Sharp: Closing a door [LWN.net]

Sharp: Closing a door

Posted Oct 5, 2015 22:21 UTC (Mon) by pizza (subscriber, #46)
In reply to: Sharp: Closing a door by rahulsundaram
Parent article: Sharp: Closing a door

> Not necessarily. It might be make you more tone deaf or less sympathetic when people complain about the development environment and even turn against the person raising the issue as opposed to changing the environment to be more suitable for everyone involved.

I hate to say it, but the ones seeking to impose a change bear the burden of demonstrating that said change would be worthwhile. (Or forcing said change, in extreme cases).

(See, that's the problem with sociology as a field of study; there are laws that prevent performing directed, controlled experiments so all you can do is analyze uncontrolled environments after the fact)

to post comments

Sharp: Closing a door

Posted Oct 6, 2015 0:36 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] (17 responses)

>I hate to say it, but the ones seeking to impose a change bear the burden of demonstrating that said change would be worthwhile.

That would be true if this was a new issue. The question of poor tone in LKML has been brought up multiple times by several kernel developers and several people have said that they have either left or not joined kernel development because of said environment or they are working on it despite it and not because of it.

Projects that are generally less abrasive welcome and sustain more diverse contributions. That has been demonstrated many times. Rust and KDE were mentioned as examples in this thread but one could easily cite more projects that do it very well.

Sharp: Closing a door

Posted Oct 6, 2015 3:09 UTC (Tue) by pizza (subscriber, #46) [Link] (16 responses)

>>I hate to say it, but the ones seeking to impose a change bear the burden of demonstrating that said change would be worthwhile.
> That would be true if this was a new issue. The question of poor tone in LKML has been brought up multiple times by several kernel developers and several people have said that they have either left or not joined kernel development because of said environment or they are working on it despite it and not because of it.

Be that as it may, would said change actually be *worthwhile* to the Kernel as a whole?

> Projects that are generally less abrasive welcome and sustain more diverse contributions. That has been demonstrated many times. Rust and KDE were mentioned as examples in this thread but one could easily cite more projects that do it very well.

I don't deny that a less abrasive environment would be be better for specific individuals, which in turn could indirectly lead to "more diverse contributions", but would said contributions actually result in an objectively better kernel?

...It's a legitimate question (and if we're being honest, an unanswerable one)

I generally take the position that diversity is a GoodThing(tm) -- But it's foolish to assume that "change this one thing" won't have additional side-effects beyond the ones you're looking for.

Personally, I don't find the LKML environment appealing -- not because of the occasional flaming (I've been the indirect recipient of it twice, incidentally) but because it's a frantic, high-pressure environment, and much like Ms. Sharp, I don't need that sort of stress in my life. Unfortunately, I'm of the growing opinion that these hostile environments are endemic to the profession of software development. In the past fifteen years I've been doing this, I've only seen it get worse and I look forward to the day when I can leave it all behind.

Sharp: Closing a door

Posted Oct 6, 2015 3:42 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] (2 responses)

>Be that as it may, would said change actually be *worthwhile* to the Kernel as a whole?

Very likely, yes. We have several specific examples of talented kernel developers leaving the kernel because of the tone in the kernel development community. Just avoiding that would have already resulted in a better project. We can also demonstrate that similar approaches in other open source projects has been positive. We can't demonstrate that it would beneficial to the kernel specifically without making the change. That is true of any changes. Even technical ones. For example: can you demonstrate that testing would result in better software for any given new software without any tests? Very likely, yes but no guarantee. Now social changes are harder than technical changes but the same underlying principles apply.

Sharp: Closing a door

Posted Oct 8, 2015 15:59 UTC (Thu) by alankila (guest, #47141) [Link] (1 responses)

> Just avoiding that would have already resulted in a better project.

Provided no other adverse outcomes occurred as a result of how this was achieved. It is unlikely that we could e.g. institute a code of conduct on LKML, and a "police force" which can enforce it without also impacting it negatively somehow, e.g. each time this police force steps in and enforces a CoC, it might apply quite mindless set of criteria to determine if CoC is violated followed by a process of escalation that eventually would remove anyone from the kernel development community, up to Linus himself. This is sort of bureaucracy and enforced niceness is always stifling in some sense even if its intentions are good.

> For example: can you demonstrate that testing would result in better software for any given new software without any tests?

Possibly, but writing the tests takes time that would be removed from other development. Maybe the quality is better, but functionality is less.

There is no free lunch. I think these repeated discussions about tone of the LKML are also a harm per se because they are also likely to scare off any new developers because they now expect to be flamed and have their mother and pet dog insulted, regardless of whether this actually would happen to them.

Sharp: Closing a door

Posted Oct 8, 2015 16:36 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]

>There is no free lunch. I think these repeated discussions about tone of the LKML are also a harm per se

If you say pollute a river, the harm is because of the pollution and not because a news outlet publishes the story.

Sharp: Closing a door

Posted Oct 6, 2015 9:43 UTC (Tue) by micka (subscriber, #38720) [Link]

I feel we're a at this point now : what if we made LKML a better place for nothing ?

(stealing from http://www.gocomics.com/joelpett/2009/12/13/)

Sharp: Closing a door

Posted Oct 6, 2015 18:23 UTC (Tue) by dskoll (subscriber, #1630) [Link] (11 responses)

Be that as it may, would said change actually be *worthwhile* to the Kernel as a whole?

That is the wrong question. If the Kernel is the ultimate good against which all other goods are subordinate, then we've lost. If that were the case, it would be possible to justify any sort of behavior up to and including genocide as long as it improved the Kernel.

What we need to realize is that people are more important than anything else and sometimes changes are needed because they're the right thing to do, and not because they can be proven to improve the Kernel.

Sharp: Closing a door

Posted Oct 6, 2015 19:58 UTC (Tue) by shmget (guest, #58347) [Link] (10 responses)

" sometimes changes are needed because they're the right thing to do, "

But that is not the case here.

Sharp: Closing a door

Posted Oct 6, 2015 20:21 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] (9 responses)

>it is obviously _not_ better for _everyone_ involved.

>But that is not the case here.

and so on. It would help to explain your perspective rather than just assert.

Sharp: Closing a door

Posted Oct 6, 2015 21:12 UTC (Tue) by shmget (guest, #58347) [Link] (8 responses)

" It would help to explain your perspective rather than just assert. "

a story about pots and kettles come to mind... the point here was to counter the often asserted claim of 'right thing'

but if you want more details.. Linus already expressed thing quite adequately
in a thread already posted here.

http://marc.info/?l=linux-kernel&m=137392506516022&...

fell free to navigate in the thread in question, most of the point and counter point have been made there already....

If that did not give a hint that what some people here claim is 'The Right Thing(tm)', may not be that a slam dunk, I do not think I can convince you otherwise.

If you care about the more fundamental rooting of my position. In this topic it boils down to my strongly held belief that 'the right not to be offended' is a moronic, impossible to apply concept.

Sharp: Closing a door

Posted Oct 6, 2015 23:51 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] (2 responses)

>If that did not give a hint that what some people here claim is 'The Right Thing(tm)', may not be that a slam dunk, I do not think I can convince you otherwise.

Yes because Linus approach is precisely what some people here disagree with. Pointing to him is not going to convince them.

>In this topic it boils down to my strongly held belief that 'the right not to be offended' is a moronic, impossible to apply concept.

True but irrelevant since noone here has claimed otherwise.

Sharp: Closing a door

Posted Oct 7, 2015 8:02 UTC (Wed) by shmget (guest, #58347) [Link] (1 responses)

"Yes because Linus approach is precisely what some people here disagree with. Pointing to him is not going to convince them."

Sure, you can decide to ignore the argument, but then you cannot complain that such argument is not given.

I certainly read that 'some people' disagree with it, that does not means that it is not correct, or that the opposing view is ipso facto the 'Right Thing(tm)'

"True but irrelevant since noone here has claimed otherwise."
Of course that has been claimed, many, many times in this debate. It is the core foundation of the argument that speech mush be judged by each and every witness to the speech... not only the one it is addressed too but also real or hypothetical bystander. and that judgement must be based on the level of offense claimed by each and every one of them.
and merely 'feeling uncomfortable' is standing enough to have a cause of action.... iow a 'right not to be offended'
you can tip-toe around the wording as much as you like, but it is still what it is.

An analogy to this ludicrous thing is these people that leave the city to get a house in farmland, and then sue for nuisance the neighboring farmers, because their cows block the road, the roster sing at dawn, the stench in the autumn when the farmer spread manure in the fields, etc... you knew or should have known that this was the reality of farmland...

If you really love so much the hypocritical, passive-aggressive political correctness of Corporate America.. good for you... you already have a huge playing field... stay there and do not try to ruin it for the rest of us.

Sharp: Closing a door

Posted Oct 7, 2015 12:02 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]

I am not ignoring it but letting you know why it is not convincing to point to linus when asked your perspective.

Also you are setting up a strawman. I will stick to the points in this thread.

Sharp: Closing a door

Posted Oct 7, 2015 2:51 UTC (Wed) by dskoll (subscriber, #1630) [Link] (4 responses)

There is no right not to be offended. Such a "right" would itself be offensive because it would stifle free speech.

That does not mean it's OK to be crude and offensive. It would not be tolerated in a work environment---a perpetrator would justifiably be fired for creating a toxic workplace---and it should not be tolerated in an open-source development environment. Especially when there's a power imbalance, aggressive behavior can cause real damage.

It's far too convenient for people to behave disgustingly and then say "well, you're the one getting offended so it's your problem." Nope, I don't buy it and neither do most civilized people.

Sharp: Closing a door

Posted Oct 9, 2015 17:48 UTC (Fri) by Wol (subscriber, #4433) [Link] (3 responses)

> There is no right not to be offended. Such a "right" would itself be offensive because it would stifle free speech.

I hate all this stuff about "rights". Your "rights" probably infringe my "rights".

Far better - it is your DUTY as civilised person, not to be offensive. As a CONSEQUENCE, you are unlikely to be offended.

I don't like "do as you would be done by" - the obvious problem with this is that lkml, and lwn, and places like that are mostly an "old boys club" and us men are only too happy to treat everyone else here as a bloke - then we wonder why people like Ms Sharp and dskoll find the place very hostile and uninviting ... :-(

At the end of the day, it boils down to respect. And sadly, it seems far too many people have no respect for other people. And yes, that does prove dskoll's point - as I understand, what she has been saying all along is "your behaviour fits the dictionary definition of a sociopath". That's not calling some ONE a sociopath - that's calling someone's BEHAVIOUR sociopathic, and imho she's right on the button. THAT is what needs to be addressed. And that's not a personal attack, that's a technical assault on a problem ...

And, of course, unfortunately we do have a "blame the victim" mentality - we have the thread elsewhere of the "wannabee developer" - HIS behaviour is sociopathic, and when some poor kernel guy blows his top and swears, HE's the guy who gets picked on for bad behaviour :-(

Cheers,
Wol

Sharp: Closing a door

Posted Oct 9, 2015 19:50 UTC (Fri) by fuhchee (guest, #40059) [Link] (2 responses)

"what she has been saying all along is "your behaviour fits the dictionary definition of a sociopath". That's not calling some ONE a sociopath - that's calling someone's BEHAVIOUR sociopathic, and imho she's right on the button."

No, just no.

She's inflating the degree of stated "I care only about code" to ridiculous extremes (as needed for a clinical diagnosis), instead of assuming reasonableness. She proceeds to label the behaviour with mental illness. (It is sophistry to claim only a behaviour and not the person is being labelled.)

If the deflection was going to be ... "oh, it's only consistent with $illness, not necessarily $illness", hold up. Conflict is "consistent with" genocide too, but it would be manipulative and dishonest to use the extreme term.

So, no. Someone who says they "care only about code" is not meaning literally that. Don't pretend to believe they do. Don't label them mentally ill.

Sharp: Closing a door

Posted Oct 9, 2015 20:27 UTC (Fri) by dskoll (subscriber, #1630) [Link]

Someone who says they "care only about code" is not meaning literally that.

Exactly, and that's my point. Saying that is ridiculous. We all care about other things.

The only points of contention are (1) the relative merits of good code vs. good behavior and (2) the extent (if any) to which bad behavior is necessary to produce good code. But saying you "only" care about code is shutting down the discussion. People can legitimately have different viewpoints about (1) and (2), but the extremes (only caring about code or only caring about feelings) are ridiculous.

Sharp: Closing a door

Posted Oct 9, 2015 20:49 UTC (Fri) by bronson (subscriber, #4806) [Link]

> Someone who says they "care only about code" is not meaning literally that.

So everyone agrees that someone who cares only about code does not actually exist.

According to Wol, that leaves nobody to get offended. According to fuhchee, if we paint with a wider brush, we can find lots of people to get offended. Both positions seem valid.

That said, a person claiming they 'care only about code' seems just as guilty of sophistry as the person labelling that behavior sociopathic. Both sides are overstating their positions and looking for a reaction.

So why not roll your eyes, give everyone the benefit of the doubt, and move on to a less hypothetical offense?

Sharp: Closing a door

Posted Oct 6, 2015 8:49 UTC (Tue) by rodgerd (guest, #58896) [Link] (3 responses)

> I hate to say it, but the ones seeking to impose a change bear the burden of demonstrating that said change would be worthwhile. (Or forcing said change, in extreme cases).

Given the constant whining from the kernel community about the broken whole architecture trees are (because people don't want the hassle of committing upstream), or that the kernel just lost the only person who stepped up to make sure we got USB3 support, I would say there's plenty of support things are broken.

This fine publication has reported that one third of embedded developers don't upstream because of fear of rejection noting that "Kernel subsystem maintainers tend to be strict and terse with contributors [...] If a contributor gets a reputation for submitting bad patches and wasting maintainer time, their life will get worse."

If a third of developers would rather, in essence, throw aay their code than have to work with the upstream kernel developers, things are broken.

Sharp: Closing a door

Posted Oct 6, 2015 23:04 UTC (Tue) by tialaramex (subscriber, #21167) [Link] (2 responses)

We might suppose that the Microsoft corporation, which won't even let you commit a fix that contains swear words or has anything less than complimentary to say about key partner brands, is a "friendlier" organisation to do business with in this rather limited sense.

A few seconds confirms that Microsoft's products do not have one third more upstreamed third party stuff. Quite the contrary, far more so than in Linux, everything is kept in-house and/ or thrown over the fence as unmaintainable binaries.

Laziness wins. Not every time, but overwhelmingly the true answer is "I couldn't be bothered" and what you can find out with a survey is "My currently favoured excuse for why I couldn't be bothered, PS I am goofing off to fill out this survey too".

Sharp: Closing a door

Posted Oct 7, 2015 7:17 UTC (Wed) by rodgerd (guest, #58896) [Link] (1 responses)

I fail to see any point attempting a constructive conversation with someone whose response to evidence is to claim everyone is lying.

So, have fun ignoring what people are telling you and making up little fairy tales. It won't fix them problem, of course, but I guess getting to treat people like shit it more important than working code.

Sharp: Closing a door

Posted Oct 7, 2015 19:37 UTC (Wed) by tialaramex (subscriber, #21167) [Link]

So, Microsoft do have all this stuff? They're so friendly, surely they'd be falling over non-scared developers who, we've established have plenty of spare time and are extremely enthusiastic about handing this stuff over, they're just _terrified_ of upstreaming in Linux. No?

People lie in surveys. They lie in an anonymous psychology study. They lie in a corporate satisfaction survey. A properly conducted census actually has to be post-processed to compensate for this, or it won't reflect reality. Surveys aren't useless, but you should take them with a large dose of salt.

We have an actual data point from the real world, and it doesn't match what we might think we should learn from the survey. That doesn't automatically mean the survey is wrong, but it means we should be far more cautious than it seems you're willing to be. It isn't my goal to "treat people like shit" even the inevitable LWN guests who don't have a dime to pay for a subscription but do want to put in their tuppence about how Linux ought to be run. However I do care about working code. So show me the code.