Philip Reames via llvm-dev
2016-May-05 23:30 UTC
[llvm-dev] Resuming the discussion of establishing an LLVM code of conduct
On 05/05/2016 11:42 AM, C Bergström via llvm-dev wrote:> On Fri, May 6, 2016 at 2:30 AM, Chandler Carruth <chandlerc at gmail.com> wrote: >> On Thu, May 5, 2016 at 2:55 AM C Bergström <llvm-dev at lists.llvm.org> wrote: >>> Chandler - I do not want to derail, hijack or change the topic of this >>> discussion - Would you be ok with me going into specific examples? >> >> IMO, no, I don't think that would be a productive direction. I also suspect >> it would have a high probability of (unintentionally) leading to exactly the >> kinds of situations the code of conduct is designed to prevent. > I'm on the fence if it could be productive..I'm a bit torn myself, but would lean towards the "too risky" side of things. The gain is minimal and the odds that specific situations escalate are too high. As a minimum standard, any specific situation that you do raise should be done only with the full consent of all the parties involved. It would be utterly inappropriate to raise a particular situation/example if not all parties involved wanted to draw the negative attention raising it in this thread might generate.> > LOL - How could a thought-out and detailed explanation of real world > circumstances lead to something so negative it would have to be > moderated... (??puzzled??)The problem is that discussing a particular situation brings specific people into what was previously an abstract discussion. Several of the comments made in this and previous threads have skirted the appropriateness as things stand. If those same comments had been poorly worded and seemed to apply specifically to a particular person... Well, that's likely to lead to exactly the type of personal arguments we're all hoping to avoid. Does my concern make sense?> > Under your regime - would I be forbidden from calling someone else out > for generally being a bully or troll.. Specifically if I went through > and found say 6 cases where X caused friction in the community and in > general their behavior was more noise than actually productive.Your point is directly contradicted by the current CoC proposal. Calling someone out on inappropriate behavior is absolutely appropriate. However, doing so without making it into a personal attack is important as well. "Hey, what you just said is not okay. I'm sure you didn't mean to be personal insulting, but that came across as..." - OK "You ***, how dare you say ___" -- NOT OK As an example taken from your email, your use of the work "regime" comes across as potentially loaded with negative meaning. I'd suggest that using a word like "proposal" would have been more neutral connotation wise and still made your point. Your word choice could be read to imply that you view Chandler as a authoritarian dictator which he clearly is not. :)> I'm still lost at what's really driving this.. *something* must have > happened that prompted all this..To my knowledge, there have been no specific recent incidents within the LLVM community to trigger this discussion. There have certainly been ones in other communities in recent years. A news search should find several. Philip
C Bergström via llvm-dev
2016-May-06 04:01 UTC
[llvm-dev] Resuming the discussion of establishing an LLVM code of conduct
On Fri, May 6, 2016 at 7:30 AM, Philip Reames <listmail at philipreames.com> wrote:> > > On 05/05/2016 11:42 AM, C Bergström via llvm-dev wrote: >> >> On Fri, May 6, 2016 at 2:30 AM, Chandler Carruth <chandlerc at gmail.com> >> wrote: >>> >>> On Thu, May 5, 2016 at 2:55 AM C Bergström <llvm-dev at lists.llvm.org> >>> wrote: >>>> >>>> Chandler - I do not want to derail, hijack or change the topic of this >>>> discussion - Would you be ok with me going into specific examples? >>> >>> >>> IMO, no, I don't think that would be a productive direction. I also >>> suspect >>> it would have a high probability of (unintentionally) leading to exactly >>> the >>> kinds of situations the code of conduct is designed to prevent. >> >> I'm on the fence if it could be productive.. > > I'm a bit torn myself, but would lean towards the "too risky" side of > things. The gain is minimal and the odds that specific situations escalate > are too high. As a minimum standard, any specific situation that you do > raise should be done only with the full consent of all the parties involved. > It would be utterly inappropriate to raise a particular situation/example if > not all parties involved wanted to draw the negative attention raising it in > this thread might generate. >> >> >> LOL - How could a thought-out and detailed explanation of real world >> circumstances lead to something so negative it would have to be >> moderated... (??puzzled??) > > The problem is that discussing a particular situation brings specific people > into what was previously an abstract discussion. Several of the comments > made in this and previous threads have skirted the appropriateness as things > stand. If those same comments had been poorly worded and seemed to apply > specifically to a particular person... Well, that's likely to lead to > exactly the type of personal arguments we're all hoping to avoid. > > Does my concern make sense?Yep. I thought people would be objective - you reminded me that people involved may just get upset again.>> >> >> Under your regime - would I be forbidden from calling someone else out >> for generally being a bully or troll.. Specifically if I went through >> and found say 6 cases where X caused friction in the community and in >> general their behavior was more noise than actually productive. > > Your point is directly contradicted by the current CoC proposal. > > Calling someone out on inappropriate behavior is absolutely appropriate. > However, doing so without making it into a personal attack is important as > well. > > "Hey, what you just said is not okay. I'm sure you didn't mean to be > personal insulting, but that came across as..." - OK > "You ***, how dare you say ___" -- NOT OK > > As an example taken from your email, your use of the work "regime" comes > across as potentially loaded with negative meaning. I'd suggest that using > a word like "proposal" would have been more neutral connotation wise and > still made your point. Your word choice could be read to imply that you > view Chandler as a authoritarian dictator which he clearly is not. :)Aha! Great nit and the opportunist in me can't resist... BUT I'm not going to change my choice of wording, or apologize and if it offends you, tough cookies. Now what? Under the old system - x() New Chandlers CoC - y() My proposed CoC - z() Under my proposal I hope to word this process below ------------- If you felt strongly that I had offended you, you should email the moderators or whatever their title is, explain the situation and get help to resolve the conflict. While the law of the land is stated vaguely as "common sense", is there anyone around here where that is too ambiguous or doesn't actually capture what we have today? I have read Chandler's CoC and the fact that I have questions about how it applies to different things means either my ability to read and comprehend is in question or at the least the wording should be "fixed".> >> I'm still lost at what's really driving this.. *something* must have >> happened that prompted all this.. > > To my knowledge, there have been no specific recent incidents within the > LLVM community to trigger this discussion. There have certainly been ones > in other communities in recent years. A news search should find several. > > PhilipPhilip - Thanks for your great reply and I hope you are ok with me trying to turn it into an example.
Philip Reames via llvm-dev
2016-May-06 18:04 UTC
[llvm-dev] Resuming the discussion of establishing an LLVM code of conduct
On 05/05/2016 09:01 PM, C Bergström wrote:> On Fri, May 6, 2016 at 7:30 AM, Philip Reames <listmail at philipreames.com> wrote: >>> >>> Under your regime - would I be forbidden from calling someone else out >>> for generally being a bully or troll.. Specifically if I went through >>> and found say 6 cases where X caused friction in the community and in >>> general their behavior was more noise than actually productive. >> Your point is directly contradicted by the current CoC proposal. >> >> Calling someone out on inappropriate behavior is absolutely appropriate. >> However, doing so without making it into a personal attack is important as >> well. >> >> "Hey, what you just said is not okay. I'm sure you didn't mean to be >> personal insulting, but that came across as..." - OK >> "You ***, how dare you say ___" -- NOT OK >> >> As an example taken from your email, your use of the work "regime" comes >> across as potentially loaded with negative meaning. I'd suggest that using >> a word like "proposal" would have been more neutral connotation wise and >> still made your point. Your word choice could be read to imply that you >> view Chandler as a authoritarian dictator which he clearly is not. :) > Aha! Great nit and the opportunist in me can't resist... BUT I'm not > going to change my choice of wording, or apologize and if it offends > you, tough cookies. Now what? > > Under the old system - x() > New Chandlers CoC - y() > My proposed CoC - z()My take would be "no action". If I was seriously offended (I'm not), I certainly could ask for the advisory committee to review the incident. My expectation is that any reasonable group of people we picked for that group would look at the phrasing involved, ask why it bothered me, and then possibly send you a note of the form "hey, might make sense to watch your wording a bit. This incident isn't serious, but a) you're potentially walking on thin ice with a couple of folks now, and b) repeat issues make even minor things more serious." (But better worded.) The advisory group would also send me a note of the form: "Thanks for raising the concern, but we don't see there being any need for action here. If you see a repeat serious of behavior which is problematic, please let us know, but generally a single debatebly poor choice of wording is not considered a CoC violation." Now, if your response to my comment had been worded as a personal attack (it wasn't), *that* would have been a more serious issue under the CoC as proposed. Generally, respectful disagreement is not a problem. Personal attacks on the other hand are.> > Under my proposal I hope to word this process below > ------------- > If you felt strongly that I had offended you, you should email the > moderators or whatever their title is, explain the situation and get > help to resolve the conflict. While the law of the land is stated > vaguely as "common sense", is there anyone around here where that is > too ambiguous or doesn't actually capture what we have today?I think this is exactly what I spelled out above right?> > I have read Chandler's CoC and the fact that I have questions about > how it applies to different things means either my ability to read and > comprehend is in question or at the least the wording should be > "fixed".Agreed. If there are problems with wording, please help us find them and fix them. This is definitely a case of "patches welcome". :) Being specific is really helpful as well.>>> I'm still lost at what's really driving this.. *something* must have >>> happened that prompted all this.. >> To my knowledge, there have been no specific recent incidents within the >> LLVM community to trigger this discussion. There have certainly been ones >> in other communities in recent years. A news search should find several. >> >> Philip > Philip - Thanks for your great reply and I hope you are ok with me > trying to turn it into an example.Given I deliberately wrote it to act as an example, I'm more than okay with that. :) Philip