|
|
Log in / Subscribe / Register

Sharp: Closing a door

Sharp: Closing a door

Posted Oct 6, 2015 22:09 UTC (Tue) by andreasn1 (guest, #88420)
In reply to: Sharp: Closing a door by neilbrown
Parent article: Sharp: Closing a door

I stumbled upon this (via twitter) from last month, and it stood out to me as a bit unnecessarily rough:
"Christ, people. Learn C, instead of just stringing random characters
together until it compiles (with warnings)."

https://lkml.org/lkml/2015/9/3/428


to post comments

Sharp: Closing a door

Posted Oct 6, 2015 22:46 UTC (Tue) by neilbrown (subscriber, #359) [Link] (6 responses)

Thanks for the link.

Linus goes on to say

> So I can see how this bug happened, and I am only slightly upset with Lorenzo who is the author of that commit.

and it is very quickly revealed that while gcc-5.1.1 gives a warning, gcc-4.9.2 doesn't.

So that attack from Linus, which you quoted, seems to be an emotional response which seems to not be justified by the facts, and is an accusation of much more incompetence than the evidence shows even in the most extreme interpretation ("just stringing random characters together").

I agree that is uncool.

Does the "Code of conflict" help us here? It only triggers if someone feels "personally abused, threatened, or otherwise uncomfortable". I wonder how Lorenzo felt. Apparently someone felt enough about it in some way to tweet it.
Is it worth a formal complaint from someone to test the process?
Would it be appropriate for the TAB to approach people involved in the conversation to check if they felt "personally abused, threatened, or otherwise uncomfortable" rather than just leaving it up to the individuals to raise the issue?
(maybe they already have and everyone is happy)

Sharp: Closing a door

Posted Oct 8, 2015 2:29 UTC (Thu) by djbw (subscriber, #78104) [Link]

The "code of conflict" is simply the wrong polarity of burden compared to a "code of conduct". You can point someone who is being abusive to a code of conduct and say "please stop". Otherwise, you're expressly telling the recipient of abusive behavior "put up with it until it rises to the point of injury that you feel you need to involve a disinterested third party to mediate". I submit no one will ever meet that threshold before simply moving on to another project.

I'm less concerned with long standing kernel developers being abusive to each other compared to the observers of those interactions. My anecdotal experience is that these observers of violence on LKML go on to emulate it thinking "that's how you command respect and get your way". They completely miss the point that the respect and "getting your way" comes from a proven track record of being almost always correct from a technical perspective.

Sharp: Closing a door

Posted Oct 8, 2015 18:43 UTC (Thu) by mgross (guest, #38112) [Link] (4 responses)

I just read that post. I see nothing wrong with it and I'm rather impressed with the amount of effort put into the email to explain things and educate people on things that are important.

Does anyone honestly think we would be better off if he just replied-all with a "please check your compiler warnings"?

I don't.

BTW I think CoC's are Marxist "right-think" tools of censorship and CoC supporters are either confused, evil or both.

Sharp: Closing a door

Posted Oct 8, 2015 23:16 UTC (Thu) by neilbrown (subscriber, #359) [Link]

Indeed, the post has many excellent qualities. After all these years Linus is still passionate about the kernel and the quality of code. It's good to see.

It is just unfortunate that it is spoiled by an irrational and unjustified accusation of extreme and deliberate incompetence.

Had Linus, on proof-reading before posting, simply deleted the line about "stringing random characters
together until it compiles" it would have done nothing to reduce the effectiveness of the email, and much to reduce the possibility of it causing harm.

I understand that Linus believes that the passion he displays is an important part of his communication style, and he may well be correct. There is a very simple rule that allows you to maintain emotional passion, but avoid (or substantially reduce) the risk of harm:

Make "I" statements, not "you" statements.

"I can't believe I still get patches which add new warnings"

This is only one line in an otherwise very productive email. But it is unarguably attacking "people" (and by implication one or two specific people) rather than attacking the code. And "attack code, not people" seems to be an informal code-of-conduct that is largely agreed too.

> BTW I think CoC's are Marxist "right-think" tools of censorship...

