Live data from Hacker News

We do not break userspace (2012)

lkml.org

41–49 of 49 posts

Re: We do not break userspace (2012)

#41

I may be an average corporate backend dev, but if anyone I work with would react in this way after any kind of mistake I did, I would report the situation / leave immediately. Honest feedback is surely key to keep quality, innovation, and skills high - the thread has indeed a few lines of honest feedback (don't break user-space, wrong error) and the rest are just insults. I am pretty sure it goes against the Code of…

Linus had a reputation for doing this, back in the day. He took some time off a few years ago, and while he's still very strict on quality, his tone has changed.

I've run a lot of dev teams, and I see two side to this.

On the one side, the worse thing you can do is build an environment where folks are scared to raise new ideas, or think of new solutions, for fear of being publicly humiliated. At best, you will get teams that do very average work. At worse, you'll loose all your good people.

On the other side, I've had devs who thought they were god's gift for code, and made everyone's life around them, a misery. e.g. one guy spent months failing to implement an API with the front end he was writing, but despite that, he'd comment on PR's "This code is shit and needs to be rewritten."

Folks like that need to be part of, what we call in the UK, a frank and open conversation, which I had one to one. This particular dev refused to back down, and left without working his notice. Going somewhere where they would, in his view, thank him for his feedback. Good luck with that. They guy that replaced him had the API working inside of a week.

tbf, the API was terrible, but that's a whole other story involving IBM and SaleForce...

Re: We do not break userspace (2012)

#42
post #23

I may be an average corporate backend dev, but if anyone I work with would react in this way after any kind of mistake I did, I would report the situation / leave immediately. Honest feedback is surely key to keep quality, innovation, and skills high - the thread has indeed a few lines of honest feedback (don't break user-space, wrong error) and the rest are just insults. I am pretty sure it goes against the Code of…

> I would report the situation / leave immediately. [...] I am pretty sure it goes against the Code of Conduct [1] Back in the early 90s, I was young and working for a startup. I had a fiery and seasoned boss, who knew what he was doing and had very little patience for bullshit. One day, I decided I knew better than him, and implemented a solution that was way outside the spec of what he had asked for. The next day,…

My foot still smarts to this day. I miss the days office workers could keep a whiskey cabinet and work things out hermano to hermano. Peace brutha.

Re: We do not break userspace (2012)

#43

I may be an average corporate backend dev, but if anyone I work with would react in this way after any kind of mistake I did, I would report the situation / leave immediately. Honest feedback is surely key to keep quality, innovation, and skills high - the thread has indeed a few lines of honest feedback (don't break user-space, wrong error) and the rest are just insults. I am pretty sure it goes against the Code of…

Imagine you are a safety officer, and you see a worker committing huge safety violation, and thereby putting in danger his coworker's lives. Which of the following ways has better chance of teaching him the gravity of the situation, and preventing this error repeated in the future: a) kindly point his error to him in a nice and friendly way. b) give him a verbal beatdown with insults and head-shaking

I wonder what HN would think about this guy's account of his medical school experience.

https://web.archive.org/web/20101218031844/http://www.medsch...

https://web.archive.org/web/20101211180021/http://www.medsch...

Re: We do not break userspace (2012)

#44
post #20

> And fix your approach to kernel programming. At times I've admired Linus's tendency to not sugar coat things. Sometimes things need to be said bluntly. In this case, however, despite being correct, his message would have been stronger without all of the personal insults. While the maintainer submitted a patch that was clearly the wrong approach, it's a bit ironic that the one who needed to fix their approach to ker…

This isn't directed at you at all, but "don't sugar coat it" may be my least favorite phrase in English. I get the theory, but, personally speaking, I have nigh exclusively seen it used to excuse aggressive negative emotion, rather than to encourage direct and honest communication.

Re: We do not break userspace (2012)

#45

Do you want nice things? Because this is how we get nice things. You may not like it. I certainly prefer a nicer mode of communication. But this is peak infrastructure maintenance. Compare and contrast: https://news.ycombinator.com/item?id=29845724

It’s absolutely possible to communicate someone has made a mistake without insulting them. Nice things don’t come from degradation.

Re: We do not break userspace (2012)

#46
post #23

I may be an average corporate backend dev, but if anyone I work with would react in this way after any kind of mistake I did, I would report the situation / leave immediately. Honest feedback is surely key to keep quality, innovation, and skills high - the thread has indeed a few lines of honest feedback (don't break user-space, wrong error) and the rest are just insults. I am pretty sure it goes against the Code of…

> I would report the situation / leave immediately. [...] I am pretty sure it goes against the Code of Conduct [1] Back in the early 90s, I was young and working for a startup. I had a fiery and seasoned boss, who knew what he was doing and had very little patience for bullshit. One day, I decided I knew better than him, and implemented a solution that was way outside the spec of what he had asked for. The next day,…

There is space between breaking your foot while kicking a filing cabinet and formal review processes.

I’ve been straightened out, for example, by a reasonable explanation of what I did wrong.

Re: We do not break userspace (2012)

#47
post #23

Earlier quoted context omitted.

> I would report the situation / leave immediately. [...] I am pretty sure it goes against the Code of Conduct [1] Back in the early 90s, I was young and working for a startup. I had a fiery and seasoned boss, who knew what he was doing and had very little patience for bullshit. One day, I decided I knew better than him, and implemented a solution that was way outside the spec of what he had asked for. The next day,…

There is space between breaking your foot while kicking a filing cabinet and formal review processes. I’ve been straightened out, for example, by a reasonable explanation of what I did wrong.

Maybe you didn't screw up as bad as I did? Maybe the stakes weren't as high? Who knows,. it's all about context. Blanket bans on behavior ignore that context.

Re: We do not break userspace (2012)

#48
Linus definitely needed anger management.

But, nevertheless, I don't know the full context. Linus might have repeatedly emphasize "We do not break userspace" to the point he is red-faced in this particular incident.

Having a bad commit to a project you are responsible for, with a large number of users, is frustrating. Not to mention that Linux is not just a responsibility, but a life-work. What should he do instead of this? Block bad commit author from contributing to Linux forever?

A programmer blaming the user for the programmer's mistake is not less callous of an action. Probably better or worse, because it is wrapped with courteous words, which imply that the programmer is right. What if this happens on a mission critical machine that deals with life-and-death situation? Whose head is gonna end up in the masses' pitchfork?

I've been in both end, the incompetent contributor and the project owner who feels there is not enough competencies in the room. I'm sure there are many others who still have to deal with this dynamic. "We do not break userspace" may be old story, but the dynamic between is still relevant to this day and there is a lot to learn from the incident.

Re: We do not break userspace (2012)

#49
post #45

Do you want nice things? Because this is how we get nice things. You may not like it. I certainly prefer a nicer mode of communication. But this is peak infrastructure maintenance. Compare and contrast: https://news.ycombinator.com/item?id=29845724

It’s absolutely possible to communicate someone has made a mistake without insulting them. Nice things don’t come from degradation.

correct

they come from this level of passion for perfection

Post reply on HN