo-d...@googlegroups.com> wrote:

Thank you Anna, very well said. -Ben

On Tue, Jul 28, 2020 at 10:05 AM anna.fr...@gmail.com <anna.f...@gmail.com> wrote:

But it is not bogus. People are offended by using these words. If the word slave is neutral to you, it shouldn't matter that we switch to a word neutral for everyone?

As a matter of fact, we have already abandoned that term in Gerrit.

Looking at: [1] it is still open. Code search still returns a lot of places: [2].

With regards to [2], many places where it is used are NOT in Gerrit core but in plugins and other associated projects.

Yes, it’s still a WIP on the eco-system. Every project owner (including me for my projects) has still to act on it.

Thomas has introduced a new annotated injection that should return the fact of running in a Gerrit replica, which should allow to remove the banned keyword in many more places, whilst keeping backward compatibility.

We will still need to keep it in 1 or 2 places in Gerrit, for backward compatibility.

And primary documentation is still using it, e.g.: [3].

Yes, for backward compatibility, we need to mention it.

We need to keep the terms respectful, but we need also to respect the integrity of the current Gerrit setup.

Breaking eveyone’s production system would be disrespectful for our users also.

I fail to see how wiping out this section: [3] from gerrit documentation

would "break eveyone’s production system".

[Luca Milanesio]

I mean, the documentation and the legacy setting :-)

I agree that we can drop it altogether (docs and code) from v3.3 onwards.

Luca.

--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/repo-discuss/6cc897e2-209e-4bc0-b9a9-c60995243291o%40googlegroups.com.

[Matthias Sohn]

I think it should be documented somewhere how to migrate from the old name to the new option name

when upgrading  from an older version. Maybe we could demote documentation of this legacy option

from primary configuration documentation [3] to migration instructions in the relevant release notes. 

-Matthias