Dear Linus, STOP SHOUTING and play nice - says Linux kernel dev A Linux developer has blasted the kernel's chief Linus Torvalds, taking him to task for his famous potty mouth and brutal putdowns of his lieutenants. It's time for Torvalds to stop "verbally abusing" his programmers, Sarah Sharp told the fiery Finn, warning him she’s "not taking it any more". The USB 3.0 driver maintainer …
Except Torvalds is obviously not an idiot.
And, tbh, he's right about 'On the internet, nobody can hear you do subtle'
As I see it the problem with big OSS projects like Linux is exactly the back stabbing LT is describing, and perhaps being a dictator is one of the best ways of handling the disparate group of people working on it. Being nice simply won't work - people will just continue doing exactly the wrong things again and again. Although it seems to me that even shouting at people hasn't stopped them committed code that simply shouldn't be committed!
"And, tbh, he's right about 'On the internet, nobody can hear you do subtle'"
I strongly disagree. The people who matter and value the things you say will hear you loud and clear, even if you do things subtle. I'm getting the feeling that it's not so much the people who matter who should hear him, it's the rest of the world too.
Lets face it: would some of the kernel mailinglist posts have made it to El Reg if it wasn't for the shouting and swearing? For example when he told RedHat off about code signing? Sure, in my opinion he was right on that call, but you can bring a message across without making it look as if you feel better than everyone else.
And there's something else to consider: what would happen if someone forks the Linux kernel, starts a new Linux-like project and actually treats people as they want to be treated? I know no one has done this (there'd be holy wars) but for the sake of argument lets say it happens.
How long before Linus' "lieutenants" bail out? On the other end of the mountain they may still end up in discussions and people disagreeing with them, but they could count on the respect they deserve.
I think this is a liable risk presenting itself right now. People representing RedHat will remain on board no matter what, there's money involved. But hobbyists may reach a point where they no longer think its cool anymore. And then a ship can sink really fast.
Are you now suggesting Linus Torvalds should be scared and kneel in front of his lieutenants just because they might fork the kernel ? I don't remember ever seeing Steve Ballmer apologizing or being concerned about what people at Microsoft might feel. Sorry, mate, but if Linus allows himself to be blackmailed by his developers then Linux kernel is toast. He knows damn well that the kernel can be forked because he designed it like that.
"Are you now suggesting Linus Torvalds should be scared and kneel in front of his lieutenants just because they might fork the kernel ?"
No, but there's a big difference between saying someone wrote horrible code and someone is a horrible person. Because this can only spiral out of control and go out with a bang, unless someone comes to his senses.
Let's assume Linus is right and that no one listens to him if he says (quote:) "please don't do that". Has no one ever stopped to wonder why that may be so?
If you start raising your voice to get your point across you will reach a point where people eventually ignore that shit ("there he goes again"). So you'll have to come up with other ways to get your point across in a way where people know you really meant it.
Let's start swearing..
Yet now you only started a new cycle. Eventually you reach a point where people get fed up again and will ignore it ("here we go again").
Let's start swearing.. in Finnish!
So where does all of this end? Don't forget that most people who tolerate that shit don't necessarily do so because they agree, but because they can't back out. There is only one Linux kernel. But take that prestige or the need away and I think things will come crumbling down, and fast too.
"Except Torvalds is obviously not an idiot."
Well I don't know. He commits code that he assumes (i.e. doesn't check) to be bug free. Then blames someody else for introducing the bug into the kernel. Hmmm. Who committed the code?
Let's face it, if Linus was so great as the Linux fanbois like to pretend he is, there would never have been a single bug in the whole history of the Linux kernel.
Except that he isn't necessarily the one who committed the changes: they may have arrived via one or more git repositories (even if also posted to lkml). Also, commits tend to have at least one sign-off by various system or subsystem maintainers, and they should be checking patches and commits too. It works, but things will slip through sometimes.
(Linus is known to have learned the hard way about last-minute what-could-possibly-go-wrong changes…)
But there's more than one way to get your point across. So what if the Internet doesn't do subtle? Whatever happened to "civil but firm"? Torvalds' rant could easily have been rephrased on the lines of:
"You committed bug-ridden code to the project such that others mistakenly assume it to be stable. In future, warn us beforehand when submitting unstable code. Otherwise, do not submit the code until it is in a stable state."
See? No profanity at all, and no "fake politeness" either. But also pretty cut-and-dry in what's being mentioned and requested: no subtlety at all.
"You committed bug-ridden code to the project such that others mistakenly assume it to be stable. In future, warn us beforehand when submitting unstable code. Otherwise, do not submit the code until it is in a stable state."
Yeah - sure. That will do the trick. Clearly all those devs had no idea and were never told that code marked "stable" shouldn't be, well, err, unstable. A bit of nice and polite talking to is all that is required. My a*se! Without knowing the full details - people who ask to be told politely what to do many times ignore it completely when they are being told politely. And then they complain when "polite" has run out. They would like to be treated "politely" so they can keep on ignoring you and do what they want.
Sometimes people just do crap until somebody puts some pressure on them. Be it that it's a "nice" (but subtly threatening) meeting with your boss or HR people in your company - which is what LT is referring to as tie-wearing back stabbing. Or in LT's case, being told to f*ck off and do things properly.
On many occasions people know exactly what should be done, but can't be a*sed for one reason or another to do things properly. On one hand harsh verbal manners are not exactly cool - but I think I prefer being put in my place that way (if I deserve it) instead of one of those "nice talking to's" after which you feel all creeped out. Sometimes there is no nice way to get things done - your choice - being shouted at or being subtly threatened.
Pushing shit into the kernel expecting roses and then making fuss about it, is something else. Yes, I know men can, and have done the same earlier. But now, silly woman, try to understand your responsibility or fuck off. It is not about you but about, apparently, shit code you try to push.
" Passion is one thing
But why do people believe it's ok to cross the line into abusiveness?"
And submitting crap code ISN'T abusive? You need to examine your priorities. Linux is used in a wide range of very critical applications from core networking to military. Sharp's "you shouldn't be mean to me" attitude glances aside form the fact that she and her people are quite literally making life harder for - abusing that is - numerous others, thousands, in fact. By that standard, Linus isn't even intemperate, and Sharp ought to go into politics where she can do more harm more efficiently. It's evident that she is not "leading" her team, or holding them to any kind of quality standard.
@OG A.C.
No one has seemed to notice that you posted as A.C., which means you won't even back up your own passive identity. I think the Linus management style would work very well with you.
Since when did being vulgar in detail about work become associated with personal feelings? Maybe "Sarah" needs to learn to deal with multiple personality types, instead of pressing her own as being correct for everyone and not just hers alone.
If having to *read* vulgar comments about me was the worst part of my day, I'd be smiling right now.
Someone committed unstable shit as stable code and wasted the guy's time. He has to deal with a myriad of commits and keep the project rolling forward. What he doesn't need is some fucking idiot wasting time that he probably has very little of. To then come back with a series of "speak to me nicely" missives probably made the Finnish volcano blow its top. If you act like an idiot and show utter contempt for somebody else's time then don't expect their reaction to contain please or thank you. Simply - do your fucking job.
If I screw up, I expect to be told exactly how, and how badly. If I've screwed up badly, I expect to be told in the bluntest possible terms. Profanity evolved because it serves a purpose.
Let me put it this way - when I am in the navigator's seat of a rally car, and I call a bend wrongly so that it risks the integrity of the car and/or the occupants/spectators, I expect to be told in no uncertain terms that I was very, very silly. Equally, if my driver doesn't listen to me with the same results, I am going to use an appropriate amount of Anglo-Saxon on him/her.
The cruellest tellings-off are the ones that are apparently polite, but play on emotion. I've never wanted to kill anyone that swore at me, but I've spent hours plotting the demise of someone who used the latter.
exactly. I've had committers to projects I've been involved in submit crap, break many many tests and obviously never run the tests, then ask them politely to fix their "work" asap, them do nothing, so I have to spend my time backing it out. The blunt and direct approach gets more results IMHO (though the level of the searing necessary to be "blunt and direct" is open to debate)
Judging by many open source projects, what you actually end up with is contributions coming only from arrogant arseholes who always believe they're in the right and are too thick skinned for even the blunt abrasive comments made by the likes of Linus to have any effect. Meanwhile lots of nice, polite and extremely capable coders are put off by the unnecessarily rude behaviour of the minority.
If you honestly think that approach makes for better software, then you're an idiot.
Acting professional is something quite different from being professional. What seems to set him off is amateurs NOT being professional and then demanding to be treated professionally.
