I've read that Linus is only mean to people who he's familiar\friends with or who senior enough that they should already know better. This was recently on HN and shows the more patient side of Linus (not sure how often this happens) - https://news.ycombinator.com/item?id=8761539
Typically when Linus has gotten involved, and when it's been made news, etc... it's long long after the offending commiter has taken a hard stance and a large portion of other senior commiters have attempted to deal with the situation. At a point he steps in and "brings down the hammer" so-to-speak.
Linus' "rants" are fairly infrequent in reality, and are never targeted at random people in the Kernel community -- always someone who has been in a senior position for a long while and ought to be behaving more appropriately.
"This is crap" (without pointing why or how to fix it - of course with ruder words)
"We're not applying this fix" (without any explanation why and the system is pretty much broken/only works for limited cases/bug waiting to happen/etc)
"What are you talking about?"/patronizing behaviour (when it's crystal clear what you mean)
And usually the ones with thinner skin are the ones who try to push fundamentally bad things.
I'm not fine with that. He's questioning whether another human being should be allowed to live or survive. I'm puzzled how calling for someones death (retroactively or not) is even remotely defensible in this context.
He's not calling for someones death. He's using hyperbole to make a point.
You're of course free to find that offensive, but if you think he's genuinely questioning whether or not another human being should be allowed to live or survive, you don't understand the message.
EDIT: It's also a point that likely stems from longstanding frustration. The amount of userspace software that does stupid amounts of system calls because people still do not realise how expensive system calls are is astounding. My favourite high profile pet peeve with this used to be the MySQL client library (don't know if it's been fixed) that used to do an astounding number of tiny read()'s which could have been cut to a tiny fraction with some basic user-space buffering. You can often find substantial bottlenecks in applications in about 10 seconds by attaching strace and look for stupid amounts of syscalls.
It's not to be taken literally. Still though, it seems over the top.
I think part of the problem with Linus' style is that people will inevitably try and imitate him. Not many will manage the intelligence and coding skills, but pretty much everyone can be a jerk if they put their mind to it.
I suppose you can say it's not his fault if people do that, but perhaps with a bit more kindness (to people, not to their bad ideas), the world would be a better place.
> but of course this is HN and some people take everything literally.
I have my pet theory that getting outraged about random things like that is not genuine, but just signalling. You can see a lot of this kind of behaviour among the social justice movement.
hyperbole or not, credible threat or not - questioning other humans livelihood is unacceptable in this setting. It's overly aggressive and attacks the person behind the code. I'd rate "you're not worth living" the ultimate insult.
> You seem to lack basic reading comprehension, come from a sheltered environment, have aspergers or is just trolling
This kind of comment has no place on Hacker News. I'm sure you know that.
All: when you see a comment that is particularly bad for HN, please flag it by clicking its timestamp (e.g. "4 hours ago") to go to its page, then clicking "flag". We monitor these flags and take action based on them.
Not ok. Personal attacks are not allowed on Hacker News, even when someone else has posted something inappropriate. Especially then, in fact, since those are the moments when a civil response contributes most to the community.
(There's also a bit in the HN guidelines against all caps.)
Because SJW culture (which has a sickeningly strong presence here) has convinced everyone that they have a right to be offended, and this right needs to be enforced through every nook and cranny of the Internet.
The problem in essence comes down to only dealing in absolutes. There may be a time or place for anything, and the sign of experience is to know when that time and place is.
Sadly the web has made it all too easy to take something out of its context.
I've been thinking a lot about this lately.
Does the end justifies the means? Should we allow us to have dick developers because the come up with great code?
My moral says no, but my craving for progress says yes.
It's a battle that's been driving me nuts for a long time.
Incidentally, this kind of incidents might slow down progress in the long term.
update: by this kind of incidents i mean him being an asshole, not people telling on him.
update 2: i don't endorse his behaviour.
I have remarked here multiple times that I regard "the end justifies the means" as the perfect definition of evil, but I'm not sure that is what's going on here.
I don't think Linus or Theo are inherently dicks. They actually seem to be awfully nice guys for the most part. Unfortunately, we seem to have a generation of people that cannot take the hint that their code needs work or they just don't seem to pay attention to what is going on. I'm not sure if its a symptom of rapidly developing code and then fixing it later. Most BDFLs are under stress and need to set clear boundaries. It seems like people only pay attention to the bombastic and not the subtle.
> I have remarked here multiple times that I regard "the end justifies the means" as the perfect definition of evil
I personally disagree with this definition; for me "end justifies the means" sounds like a perfectly rational conclusion, but not for humans - it's too common a failure mode for a man to follow this maxim, and it almost always ends bad; hence we have that saying.
> Unfortunately, we seem to have a generation of people that cannot take the hint that their code needs work or they just don't seem to pay attention to what is going on.
I agree. Moreover, I think we have a generation of people who like to take offense at anything and everything, mostly for signalling reasons. "I am offended by what you wrote, therefore I am a better person and you are wrong".
> I personally disagree with this definition; for me "end justifies the means" sounds like a perfectly rational conclusion, but not for humans - it's too common a failure mode for a man to follow this maxim, and it almost always ends bad; hence we have that saying.
I have never heard a person use that phrase and not violate the rights of other humans and their own morality. It is the gateway phrase people use to justify some act that they admit, by using that phrase, is not honorable. It's the phrase you say right before walking on that road paved with all those good intentions.
This post elaborates on the topic a bit more, and after reading it some time ago I ended up agreeing with its conclusion - that the phrase is OK from theoretical standpoint, for perfectly rational actors, but not for humans - hence the heuristic, "end does not justify the means".
> Should we allow us to have dick developers because the come up with great code?
Even if we're making this "the only thing that matters is how great the code is" argument, there are good arguments against allowing people to be hostile.
You never know whether the size of your developer community could be larger. If you're turning 10 expert but conflict-averse developers away for every one asshole genius, it is unlikely that your code output/quality is being maximized.
There's no shortage of people -- even people who are paid to hack on the kernel every day! -- who say that they refuse to engage with the upstream community due to the hostility they expect to receive or have received.
The key thing is that size is not the most important measure of a dev community. A larger community of low-skill low-time-commitment devs may be substantially less valuable in practical terms than a few highly skilled and highly committed devs.
There's also the question of if the people to whom you refer would change their behavior if the item they point to was removed. As we all know, what people say they will do and what people actually do often diverge.
A better question, perhaps, is if a warm, accepting, sensitive community produces technically superior products.
A community that is all those things is of immense social value to those who feel accepted as a result, but warm fuzzies aren't the same as working software.
If you look across history, genius and caustic personalities seem to go hand in hand. Only that Torvalds is venting on a public and logged mailing list, while in previous times it was mostly stated verbally in private conversation.
Is Linus euphemistically referring to someone specific there that everyone taking part in the conversation knows the real name of very well and who probably read that comment and recognized themselves, or is the idea of whomsoever decided that that thing should work "one byte at a time" a more diffuse fact that's obscured in history? That makes a difference in that particular situation, I feel.
Not at all. Both are about not conforming to unstated conventions. A lot of what's been written about "unprofessional" applies equally well to "offensive".
You have a point. While I don't think being an asshole is a part of hacker culture, I can see a lot of things hacker say and do be labelled as "offensive" or "unprofessional", because they break unwritten rules about behaviour. Most of the time I saw that happening it was because hackers tend to state things directly instead of engaging in political correctness that is perceived to be a waste of time.
Because Stallman food habits are just ugly, not something one can point at and say "I'm offended!". There's a difference between saying "Yuck! I don't feel comfortable around that person." and "I am offended, you need to change yourself!". People are defending Torvalds from people with second attitude.
There are a lot of rude people, people who answer in several ridiculous ways, etc.
Linus answers may be rough, but it's usually a clear answer, and well founded. There's a lot of worse behaviour out there (especially in subsystems)