50 comments

[ 6.7 ms ] story [ 109 ms ] thread
From one of the emails:

I care not one whit about the reason for them. In fact, if the reason is stated as "it makes debugging easiler", then I fart in your general direction and call your mother a hamster.

And when confronted about it:

> That is plain harassment. I ask to stop it!

The correct reply is

"Is there someone else up there we can talk to?"

just google for it if you haven't seen the Holy Grail.

(And I got the quote wrong too. I forgot about how your father smelt of elderberries)

Disagreeing about a piece of code is one thing, abusing the person you disagree with is another. Such language doesn't belong in an open source forum.

Calling "I fart in your general direction" abusive feels like a pretty big stretch to me.

It's rude or immature at worst.

"Call your mother a hamster" was the part I wanted to highlight
Is not abuse. And clearly you get to decide what belongs and what doesnt in an "open source forum" (as if that was a thing).
I don't get to and don't want to. But I certainly don't want to debate with a person who calls my mother a hamster or how my father smelled.
(comment deleted)
Linus has mentioned abuse of BUG_ON on multiple occasions. What is the deal with whiny devs wanting to make Linus look like an ogre when they are CLEARLY in the wrong?
Communication is a foundation of software engineering, and the problem is here is about communication.

If you read the discussion (https://lore.kernel.org/lkml/20180813214328.GA15137@beast/T/...) you'll find how the devs are trying to tackle the subject constructively, while Torvalds rants and (arguably) abuses more than focusing on the technical side.

All of his bantering (this is a very charitable definition) is completely redundant to the discussion, and at least irritating (very charitable definition as well).

The extract of the twit `I and @kees_cook had hard times extracting the rationale from his wild abuse.` is the core of the problem, and I find it very accurate.

I have (of course) read the discussion. The rationale is perfectly clear from Linus first response!

“Extract of the twit”? ROFL. Sadly I think that is the whole problem.

He's such a jerk.
He can be for sure... but look what he has to manage. and It wont matter in a while anyway when systemd has its OWN kernel, and people have to deal with Lennart.
Definitely. Still, I have to respect that he's worked on one project for 26 years. That is a long time!

Also, I suspect in real life he is not such a dick, but it's easy to rant in an email and he probably doesn't realise how it comes across.

He's the jerk we dont deserve, but the jerk everybody needs.
from his reply:

>I absolutely refuse to take any hardening patches at all that have BUG() or panic() or similar machine-killing in it. I care not one whit about the reason for them. In fact, if the reason is stated as "it makes debugging easiler", then I fart in your general direction and call your mother a hamster.

yes "this is typical Torvalds", but to me the tone suggests his personality is still stuck in the mid/early 90ies. It was perfectly ok to use insults dressed as humor in order to win arguments with the "unwashed masses" on Usenet. It was funny back then but seeing this in a middle-aged man in 2018 tells me the guy still hasn't grown up.

Also nobody expects the Spanish Inquisition, but I'd expect at least some fresh jokes, because the Monty Python references were already old back then.

If you don't like Linus, you can always fork Linux and create DyslexicAtheist OS :), but no you are attacking, he clearly says few mails later, that killing machine without very good reason is big no:

I absolutely refuse to take any hardening patches at all that have BUG() or panic() or similar machine-killing in it.

I think the bulk of the argument is about the insult. Creating a new fork doens't really help anyone. Creating a new standard just increases the number of standards.
sad to say but Linux security is shit mostly because of Torvalds posturing. He has a history of lashing out to anyone contributing security and compared security researchers and the BSD camp to a bunch of masturbating monkeys. Grsec is dead too.

I worked as a Linux Evangelist in the 90ies for a major distro and was a big fan and later as developer would only work in "linux shops". We have ridiculed and hurled abuse at Microsofts poor security at every opportunity (for over a decade), which has IMHO massively strengthened Microsoft - but all this Linux superiority posturing did nothing to improve Linux security.

It's kind of annoying (and somewhat a relic of the 90ies) that people still claim Linux has good security. It was an easy claim to make when the thing that Linux compared itself to was a Win95/Win98 Desktop installation. The thing Linux community keeps illustrating is that it's easy to stand on the sideline shouting at your competitor but much harder to actually design better security. All the abuse (or was it tough love?) targeted at Microsoft did a lot of good for Microsoft but it did the opposite to Linux.