I once tried to walk this line in a volunteer organization. I thought we needed to treat our supporting members (aka customers in standard business speak) professionally. Others in our organization argued that professionalism would degrade into the sorts of corporatist behavior none of us like in big organizations. Without yielding on the point that we needed to treat our supporting members with respect, I must admit that while we were able to maintain the balance I expected for a short time, in the end those who argued against me were correct about the direction the group eventually went.
I personally would give Torvalds seven kinds of shit back if he took that approach. I respond well to calm, concise explanations as to where I made a mistake. If it's made into a piss-take based joke, even better. I get onto the problem with a smile, eager to resolve it. Much better than through gritted teeth and a fuck you for raging at me attitude.
But that approach works for me, others possibly not. That's where management skills come in, so you can identify which approach works for each person. You can't use a hammer for all the jobs!
Good luck with that, your code would get rejected and not integrated into the kernel and if you are submitting it on behalf of a hardware vendor I think you would be out of a job.
Linus is trying to manage all of the code that people are sending him, why shouldn't he be angry if he is being sent rubbish? your manager would shout at you if you sent them rubbish. You would fail any exam or degree if you submitted rubbish, so why should Linus accept it?
You are assuming that I am making a mistake every time. You are also assuming that shouting works because someone doesn't like something.
Like any human being I make the occasional mistake, certainly within normal expected parameters because I have had the same coding job for over 6 years. Shout at me and you WILL get worse results because I want to rip your head off and shove it up your own ass. Treat me with respect and I'll move heaven and Earth to get it right. I'll say it again: You can't use the same hammer for every job.
You are assuming that I am making a mistake every time. You are also assuming that shouting works because someone doesn't like something.
How many times has Torvalds gone off like this when someone didn't make a mistake? That's the key question here, in my book.
"when was the last time you had an examiner call you a fucking idiot?"
Personally I've never had it happen, but I saw it happen to other people when I was at university. And I will be frank, they deserved it.
(not the AC for the record, just a guy who was smart enough to not be called a fucking idiot)
"when was the last time you had an examiner call you a fucking idiot?"
So it would have been better if instead of calling her (and her team) on submitting untested code as STABLE... they should've all gotten a pink slip and be fired?... that's what happens on tests, you get almost no feedback and failing grade
Submitting untested code as stable is a HUGE problem, it is not a difference of opinions, it is not a difference in programming styles... it is serving rotten fish hoping the sauce will cover the stink and THAT is unprofessional...
"when was the last time you had an examiner call you a fucking idiot?"
We aren't allowed to any more. It apparently reflects badly on the organisation to be honest.
To my mind, calling someone a "fucking idiot" is fine, as long it is proportionate to the error, you explain why you said it, and how it can be avoided next time. Torvalds ticked all these three boxes, and it is his organisation so he sets the rules. What is the problem with being hones and clear.
So many delicate little flowers in positions of responsibility - no wonder Western civilisation is on the skids.
More to the point, in the grand scheme of things I expect that someone will occasionally check in the wrong segment of code. (Oops, I sent the code dated 6/30 instead of 7/30.) At which point the manager of the code base should be the one noticing and taking the corrective action with the coder. In this instance not only did the coder check in untested code, the manager (whose job it is to check for that so Torvalds doesn't have to) just forwarded the untested code for inclusion as part of the final build.
Disasters are rarely the result of single mistake and almost always the result of a chain of mistakes.
> “You may need to learn to shout at people,”
Personally, I've been shouted at by experts: people who do it professionally and can verbally abuse you for 5 minutes, fluently, without hesitation, repetition or deviation (although some of the deviations you get accused of would make your hair curl and IIRC, that was one of them!). So some guy shouting at me, either in person, over the phone (ineffective: you just quietly lay the receiver down 'till they've fininshed - it's hard to argue with someone who won't answer back) or by email (even more ineffective and with the added disadvantage that it leaves evidence to be used against the shouter) might make them feel better but it will have zero effect on me.
At best you suffer from raised blood pressure and fear. Although fear is a great motivator in the short term, it's effect needs to be continually escalated to continue achieving results. At worst you get someone like me who ignores it all and just thinks of the shouty git as a complete 'hole and ends up with an even lower opinion of them.
If you want to motivate someone, someone who you have no real control over as FOSS developers can stop FOSS-ing and just walk away, you need to adopt an air of professionalism, gain respect and lead by example.
Sounds like Linus has pretty much lost it.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