I'm a little surprised that you are agreeing to use the same language as the rest of us - I'm sure it restricts your ability to think freely....
FWIW, I think CoC's are useful tool for unambiguous communication of group norms. It is intrinsic in the human condition that we sacrifice individual freedom to benefit from group support. In situations where body language cannot communicate/mediate those norms, a CoC can help.

Sharp: Closing a door (on the Marxist right)

Posted Oct 10, 2015 1:12 UTC (Sat) by neilbrown (subscriber, #359) [Link] (1 responses)

> BTW I think CoC's are Marxist "right-think" tools of censorship and CoC supporters are either confused, evil or both.

I kept thinking about this ... partly because I wasn't really happy with my flippant response (sorry), partly because seeing "Marxist" and "right" in the same sentence seems odd, but mostly because I need something to think about as I cycle around the bay.

I guess your reference to "Marxist" draws a parallel with the contrast between a controlled economy and a free market. Having a Code of Conduct is a bit like a controlled economy - you are only allow to do that which is within the rules. You, I presume, would prefer a more free-market approach where people can say what they like but must suffer the consequences.

While the free-market has a lot of strengths, it has requirements that are not always met - particularly various freedoms and transparencies. There are some edge cases where it doesn't work at all, such as fly-by-night confidence tricksters and monopolies. Most (all?) free markets still have regulation to try to avoid the worst excesses of these.

In the kernel community, a senior maintainer - and particularly Linus - has an effective monopoly. The result is that market forces don't really work ... or they work by people leaving the community and going to work in some other project which isn't healthy for the kernel community.

So people suggest the Marxist controlled economy of a Code of Conduct. Is that the only way?

One idea I came across a while ago that applies to leadership teams is "Ask for forgiveness, not permission". The idea is that if all the team members keep having to ask permission to do things, nothing much gets done. If instead they are given freedom to follow their own ideas, with the understanding that forgiveness can be negotiated if things go wrong, then the net result is much more positive.

Applying that to the kernel community, I imagine the community as a whole being the source of permission or forgiveness, and the individual kernel maintainers (from Linus out toward the fringes) being the team members who might mess up occasionally and so might need permission in advance, or forgiveness after the fact.

A CoC promotes the "You need permission first" approach and maybe that it a good reason to reject it (in the context of a smallish community of maintainers - it may be more appropriate else where).

So how would the "ask for forgiveness" approach apply? It would mean that team members who (whether through carelessness or malice) harm the community (or part of the community) would need to acknowledge the harm and seek forgiveness, and the community should be ready to forgive.

A lot of the infractions that are mentioned in these discussions are from years ago. This suggests the community is not eager to forgive. But then there has been little or no acknowledgement of harm (not even "I don't see how what I said was hurtful, but I can see that you are hurt, and I'm sorry for that hurt whatever its cause") or requests for forgiveness.

Are we, as a community, capable of compassion? Capable of acknowledging we may be complicit in some hurt? Capable of seeking and providing forgiveness? Or do we drive away all the people with those abilities?

Sharp: Closing a door (on the Marxist right)

Posted Oct 11, 2015 17:11 UTC (Sun) by nix (subscriber, #2304) [Link]

Just a note to say that analysis like this is why I still read LWN comments, despite their increasingly cesspoolish nature. If LWN had an adaptive scorefile you'd be at or near the top :)

Sharp: Closing a door

Posted Oct 10, 2015 3:16 UTC (Sat) by bronson (subscriber, #4806) [Link]

> CoC supporters are either confused, evil or both.

Add 'desperate'? In the past six months the Atom editor community grew faster than the stewards could keep up. I'd guess that the leaders felt like they needed to do something and do it fast. Their first choice blew up in their face (IMO rightfully so), but the second one seems pretty harmless.

Maybe this falls under 'confused.'

In general, I agree: CoCs end up getting gamed by trolls, encouraging SJWs, and causing way WAY too much useless discussion. But it's a very real problem that they're trying to solve, and they're among the very few tools that project leaders can turn to. If you have a workable alternative, I'm sure a lot of people would be interested.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds