Observed effects on Kernel code from Code of Conduct (CoC)

Synesthesia · 04-17-2019, 09:04 PM

As most of you know, the new code of conduct for the Linux kernel was written and implemented by self-proclaimed social justice warriors (SJWs).

I would like to hear from anyone who knows about the effects it has had on kernel development and code quality so far. Some have claimed that it has no effect; just to appease some people - but it has quite a lot of potential to cause insidious effects, I think.

Without going into theory, I would just like to know the real changes to Linux you think or can demonstrate have been a result of the CoC.

hazel · 04-19-2019, 09:50 AM

What kind of thing do you have in mind? PC slogans in the comments where there used to be effing and blinding? Not very likely!

Or a rash of CVEs perhaps. If there was a bad kernel bug, I suppose you could say, "Well, Linus wouldn't have let that one through." But you or I wouldn't be able to identify it. We don't know enough about programming.

And we still wouldn't know if it had anything to do with the new CoC unless the programmer responsible turned out to be a black transgender Muslim with learning difficulties who had been recruited to make the team less meritocratic.

Synesthesia · 04-21-2019, 01:38 PM

Sarcasm noted. I have in mind exactly as I said: changes anyone thinks would not be present in code without the new policy.

RickDeckard · 04-21-2019, 03:08 PM

But wouldn't that factor only be known if the developer in question made a huge issue of their changes being tied to who they are and threatened to dox half the LKML?

berndbausch · 04-21-2019, 09:24 PM

Quote:

Originally Posted by Synesthesia

Changes anyone thinks would not be present in code without the new policy.

You seem to think that rudeness is a necessary requirement for generating good code. Since you don't offer any examples yourself, I have the strong feeling you are chasing non-existing ghosts.

Quote:

the new code of conduct for the Linux kernel was written and implemented by self-proclaimed social justice warriors (SJWs).

Why don't you name those SJWs?

Code:

author	Greg Kroah-Hartman <gregkh@linuxfoundation.org>	2018-09-15 20:26:44 +0200
committer	Linus Torvalds <torvalds@linux-foundation.org>	2018-09-16 11:42:28 -0700

and

Code:

Signed-off-by: Chris Mason <clm@fb.com> 
Signed-off-by: Dan Williams <dan.j.williams@intel.com> 
Signed-off-by: Jonathan Corbet <corbet@lwn.net> 
Signed-off-by: Olof Johansson <olof@lxom.net> 
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org> 
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Disclaimer: I am not involved with kernel development.

Turbocapitalist · 04-22-2019, 12:13 AM

Look, the question is important even if you might think that the policy changes were meant to help. Don't be distracted by bringing name-calling and other stuff, that's what the opponents like. Here's why:

0. The LF is a trade association representing its members' interests.
1. The LF has had many opponents/competitors join as members.
2. Its harshest long-term opponent is now on the board.
- 2a. The board supervises the executive director.
- 2b. The executive director supervises the technical lead.
3. The technical lead was replaced temporarily back in the autumn.
4. A large number of patches he was holding back went through during his absence.
5. The technical lead is back but in a lesser capacity, informally.

Assuming that there have not been efforts to obfuscate the companies of origin for patches, it might be possible to identify if there has been a shift, large or small, in which corporations have been committing and / or what new areas are suddenly experiencing more or fewer commits than before.

Because of the scale of the code base, these are not anomolies that will be detectable manually.

For example, if you want to see the scale, take a look with Gource. A full run will take hours, but you can look at recent years just as well and that will only take tens of minutes:

Code:

git clone git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git

cd linux-stable

`gource --log-command git` > ../linux-foundation.log

cd ..

gource --fullscreen --start-date 2017-01-01 ./linux-foundation.log

That is not the right tool for the job. That won't show the trends that should be exposed, if present, but it will show the scale of the beast.

Someone with statistical skills who can work with the right tools is needed.

berndbausch · 04-22-2019, 01:00 AM

Quote:

Originally Posted by Turbocapitalist

Assuming that there have not been efforts to obfuscate the companies of origin for patches, it might be possible to identify if there has been a shift, large or small, in which corporations have been committing and / or what new areas are suddenly experiencing more or fewer commits than before.

A power struggle over the control of the world’s most used operating system kernel? Commercial interests in, for and against Linux? Not really surprising. It would surprise me if the struggle had an “SJW” agenda at its core.

frankbell · 04-22-2019, 08:34 PM

I love it when courtesy, decorum, and professional conduct are libeled "political correctness."

Turbocapitalist · 04-22-2019, 10:39 PM

Ignoring the malicious posts above, there seem to be some tools for analyzing Git but not much choice in regards to actively developed tools:

Hercules -- active but not in the repository
Gitstatistics aka Gitstats -- last updated 2009
Gitinspector -- last updated 2017

Are there any others which may be better?

