Harsh or soft, criticism should always be actionable. The way you describe Torvalds' feedback, that's actionable, when someone is told what makes their code unacceptable, so they can correct course and try again.
EDIT: I should edit to add, the post that the quote is from -does- in fact then go on to point out what it finds unacceptable. But I'm commenting just on the quote.
Telling someone to kill themselves is not a sound argument under any circumstances.
But yes I think it's literally absolutely always wrong to tell anyone to kill themselves, no matter how evil they are. Personal abuse is always wrong.
In the scope of any realistically possible pull request discussion, yes, you're undoubtedly right.
There are a few misconceptions at play:
1. that expressing negative emotions implies being offensive
2. that being offensive strengthens an argument
In my experience, being able to separate the offensive part, which is certainly difficult, leads to much more productive outcomes, without losing argumentative strength - besides, being also (considered) more professional.> some think themselves or others must be protected from "harshness"
In the above perspective, harshness is simply improductive in any context - being firm and persuasive is always a better choice. It's just very difficult :^)
> I'd rather be told I'm working on a dead end than a "nice comforting letter" explaining how my patch can't be accepted _right now_ because of x and y with wording that could be interpreted as "it could be accepted later"
This is a dichotomy between an arguably harsh approach and a false/polite one. There are other different approaches (besides, I think that there's nothing offensive/harsh in being told that one is working on a dead end).
> I don't see anything wrong with that quote. Can it be expressed more "softly"? Sure. But that person is expressing his opinion and I don't see any problem with that.
Being considerate towards other people leads to a much more productive collaboration. This vision:
> You could also argue it lays the foundation for further escalation but that's on whoever chooses to engage in it IMO.
is indeed non-collaborative - essentially, not being interested at all about the other person. The point is then just venting frustration rather than communicating (in general, not the specific case).
I doubt that anybody can hold a lead/management position with this attitude (without being considered at least an unpleasant manager).
> For example Linus Torvalds gets a lot of flak for his opinions on some pieces of code he doesn't like but every time I read those emails, I feel like people are blowing it way out of proportion. I'm sure it's not fun to be on the receiving end but his arguments are typically sound and likes to write a lot to really drive his point home.
This argument actually proves the point. Torvals made a public apology:
http://lkml.iu.edu/hypermail/linux/kernel/1809.2/00117.html
so definitely there is something (to say the least) inappropriate with that attitude. There are indications that he went to therapy specifically because of that: https://twitter.com/lexnfx/status/1409185767348310020
All in all, I definitely support (in a more verbose way :^)) the argument:> top comments make clear there is a dire need for professional communications training among these programmers
Whilst true, what often happens is that people emotionally express disapproval. This then often includes harsh judgement not just of someones work, but of their person as well. At that point it becomes insulting. Even when not moving on to harshly judging a person, totally destroying someone's work is likely to put them on the defensive. For the purposes of keeping up conversation, it can be equivalent to being insulting.
Consider remarks like.
"This code is riddled with bugs and lacks insights form the last 10 years. Whomever wrote it must be willfully ignorant or a total moron with their head in the sand".
It reads like a statement made from emotion, it technically only includes "making a point" but it goes so far as too become insulting. Even if you drop the second sentence it can still be insulting.
This sounds like victim-blaming. I can imagine neo-nazis using this type of statement to defend someone who yelled the N-word in a black neighborhood.
The truth is that inflammatory language, which GP's quote was, is rarely constructive.
There seems to be a more general trend going on, some think themselves or others must be protected from "harshness" (even though I consider the above quote to be only a mild example - see the rest of the internet...). For example Linus Torvalds gets a lot of flak for his opinions on some pieces of code he doesn't like but every time I read those emails, I feel like people are blowing it way out of proportion. I'm sure it's not fun to be on the receiving end but his arguments are typically sound and likes to write a lot to really drive his point home. I'd rather be told I'm working on a dead end than a "nice comforting letter" explaining how my patch can't be accepted _right now_ because of x and y with wording that could be interpreted as "it could be accepted later" just to be friendly even though the code is structurally bad and needs to go back to the drawing board. It would give hope where there is none and I think that is worse than "no your code sucks because of x y and z".
I think in this case here it was ok for disappointment to be expressed.