Code, conflict, and conduct [LWN.net]

Code, conflict, and conduct

Posted Sep 19, 2018 13:44 UTC (Wed) by Tara_Li (guest, #26706)
In reply to: Code, conflict, and conduct by k8to
Parent article: Code, conflict, and conduct

And that's *ALL* I asked for. Personally, from the outside looking in, I didn't see any issue with Linus' behavior. The developer community for the kernel seemed to still be growing, development of new features was happening apace - I didn't see any significant numbers of people throwing up their hands in public and saying "I can't work like this." But of course, like the use of handguns to prevent a crime, it's pretty much impossible to judge whether or not people were never bothering to start kernel developing because of Linus' behavior. The *only* way to judge is by comparing before and after records, which is what I asked about. Someone brought up Django, I mentioned Drupal as a highly public case where someone was kicked out of a community for activities that (as far as I could tell) didn't actually take place within that community... But I'm sure there's a lot of other communities that have switched to this and found things changed either for the better, or for the worse - or the interesting cases, where as far as anyone can tell, the community didn't change at all after adding a CoC (which would be illuminating in its own way - if the CoC doesn't change anything, why bother?)

Oh, yeah - one other case comes to mind - the recent Lerna license flap, where the CoC was actually used to kick the guy out who changed the licensing.

I'm not sure how the minor flap in the Python community over the push to eliminate the terms "master" and "slave" is going - I found it interesting to see this arise once more after the first time it came around in reference to PATA IDE drives. Still, someone was apparently pushing against it on "diversity" grounds - they have something against the BDSM community? That seems a remarkably undiverse definition of diversity.

I'm just really not sure of the utility of these things. They just seem like one more set of rules to be used to play "gotcha" with.

to post comments

Code, conflict, and conduct

Posted Sep 19, 2018 15:29 UTC (Wed) by JoeBuck (subscriber, #2330) [Link] (18 responses)

Read the recent article Trying to get STACKLEAK into the kernel for an example of Linus' behavior causing problems. People who are trying to solve tough problems in contentious areas shouldn't have deal with torrents of personal abuse, and rigorous code review does not require it.

Code, conflict, and conduct

Posted Sep 19, 2018 16:01 UTC (Wed) by Tara_Li (guest, #26706) [Link] (10 responses)

Went through it. Checked linked items. Checked linked items from the linked items. Found one case of Linus using an insult out of MP&THG (https://lwn.net/ml/linux-kernel/90a49b4e-7b17-0262-e358-2...) and Popov calling it harassment, but ... I'm not seeing it.

Code, conflict, and conduct

Posted Sep 19, 2018 18:02 UTC (Wed) by nix (subscriber, #2304) [Link] (1 responses)

I think this is a case of the seriousness of the insult depending on cultural signifiers (and knowledge of Monty Python, I guess, though it appalls me to realise that Monty Python is now a historical artifact like Shakespeare or Chaucer that not everyone can be expected to have word-perfect as a matter of course.)

Code, conflict, and conduct

Posted Sep 19, 2018 18:13 UTC (Wed) by jerojasro (guest, #98169) [Link]

One more symptom of kernel development not being a village anymore (with common culture), but a city (with people most likely without a common cultural background.)

Might look apalling, but I see it as "yay, more people collaborating".

Code, conflict, and conduct

Posted Sep 21, 2018 2:43 UTC (Fri) by jbicha (subscriber, #75043) [Link] (7 responses)

See this and this which are more abusive than simple Monty Python jokes.

Code, conflict, and conduct

Posted Sep 21, 2018 16:29 UTC (Fri) by Tara_Li (guest, #26706) [Link] (6 responses)

Mmmm - if those are considered abusive... Frankly, I don't know what to think. I just can't comprehend the standard by which those are judged abusive. Is it "That makes me feel bad?" If so... this whole discussion is abusive, because it makes me feel bad for the future of the human race.

Code, conflict, and conduct

Posted Sep 24, 2018 15:55 UTC (Mon) by JFlorian (guest, #49650) [Link] (5 responses)

I'm with you. In fact, I'd go even further and wish that more discourse was this direct and obvious. Maybe Linus lacks tact. I'm probably the same, but I *know* I often don't get the point on the receiving end simply because I'm corresponding with someone who uses too much tact.

Mean what you say, say what you mean. Be strict on your output, forgiving on your input.

Code, conflict, and conduct

Posted Sep 24, 2018 16:19 UTC (Mon) by jbicha (subscriber, #75043) [Link] (4 responses)

It's possible to be direct and obvious without cursing and without saying thing like "have you learnt *nothing*?"

I believe the last 2 sentences of this email demonstrate emotionally abusive language.

Code, conflict, and conduct

Posted Sep 27, 2018 18:30 UTC (Thu) by marmalade (guest, #127517) [Link]

Seriously? That's "emotionally abusive"?

That's not emotionally abusive, it's direct. In a huge project people need to be told "no" sometimes in a manner that lets them know it's not okay.

I've seen people in the workplace who don't learn because nobody is willing to be "mean". Sometimes you need to be "mean". Feeling bad is part of growing up and improving yourself. It's incentive. If everyone just tells you "it's okay" nothing changes, you just coast along thinking that it's okay if you mess up or do something wrong or stupid, because there's no consequence.

Negative emotions are not to be dispensed with. You *should* feel bad sometimes. It's a critical part of the human experience.

Code, conflict, and conduct

Posted Sep 27, 2018 21:08 UTC (Thu) by neilbrown (subscriber, #359) [Link] (2 responses)

Those last 2 sentences being:

> I'm, disappointed in the whole feature, but I'm also tired of having
> to go and even look for these things.
>
> Then actually *finding* them makes me just pissed off.

I think these are appropriate, not abusive. They are "I" statements. Linus. it talking about how he feels. This is always a reasonable thing to do. If hearing about Linus' feelings makes you upset, then that really is your problem, not his.

The previous sentence:

> Dammit, have you learnt *nothing*?

Isn't an "I" statement, so it would probably have been better to leave it out. However it is phrased as a question, not as an attack, so it should not be too hard to read it objectively, and respond to it accordingly.

Code, conflict, and conduct

Posted Sep 27, 2018 22:08 UTC (Thu) by apoelstra (subscriber, #75205) [Link] (1 responses)

> However it is phrased as a question, not as an attack

Honestly that line seems like one of the most hurtful things I've seen cited in these comments. The implication is clearly that Kay *hadn't* learned anything, out of negligence or stupidity or whatever.

I don't know whether Linus can distinguish hurtful things from non-hurtful things (his statement about stepping away "to learn about human emotions" suggests that he thinks he can't), but it's clear that many people in this space lack this ability (I don't mean you, just in general). It's interesting to read comments like these because helps me, and hopefully others, to see what this distinction is, because it's often unclear over text and across cultures.

It's also interesting to see mainstream reporters, whose job it is to communicate, get this wrong. For example, I think counting f-bombs or slurs is a total red herring, and I also think any derogatory terms about *code* rather than *people* are fine. (Though in fairness, the latter point is somewhat unique to open source development -- it is definitely not true in most cultures that you can criticize someone's work, even on purely technical grounds, and expect it to be non-personal.)

Code, conflict, and conduct

Posted Oct 2, 2018 6:50 UTC (Tue) by neilbrown (subscriber, #359) [Link]

> > However it is phrased as a question, not as an attack
>
> Honestly that line seems like one of the most hurtful things I've seen cited in these comments.

Thanks - that's really valuable and on-point. I genuinely didn't think that line would be very hurtful at all, but clearly not everyone shares my opinion.
This re-enforces the importance of not saying anything personal, even if I think it should be harmless (address the code, not the coder).
Hopefully it is OK to say personal things when they are positive: "You did a good job, thanks!" ????

Code, conflict, and conduct

Posted Sep 19, 2018 19:16 UTC (Wed) by rodgerd (guest, #58896) [Link] (6 responses)

Or that one-third of embedded Linux devs are afraid to try and upstream (source: https://lwn.net/Articles/647524/) their code.

Of course, the contemptuous (and contemptible) responses to that information merely reinforced its veracity.

Code, conflict, and conduct

Posted Sep 19, 2018 19:44 UTC (Wed) by mtaht (guest, #11087) [Link] (3 responses)

I went and read that link... being afraid of rejection barely made the top 7.

Reason % agreed
Developing against an older kernel 54%
Work depends on out-of-tree code 50%
It's too hard 45%
Unable to test 41%
Employer does not allow time 40%
Patch not good enough 35%
Afraid of rejection 33%

Code, conflict, and conduct

Posted Sep 21, 2018 2:12 UTC (Fri) by tedd (subscriber, #74183) [Link] (2 responses)

You don't think 1/3 is statistically significant?

Code, conflict, and conduct

Posted Sep 21, 2018 14:10 UTC (Fri) by jwarnica (subscriber, #27492) [Link] (1 responses)

It is, but then, why are they "afraid" of rejection?

"Good enough to work", "good enough to ship" and "good enough to be included upstream" are three different technical thresholds. It is quite possible that many or most people "afraid of rejection" are aware that their code isn't up to the required quality.

Code, conflict, and conduct

Posted Sep 23, 2018 3:30 UTC (Sun) by giraffedata (guest, #1954) [Link]

"Good enough to work", "good enough to ship" and "good enough to be included upstream" are three different technical thresholds.

Don't forget "good enough to be accepted by whoever's in charge of upstream," because that's not the same as "good enough to be included upstream." A potential contributor may be afraid that his perfectly good work would not be acknowledged as such. Thus, it could be a truly emotional "afraid" based on lack of validation by another human being as well as a practical "afraid" just meaning you don't want to risk wasting your time.

I had many patches I thought were great rejected by various projects - some just ignored, others explicitly rejected, which led to a decision many years ago just to keep my code to myself from now on.

Code, conflict, and conduct

Posted Sep 19, 2018 20:00 UTC (Wed) by excors (subscriber, #95769) [Link] (1 responses)

It's not clear to me that "Afraid of rejection" means they are personally worried about a rude or abusive response - if I was answering that survey, I'd interpret it as simply a rational concern that I might put a lot of effort into upstreaming my work and then it's still not good enough and gets rejected so I get no payoff for my effort. I'm just afraid of wasting my company's time. But it's hard to tell exactly what was meant without more context.

Code, conflict, and conduct

Posted Sep 21, 2018 17:34 UTC (Fri) by tbird20d (subscriber, #1901) [Link]

As the author of the survey, I can say that the survey was not given in the context of querying for abusive behavior. I do not believe it can be used as evidence of the number of people who have been abused. Here is the wording that was used:

12. Please think about a single change that you made to the Linux kernel, that you did NOT submit upstream. Please answer the following questions to describe your reasons for NOT submitting your change:

...

15. My change was not good enough to submit to upstream (e.g. it was a quick hack or workaround, not suitable for upstream).
Strongly agree, agree, neutral, disagree, strongly disagree

16. I was afraid my patch would be rejected.
Strongly agree, agree, neutral, disagree, strongly disagree