There are also raw numbers about which can be gleaned with git itself, which :

Code:

man git-shortlog
git shortlog -s -n -e --no-merges --since "2018-07-01" --until "2018-07-31" \
| perl -n -e 'chomp;($c,$e)=m/([0-9]+).*<([^>]+)>/;$e=~s/^.*\@//;print $c,"\t",$e,"\n";' \
| perl -n -e 'chomp;($c,$e)=split("\t",$_); $m{$e}+=$c; END{foreach $n (sort{$m{$b}<=>$m{$a}} keys %m) {print $m{$n},"\t",$n,"\n";}}'

However, those numbers without more context would provide little useful information beyond which company is committing. Number of commits by itself is not a valuable metric, and there is also the more-complex-than-necessary problem of normalizing the host names. For example, linux.vnet.ibm.com
linux.ibm.com, de.ibm.com, us.ibm.com, and au1.ibm.com are all ibm.com I am guessing that git-log would be the next place to look along that route.
That gets too close to re-inventing the wheel. Hercules is further along.

The code base is what matters so it is far more useful to look at what parts of the code itself have been changing.

berndbausch · 04-23-2019, 12:18 AM

Quote:

Originally Posted by Turbocapitalist

Ignoring the malicious posts above, there seem to be some tools for analyzing Git but not much choice in regards to actively developed tools

You are off-topic. While the tools you propose allow us to analyze corporate commits (which I consider very interesting), they don't show the SJW impact on the Linux project. My malice, which I readily admit, was not directed at your laudable attempt to find evidence for a corporate and hostile take-over of Linux, but at the idea of an SJW conspiracy that replaces the testosterone-laden bloke-culture and locker-room humor of software development with an LGBTQ+whatnot dictatorship.

Turbocapitalist · 04-23-2019, 12:35 AM

M$ has worked by proxy for decades and there is no reason for them to stop now. Similar for other harmful interests. Don't be distracted by the SJWs. As noisy and harmful as they are they are just catspaws in this mess.

RickDeckard · 04-24-2019, 02:18 PM

There's a difference between basic human politeness and authoritarian mandates of oversensitivity in service to a worldview. I strongly believe that social justice warriors traffic in the latter.

It's fine to be understanding, it's good to be a decent and loving human being, but not to force it with threat of sanction as other institutions (Evergreen State College, for instance) have experienced from these types of people. Given their propensity for violence in one form or another when they don't get their way, I've always had bad feelings about this.

To have these feelings is not to "consider rudeness a requirement for writing good code" or any such garbage.

berndbausch · 04-24-2019, 06:51 PM

Quote:

Originally Posted by RickDeckard

There's a difference between basic human politeness and authoritarian mandates of oversensitivity in service to a worldview. I strongly believe that social justice warriors traffic in the latter.

I am sure that there is ample supply of people who want to force their worldview onto others. However, while the CoC was drafted and implemented by SJWs like Kroah-Hartman(*), this thread is not about SJWs but the impact of the CoC. And this is the kind of interaction that the CoC discourages:

Code:

+* The use of sexualized language or imagery and unwelcome sexual attention or
+  advances
+* Trolling, insulting/derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others’ private information, such as a physical or electronic
+  address, without explicit permission
+* Other conduct which could reasonably be considered inappropriate in a
+  professional setting

Do you see this as an authoritarian mandate of oversensitivity?

Quote:

To have these feelings is not to "consider rudeness a requirement for writing good code" or any such garbage.

My thinking is that the above points are examples of rudeness and that asking people to avoid them is not oversensitive. Others might disagree.

Quote:

not to force it with threat of sanction

It is true that the earlier CoC did not spell out that sanctions might be applied. I don't know if there are examples of developers that were threatened by the Technical Advisory Board, either under the former or the current CoC.

(*) in case this isn't clear, I don't think that Kroah-Hartman is an SJW. He is however the author of the new CoC, which was written and implemented by self-proclaimed social justice warriors (SJWs).

Turbocapitalist · 04-25-2019, 12:36 AM

In contrast to Linus Torvalds not wishing to change Linux to please M$, Greg K-H has been (mis-)using his skill and position to advance M$ interests within the kernel for many years. That he is also behind the CoC makes it that much more suspicious. I am now quite curious as to what went through in the 4.19 kernel that he oversaw and what, if anything, strange new additions are in the subsequent releases made after Linus' "return". Something with UEFI maybe? Or HDCP or other digital restrictions?

ondoho · 04-26-2019, 11:53 AM

Quote:

Originally Posted by berndbausch

the CoC was drafted and implemented by (...) Kroah-Hartman

i beg to differ.
https://web.archive.org/web/20180925...ode-of-conduct
iirc it is not the exact same text, but it is very much inspired by that.
in light of other events when all this went down this is very obvious to me.

Quote:

Do you see this as an authoritarian mandate of oversensitivity?

not in the COC itself, but some serious sh!t has been hitting the fans that for a moment made me scared to be a male white heterosexual in an online world.
I do not want to regurgitate, but it's all there, here on LQ (welcome to check my posting history) and everywhere. IIRC it was half a year ago?