Not just any Code of Conduct
Not just any Code of Conduct
Posted Sep 17, 2018 1:13 UTC (Mon) by anotheruser (guest, #127270)Parent article: Kernel prepatch 4.19-rc4; Linus taking a break
This is how the author of the Contributor Covenant is reacting to the news: http://archive.is/6AyZy
(Log in to post comments)
Not just any Code of Conduct
Posted Sep 17, 2018 3:06 UTC (Mon) by anotheruser (guest, #127270) [Link]
Not just any Code of Conduct
Posted Sep 17, 2018 3:21 UTC (Mon) by himi (subscriber, #340) [Link]
Not just any Code of Conduct
Posted Sep 17, 2018 3:40 UTC (Mon) by anotheruser (guest, #127270) [Link]
Curious though, if you put ESR's warning with Ehmke's gloating and Torvalds' sudden capitulation and seeking of "assistance," do you not see any potential connection?
Not just any Code of Conduct
Posted Sep 17, 2018 4:05 UTC (Mon) by himi (subscriber, #340) [Link]
To Linus' great credit his response has been extremely positive and constructive, and it leaves me optimistic that the kernel community (and the wider Linux community) will move forward quickly on what has long been a serious issue.
Not just any Code of Conduct
Posted Sep 17, 2018 10:09 UTC (Mon) by nix (subscriber, #2304) [Link]
Not just any Code of Conduct
Posted Sep 17, 2018 12:44 UTC (Mon) by flussence (subscriber, #85566) [Link]
Not just any Code of Conduct
Posted Sep 17, 2018 19:47 UTC (Mon) by anotheruser (guest, #127270) [Link]
Not just any Code of Conduct
Posted Sep 18, 2018 16:34 UTC (Tue) by flussence (subscriber, #85566) [Link]
But if you weren't contributing to begin with, who cares how offended you are?
Not just any Code of Conduct
Posted Sep 18, 2018 22:48 UTC (Tue) by anotheruser (guest, #127270) [Link]
According to your logic, if I don't approve of the CoC that you insist on, I approve of indecent behaviour. This kind of rhetoric is why we can't have reasonable discussions about these issues.
> But if you weren't contributing to begin with, who cares how offended you are?
One of the primary rationales for the CoC is to avoid offending others who might otherwise contribute. See nix's comments on this page for a personal example. So I don't see how you can support the CoC while saying that you don't care how offended I might be. The whole point of the CoC is to care about offending others. Are you saying that it's okay to offend me, but not him?
Not just any Code of Conduct
Posted Sep 18, 2018 23:15 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]
Well, in your universe a mild CoC which basically boils down to: "Be polite to others and keep your political/religions/... opinions to yourself in project-related forums" is just like Marxism.
So why not?
Not just any Code of Conduct
Posted Sep 19, 2018 20:11 UTC (Wed) by anotheruser (guest, #127270) [Link]
It's almost humourous how religious your arguments sound. If only you could hear yourself.
Not just any Code of Conduct
Posted Sep 20, 2018 23:06 UTC (Thu) by flussence (subscriber, #85566) [Link]
It sounds like you're the one who wants the CoC, because you seem to be demanding others respect you for nothing but poison in return, which isn't how a meritocracy works.
Not just any Code of Conduct
Posted Sep 17, 2018 14:20 UTC (Mon) by anotheruser (guest, #127270) [Link]
Not just any Code of Conduct
Posted Sep 17, 2018 15:04 UTC (Mon) by smurf (subscriber, #17840) [Link]
Sorry, but if I saw a problem, tried to fix it, got stonewalled left and right, and then – after literally years of fighting windmills – the mill's owner *finally* admits that my way is the right way after all, and thus I can let Rosinante rest for a while, well, I'd probably snark-or-gloat-or-whatever for a bit, too. That's basically human, and doing so doesn't hurt anybody.
But: please note that she doesn't snark about Linus, and she doesn't gloat about Having Won. Instead, she snarks about the guys (I seriously doubt this set includes any not-guys) who will now leave the kernel project because OH NOES MY FREE SPEECH RIGHTS GET RESTRICTED!!! PREPARE FOR THE END OF THE WORLD AS WE KNOW IT !!!!!!!11! and she gloats about a situation which is going to drive off the idjits, who IMHO deserve to be driven off.
Again IMHO: Good. Let them leave. The kernel is better off without them, long-term, and people who cannot (or do not want to) realize that their behavior drives off more contributors than they're themselves worth, long-term, need to leave. Or, if they won't change, to be left.
Not just any Code of Conduct
Posted Sep 17, 2018 16:04 UTC (Mon) by kmweber (guest, #114635) [Link]
No, don't you understand? This is a WOMAN! Drinking ALCOHOL! After being ASSERTIVE! How dare she!
Not just any Code of Conduct
Posted Sep 17, 2018 21:36 UTC (Mon) by nix (subscriber, #2304) [Link]
They might be back when they realise they can't really get employed anywhere if they act like they've been acting heretofore, and fix themselves as Linus is trying to.
Not just any Code of Conduct
Posted Sep 17, 2018 17:51 UTC (Mon) by liam (subscriber, #84133) [Link]
The only context that came to mind where that tweet might be seen as having been gloating was if Linus had also been an outspoken teetotaler.
Not just any Code of Conduct
Posted Sep 17, 2018 10:59 UTC (Mon) by dskoll (subscriber, #1630) [Link]
I never much respected him. I respected some of his work, but never cared for his personality or politics.
Not just any Code of Conduct
Posted Sep 17, 2018 11:02 UTC (Mon) by dskoll (subscriber, #1630) [Link]
By "him" above, I mean esr. It may not be obvious because of the intervening comments.
Not just any Code of Conduct
Posted Sep 17, 2018 12:28 UTC (Mon) by ncm (subscriber, #165) [Link]
Late is better than never.
Not just any Code of Conduct
Posted Sep 17, 2018 16:18 UTC (Mon) by Wol (subscriber, #4433) [Link]
As a psychologist, a documenter of Linux, esr wasn't bad at all. I don't know of any recent work.
As a programmer, I'm not sure. He wrote fetchmail (or rather, largely rewrote it). Like many programmers, I think his ego was a bit larger than his ability.
Talk about guns ... well there's a lot of gun-nuts out there. He's just another idiot.
Respect him for his strengths and recognise we ALL have failings. There's a fair few here suspect I'm all failings, and a fair few who think I'm reason and fairness incarnate :-)
Cheers,
Wol
Not just any Code of Conduct
Posted Sep 17, 2018 20:13 UTC (Mon) by roc (subscriber, #30627) [Link]
Not just any Code of Conduct
Posted Sep 17, 2018 20:34 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link]
How so?
Not just any Code of Conduct
Posted Sep 17, 2018 21:40 UTC (Mon) by roc (subscriber, #30627) [Link]
It was far more optimistic than warranted about the power of voluntary community contributions to drive the development of large-scale projects. This led many projects (e.g. Netscape/Mozilla) to try to harness this power that did not, in fact, materialize.
CatB suggests that given a set of contributors, the extremely decentralized "bazaar" model not only works but is *efficient*. For example:
> In practice, the theoretical loss of efficiency due to duplication of work by debuggers almost never seems to be an issue in the Linux world.
Does anyone still believe this?
The related concept "given enough eyeballs, all bugs are shallow" has proven to be nonsense.
In retrospect I think it's clear ESR's libertarian politics and antipathy to Microsoft overly skewed his observations.
Another effect that can be blamed on CatB, or at least the ideas behind CatB, is that many open source projects (including Mozilla and Linux) were late to realize the importance of automated testing. CatB is completely ignorant of automated testing (fair enough, it was little practiced in that era), but by encouraging projects to rely on distributed testing by downstream users, it discouraged the development of continuous automated testing, which needs centralized infrastructure and developer attention to be most effective.
Not just any Code of Conduct
Posted Sep 17, 2018 21:50 UTC (Mon) by mtaht (guest, #11087) [Link]
Having lived through the pre-open-source era, faced with firing over using or contributing to free software when I worked at Borland and SCO...
I really would love to read a CatB, annotated, with what we've learned since 1998. I certainly remember what a breath of fresh air eric's books were and how they sparked the world we live in today - and I'm willing to forgive him (and everybody else) for not getting it all right.
In part it was the new set of analogies (notably the successful attack on the mythical man-month) that resonated then and continue to resonate now.
Not just any Code of Conduct
Posted Sep 17, 2018 22:12 UTC (Mon) by roc (subscriber, #30627) [Link]
Not just any Code of Conduct
Posted Sep 18, 2018 14:44 UTC (Tue) by Wol (subscriber, #4433) [Link]
And like most nonsense, it's actually a straw man - a false argument set up by detractors so they can pretend to destroy it.
Given enough eyeballs, all bugs REALLY ARE shallow. How long, from discovery to initial fix, does the typical serious bug in linux survive? HOURS!
The problem is that detractors assume that there are a lot of friendly eyeballs scanning open source code looking *for* bugs. Guess what, there aren't! But once someone stumbles across a bug and reports it, the more eyeballs it attracts (people it affects), the more eyeballs looking *at* the bug, the quicker it gets fixed.
Can you name EVEN ONE serious bug that, once it was spotted, survived more than a day or so without being dissected to death?
Cheers,
Wol
Not just any Code of Conduct
Posted Sep 18, 2018 15:11 UTC (Tue) by farnz (subscriber, #17727) [Link]
Spectre/Meltdown mitigations. Months from the bug being spotted in Linux to the initial fix being landed, and even then, the initial fix was only accepted because the bug was so serious that something had to be done.
Not just any Code of Conduct
Posted Sep 18, 2018 15:44 UTC (Tue) by mtaht (guest, #11087) [Link]
Still:
I've long had a corollary to Linus's law: With enough bugs, all eyeballs are shallow.
The enormous expansion of the Internet, the codebases, and the users has been far greater than the corresponding expansion of the developers capable of coping with it.
Folk with the core architectural knowledge to systematically design robust systems are few and far between. Having a holistic view of the flaws in the whole system is rare. The core devs are aging. The internet culture could use something like a fire department, that could pull in (and PAY) volunteers when an emergency like spectre happens. It could use a civil engineering department, also.
I lost 3 weeks of vacation and a great deal of sleep during spectre. I was burnt to a crisp for about three months, (I'd needed the vacation in the first place), while losing time pulling some core assets out of the cloud completely (because it still ain't fixed enough) and in the end I was late on a paid project and got fired for it.
I look at how firemen are respected, and compensated, and dearly wish we had that kind of societal infrastructure holding the net together also.
Not just any Code of Conduct
Posted Sep 18, 2018 22:22 UTC (Tue) by rodgerd (guest, #58896) [Link]
Suspend/resume/hibernate, many "write data/fsync()" bugs that degenerate into blamestorming and nitpicking whether the POSIX spec makes it OK to corrupt people's data, data loss bugs in early iterations were not trivial, btrfs is a series of data loss problems. And that's just off the top of my head.
Many of these problems are hard, and can only be solved by a fairly small group of people.
Not just any Code of Conduct
Posted Sep 19, 2018 13:42 UTC (Wed) by Wol (subscriber, #4433) [Link]
The problem here though often seems to be political, and such the code isn't buggy. "It's not working!". "It's working as per spec!". "Then fix the spec!". "Can't - I don't have the power!".
How many of your examples are actually "features" - where the unwanted behaviour is actually an inevitable side-effect of poor design?
It's not a bug IN linux, if it's trying to work around someone else's problem (which is one of the big arguments in favour of Open Source ... :-)
Cheers,
Wol
Not just any Code of Conduct
Posted Sep 19, 2018 14:14 UTC (Wed) by farnz (subscriber, #17727) [Link]
That's not what I see when I look at the long standing bugs I have personally fixed.
For example, this fix to PPPoE was something that had been buggy for ages, but no-one had bothered to look for a good fix. this IRQ issue in VIA drivers was a straight foul-up that had been present in Linus's tree for at least a year. Here's a driver that kept locking up at random until I spent a few weeks digging into exactly what was wrong.
And I'm not a prolific contributor to the Linux kernel - these are the bugs I fixed where (a) I know that the bug had been present for at least 12 months, because I could find reports of them dating back to at least 12 months before I fixed them (in some cases, my co-workers had made the reports), and (b) the bug is entirely in the Linux kernel, and not (e.g.) in external hardware (although I encountered other bugs where with buggy hardware, things just failed, but the kernel was being slow for bad reasons: e.g. this DisplayPort bug ; the kernel used suboptimal transfer sizes, but had a comment saying that optimal transfer sizes don't help. That comment was outright wrong, and meant that I spent more time than I wanted to working out just WTF Linux was doing that the hardware did not expect).
Not just any Code of Conduct
Posted Sep 20, 2018 11:27 UTC (Thu) by Wol (subscriber, #4433) [Link]
In other words, "no eyeballs" :-)
So yes, linux development failed in this scenario, but it was lack of motivation and eyeballs, not a failure of "given enough eyeballs".
Classic abuse of a saying - it's why I hate most people using the saying "a bad craftsman blames his tools" and "the exception proves the rule". In original use, the craftsman made his own tools, and the exception is a scientific proof - find an exception and your rule has to be rethought.
Cheers,
Wol
Not just any Code of Conduct
Posted Sep 20, 2018 11:35 UTC (Thu) by farnz (subscriber, #17727) [Link]
Only if you redefine "given enough eyeballs" to mean that attempts to fix the bug that fail are also not "eyeballs" on the code. Other people (some of whom had far bigger incentives than I did) had looked at the code and tried to fix it before me. I happened to be the person who fixed it because my Linux experience (fixing random bugs that affected me or my employer) gave me a much better global view of what was possible than previous people who'd been focused on just the networking subsystem.
People had looked for a fix, failed because they were unfamiliar with workqueues and thus could only produce bad fixes, and then given up because it was a hard problem if you exclude the facility meant to make it possible. No-one before me took the step further to create a good fix - I just had the knowledge needed to take the idea behind someone's bad fix and do a good job of it.
And remember that the original context in which I'm replying was Wol asking "Can you name EVEN ONE serious bug that, once it was spotted, survived more than a day or so without being dissected to death?". I can name three which were spotted at least a year before I fixed them, where people before me made attempts to fix them, and in one case actually introduced the bug I fixed, while in two others they simply failed to fix them and the bug got left alone for over a year (which is considerably more than a day or so) before I fixed it.
Not just any Code of Conduct
Posted Sep 18, 2018 16:08 UTC (Tue) by bfields (subscriber, #19510) [Link]
I feel like I need to introduce you to bugzilla. Or my inbox. If you're feeling bored, I'm sure I could find a few bugs for you to work on.
Not just any Code of Conduct
Posted Sep 18, 2018 21:42 UTC (Tue) by roc (subscriber, #30627) [Link]
It's not a straw man, it's a direct quote from the document.
> Can you name EVEN ONE serious bug that, once it was spotted, survived more than a day or so without being dissected to death?
Sure, my Skylake laptop's Intel graphics hang temporarily (sometimes permanently) under heavy load. This has been going on for years, lots of bugs have been reported on freedesktop.org, lots of people have been involved, the problems have sometimes gotten better, sometimes worse, but never gone away. It is probably a combination of different bugs that change over time.
The problem is of course that a lot of those "eyeballs" aren't actually developers, and a lot of the rest (including me) are but don't have the time or inclination to get up to speed on the Intel driver code. I even looked at it, and found it would just take me a long time to learn enough to contribute anything useful.
You may say "ESR meant 'trained eyeballs who know the relevant code and who actually do look for the bug(s)'". To that, I say in that case his maxim has not scaled and cannot scale to be meaningful in a project the size and complexity of the Linux kernel today.
Not just any Code of Conduct
Posted Sep 21, 2018 1:10 UTC (Fri) by flussence (subscriber, #85566) [Link]
(My i945 also has the exact same issue… bought it back when Intel was hyped up as having the best Linux graphics drivers and almost immediately got burned. Nowadays I have to choose PC hardware based on who's creating less horror stories and flag days.)
Not just any Code of Conduct
Posted Sep 18, 2018 18:51 UTC (Tue) by k8to (guest, #15413) [Link]
I think it's a matter of being oversold rather than entirely wrong in vision.
Not just any Code of Conduct
Posted Sep 17, 2018 13:16 UTC (Mon) by jubal (subscriber, #67202) [Link]
If that is a joke on your part, it's a very bad joke, and rather badly executed.
Not just any Code of Conduct
Posted Sep 17, 2018 5:30 UTC (Mon) by Duncan (guest, #6647) [Link]
Hmm... the HTTPSEverywhere extension has a rule for that site that redirects to https (and if it didn't, SmartHTTPS would try, as it does for all unencrypted links, tho if it it fails to get an https connection it whitelists it as http-only and doesn't try again), but neither Chromium nor Firefox like the encryption -- presumably it's either outdated and both browsers are phasing it out.
Chromium (69.0.3497.81, gentoo ebuild): Unsupported protocol: ERR_SSL_VERSION_OR_CIPHER_MISMATCH (Details: The client and server don't support a common SSL protocol version or cipher suite.)
Firefox (62.0 upstream binary as provided by gentoo firefox-bin ebuild): No common encryption algorithms: SSL_ERROR_NO_CYPHER_OVERLAP.
Links and Lynx yield similar errors when trying https.
But trying it unencrypted/http doesn't get much farther. All four browsers give me a DNS error, error 1001 or 1004 from Cloudflare, depending on whether cookies are allowed from archive.io.
The contributor-covenant.org link to the document worked, tho, just not the one to the author's opinion at archive.is.
Not just any Code of Conduct
Posted Sep 17, 2018 10:23 UTC (Mon) by jwilk (subscriber, #63328) [Link]
TLS_ECDHE_RSA_WITH_AES_128 GCM_SHA256, 128 bit keys, TLS 1.2
Anyway, this is the original page: https://twitter.com/CoralineAda/status/1041441155874009093
Not just any Code of Conduct
Posted Sep 17, 2018 18:55 UTC (Mon) by csos95 (guest, #123148) [Link]
There's some issue with the archive.is nameserver responses to cloudflare requests https://community.cloudflare.com/t/archive-is-error-1001/...
Not just any Code of Conduct
Posted Sep 21, 2018 8:32 UTC (Fri) by Duncan (guest, #6647) [Link]
Not just any Code of Conduct
Posted Sep 18, 2018 13:55 UTC (Tue) by dskoll (subscriber, #1630) [Link]
That "Hahaha" tweet looked like a joke or satirical remark to me.