He really is quite abusive. It's remarkable to me that the Linux kernel ecosystem is thriving with this kind of leader.
Reading the full thread it seems everyone walks in eggshells around him, speaking so sweetly to him while he insults them and their work.
Yes, I agree. The responses were actually more disturbing than the initial outburst. It sounded very much like what I expect from an abusive relationship.
What I find interesting is that they must at some level realize that it does not work: Torvalds is not going to respect you more, or cut you some slack because you are incredibly polite and careful in your wording. It strikes me as rather dumb. If the entire conversation can be reworded as 'Hey I think it's okay to break the machine in such-and-such a case' - 'No, I fart in your general direction, this fucks up my standards for quality software' - 'Okay', why are both Alexander and Kees trying to be so incredibly polite? It's just a waste of time.
I don't get that feeling at all. The thing that stands out to me is how they are constantly mentioning something along the lines of "we discussed this with him/her on the maliing list", "they said it was okay", ... When Linus from first to last email makes it clear that he does not want any BUG() and discussion about it is moot.

Reminds me of breaking rules as a child and then justifying it by saying that a friends mother said it was okay, or asking my dad about something my mom had already forbidden...

I can see how that is frustrating from his point, when he made it clear multiple times and is basically reiterating it that whole thread, until in the one of them even thinks of a solution.

Puzzling to me is the number of people who extoll, and even strive to emulate, those qualities in Steve Jobs.
Yes, he should be more polite like Guido Van Rossum to ensure longevity.
Lots of leaders are abusive.

Steve Jobs would go into meetings and shout at people calling them "fucking dickless assholes". Did you find it similarly remarkable that Apple thrived under Jobs? What's the difference?

I can think of two things:

1. Linus conducts his business on the public LKML so it's much more visible.

2. Technical people are expected to shut up and be polite.

It thrives, but with a subset of the people it could have. I'm reminded of both the low proportion of women (commits on the Linux kernel are 94-96% men) and kernel developer's like Sarah Sharpe citing the toxic culture for leaving.

Sarah Sharpe's explanation for leaving:

http://sage.thesharps.us/2015/10/05/closing-a-door/

Gender analysis of Linux kernel technical contributions:

https://blog.bitergia.com/2016/10/11/gender-diversity-analys...

It's so unfortunate that he continues to be so abusive, inspiring a toxic culture with it, and that he has been completely unwilling to change. This community could have so many more bright minds than it does.

From Linus we apparently accept it, but not from any other $RANDOM developer.
Is this the tone developers have to adopt in order to contribute?

F that then.

I don't think the developers are required to adopt any kind of tone.
You don't have to adopt any tone to contribute, but if you can not handle a Monty Python reference maybe you shouldn't contribute.
Can you imagine that there are people who did not watch Monty Python at all? I know it is not always easy for a western person to accept a fact that there are different cultures around the world.
On the other hand, when I interact with people from wildly different cultures than my own and they say something that sounds very strange to me, my first reaction is not to take offense but rather to try to figure out if there's a cultural difference I'm oblivious to.
I completely agree it is ok to hear strange jokes and not to be offensive. But to hear 'if you can not handle a Monty Python reference maybe you shouldn't contribute' is too much for me.
If you can't do constructive criticism (even in the face of repeated mistakes), normal human communication (meaning apologizing when your well-intentioned joke was poorly received, for example) then you shouldn't be a maintainer accepting contributions. Or working with humans at all.
No to contribute you have to stay polite, like the non-Linus devs in the linked thread. And if your patch is good the maintainer might merge it.
Linus' reasoning echos his past criticisms of hardening patches:

>if people run things on real machines, then BUG() is absolutely the last thing you EVER want to do for "debugging".

>This is why I scanned your pull request for BUG() and similar. Because I simply will not take "hardening" that kills the machine. That's a hard requirement. No excuses, and absolutely zero exceptions.

>After a year or two, when the hardening has actually been in place, and you can say "hey, look, none of the warnings happened", I may be ok with turning them into BUG() calls.

The monty python drama is a distraction, this is an issue of principle that Linus has had to explain multiple times.

Reading through the responses here, is seems that we're forgetting a. Torvalds' well known demeanor, and b. that he's protecting linux.

Yes, he's unpleasant much of the time. But he has to be. Imaging if he doesn't take this hard line and something broken worms its way into linux. You know how grumpy he's going to be when he sees the headline "Gajillions of linux servers suffer random lock-ups" on Hacker News?

He doesn't have the luxury of suffering fools. I'm generally considered a nice person - if you put me in charge of linux, it'd be a mess.

He doesn't have the luxury of accepting bad pull requests. I don't agree that this means he has to be abrasive or abusive.
Arguably it reduces his workload. If you're terse and abrasive people will try to spend less time talking to you except when absolutely necessary, which frees up more time for reviewing and writing patches.
Linus has a handful of points he doesn't want to compromise on: "breaking user space", "machine-killing", "using C++ in kernel", "writing poor code", ... Around the world, there are many people who disagree and try to push patches that violate these principles. I think he exagerates his reactions to remind people about these principles.
When you point a gun at someone while saying something, they never really remember what you said. The just remember the gun.

That sort of thing might go a long way toward explaining why Linus has such a hard time getting people on board with his ideas about these things.

Why do people keep tiptoeing around this man? There is ZERO times when a man's "influence", "brilliance", "success" or whatever the hell you want to call it means that it's right to be abusive. And he doesn't just have a "special jargon" or something. He is downright abusive. It doesn't matter whether Linux is important, whether Linus is successful. It doesn't matter that "Steve Jobs was also an asshat". It doesn't matter.

Just don't take crap. I'm not blaming Linus for being an asshat - there will always be asshats. But asshats shouldn't be in positions where they can be asshats. They should be trolling on a forum somewhere.

I'm blaming the people tiptoeing around him which enables him and propagates the myth that it's ok to be abusive in tech in 2018. Just walk away. The people who say "Linus will be Linus" or "He's just protecting Linux", or "If you don't like it, make your own software project" are the real problem. Stop that.

>Why do people keep tiptoeing around this man?

Because they can't write a rule-abiding patch set and yet feel the pressure to deliver off of their half-baked code. Under the pressure from their customers/bosses/whomever, they try to weasel their way in with nice words rather than with merit.

Patches that go by the rules and merge cleanly are graciously welcome. No tiptoeing ever needed or expected.

Also, they could "just fork Linux", but nobody would take such project seriously without a well-established leader always on guard to reject any code smell. You don't get repeatable success spanning decades, and “World domination. Fast.” [1] without a focused and stern leader like Torvalds, Jobs, Bezos, or Musk. Polite committees give us C++11 and XHTML 2.0 and POSIX and other such footguns, because nobody stood up and called out bullshiters on their bullshit. On that note, watch Python, now that Guido the BDFL is no more.

[1] https://www.linuxjournal.com/article/36

i mean...considering this is torvalds turning down a pr...I think he kept it quite civil this time around. what's the fuzz about? someone missing the holy grail reference? c'mon...
A lot of people tend to focus on the angry or inflammatory nature of Linus's patch reviews. I don't really understand why. From the reviews I've read, Linus always comes across very high signal--he cuts straight to the big issues with a patch, and it usually seems sort of easy to separate that signal from something inflammatory.

So far in my career, the people like Linus: very direct, high signal with lots of knowledge, perhaps a bit abusive (typically because they have to act as gatekeepers to keep bad code or features out) have taught me more than really anyone else.

My feeling is that people nowadays put too much responsibility on the communicator and not enough on the listener. A good listener needs to be able to extract signal out of communications and put it in context (sort of like the classical meditation: being able to perfectly balance focus and mindfulness). It takes both a good listener and a good communicator for a good conversation: each person will often not be perfect and will rely on their counterparty to fill in the gaps.

Being a less-than-stellar - even poor - communicator is absolutely something people should cut each other slack for.

Being blatantly abusive, though, should never be acceptable.

Is it just me or have these posts to the LKM become a lot more frequently? I feel like there is one every other week now and I can't really imagine that Linus cursing has increased that much in frequency.

What's the point/agenda here?

I watched Linus's TED Talk where he's being interviewed (which is unusual in TED Talk format). [0] He seemed so nice and I thought he was so cool, like a hero in the CS world. I want to rationalize and defend him because his behavior in this email thread doesn't match my preexisting conception of him but I realize that would be a mistake. I don't think he needed to behave that way. I don't want to start an internet mob or turn this into a binary thing like "he's great!" or "he's awful!" are the only two possibilities. His accomplishments are really cool, but he doesn't have to be a dick.

[0] https://www.ted.com/talks/linus_torvalds_the_mind_behind_lin...