Counterexample: Linux, GCC, Python, and practically the entire free software ecosystem from before the current crazy for hall-monitory-y supervision from above.
It is simply demonstrably, factually, clearly not true that a growing community needs the kind of structures that Rust imposed on itself.
It really makes me sad that a certain kind of person these days sees some kind of censorious overlord as essential for the formation of healthy communities.
> There's no shortage of guidance to "stay away from X tech because the community is toxic and non-serious"
A disaffected and loud minority says things like that, and the rest of the world goes right on ignoring them. Zero people in the real world avoided using the Linux kernel because Linus was brusque.
> Zero people in the real world avoided using the Linux kernel because Linus was brusque.
Actually, the best example of a project where the leadership of the project was so toxic as to drive away potential contributors would probably be glibc under Ulrich Drepper, which got so bad that most distributions abandoned glibc for the eglibc fork. (See https://lwn.net/Articles/488847/ for a high-level discussion).
Drepper was an asshole and people eventually routed around him, yes. The system worked. What people forget, too, is that Drepper's problem wasn't just an obnoxious personal style, but a ridiculous level of technical conservativism that led to critical bugs remaining unfixed for years. It was the latter problem that eventually prompted people to fork glibc, not the former.
I know very little about NetBSD and am a great fan of OpenBSD.
But Theo is a very horrible person who often attacks people in a personal way for disagreeing with him. I got into a stupid argument with him over fundraising - a subject where I have deep experience - and it was quite bizarre. He had a set of assumptions and disagreeing with them got abuse from him, and some of his minions on the OpenBSD-Misc list, and (I was astonished) abuse in my INBOX from throw away accounts.
Perhaps no user avoided it (though that seems unlikely), but can't you imagine why some contributors may have avoided it? Wouldn't that lack of potential contributions be a material loss for the project?
Who's to say that more contributors are turned off by Rust-style behavioral micromanagement? It's impossible to prove counterfactuals.
All we can say for sure is that dozens of critical projects in the past reached an amazing level of quality and importance to humanity without tone police lurking in the background and supervising it all.
I do wonder whether there's not some implicit benefit to this kind of management. Is it possible that by dissuading all but the most confident committers the project's contribution team self-selected for stronger devs? That would probably be bad for small OSS projects and good for kernels.
These "internet activists" you spend so much effort maligning are simply reminding you that there is a human being on the other side of that screen.
I'm not inside the Linux developers' world, but from the outside it seems like a much healthier, more vibrant place since Linus realized that he works with human beings.
> These "internet activists" you spend so much effort maligning are simply reminding you that there is a human being on the other side of that screen.
Actually, primarily they're reminding me that there's a human being on the other side of the screen watching everything I do in case I fuck up. How this is supposed to make anyone want to participate is a mystery to me.
edit: The weird thing to me is that these people are so hyper-vigilant to the damage bad behavior can do, and utterly blind to the idea that their own behavior can also be damaging. If I ever read a sentence like "we know that overmoderation and tone-policing can create toxic communities, and we're watching out for that" from a moderation team, I will know that this is a community that I can trust to be administered in an even-handed and fair manner. So far I have seen this once.
Well, the question is rather- what is a "fuck up" to them?
And to that- who knows? Certainly the point of a CoC is supposed to be to codify this, but I believe experience shows that its interpretation tends to be expansive, when the wording is not already expansive to begin with.
At the end of the day, events like Curtis Yarvin, a person who has never harmed a fly, almost getting banned from Lambdaconf over "safety" concerns, demonstrate that the fuck-up may just be having a political difference of opinion with the group in question.
(Analogously, and I say this as somebody who would vote Dems every time if they lived in the US, a moderation team that included at least one Trump voter would also assuage such concerns. Consider it a commitment to diversity.)
edit: To be clear, I am not asking for anything resembling quotas; just any demonstration of the ability of the team to coexist with a person they have serious ideological disagreements with.
I wasn't familiar with Curtis Yarvin, but in looking him up, you can't be serious, right?
>Yarvin's online writings, many under his pseudonym Mencius Moldbug, convey blatantly racist views. He expresses the belief that white people are genetically endowed with higher IQs than black people. He has suggested race may determine whether individuals are better suited for slavery, and his writing has been interpreted as supportive of the institution of slavery.
Curtis Yarvin is a bellwether - the sort of person that any group that starts excluding people for ideological disagreements, would probably exclude first precisely because his position is so problematic. So any group that accepts his technical contribution can obviously be trusted to tolerate any less-severe ideological disagreement. Conversely, any group that doesn't, especially when they have to make up nonexistent concerns to do it because their rules didn't cover this "obvious" reason to kick someone out and couldn't be hastily adjusted, must be viewed with caution.
I personally don't hold any beliefs nearly as objectionable as that. But I do hold objectionable beliefs - as I believe any halfway interesting person does. And those who don't, probably will eventually. Just stand by your convictions and give it time.
> Call me crazy, but I believe that you don't have to actively champion and invite openly racist people to conferences to show that you tolerate difference of opinion.
Sure you don't have to, but if you do, it's a hell of a signal. (At any rate, Curtis Yarvin was invited for his semi-esoteric functional-based distributed operating platform, Urbit.)
> If you're protecting personnel, even after a number of others in your community have shown disagreement with the person's actions (and protections afterwards), just admit you agree with those thoughts.
Sorry, I don't. Of course, you'll believe that I do anyway, and that's fine. I do think it's a bit sad that you think that the only reason someone could want somebody to be included, is because they were your ideological compatriots.
In fact, the only reason I want anyone to be included in a conference is if they have contributions to the conference's topic.
No, I think someone should be *excluded* from talking at a conference because they literally write Eugenics theory, regardless of the brackets, semicolons and spaces they write in a text editor.
It sounds like we just have a difference in moral standards.
I just disagree that standing in his presence is a matter of safety for anyone. It's possible to hold abhorrent views and still be a useful contributor.
Yes? They might have other useful ideas or opinions that I may benefit from being exposed to. People are multidimensional.
If there is a conference being organized about some technology, I'd like to see speakers who have the most to contribute, on that merit only. I couldn't care less if they march around with armbands on in their spare time. I'm suggesting that more people learn to compartmentalize.
To me, if they can keep it to themselves, they can believe whatever they want. Up to and including that I shouldn't have been born, though I may draw a line at believing I should be killed, depending on how mentally stable I believe them to be.
>I am not asking for anything resembling quotas; just any demonstration of the ability of the team to coexist with a person they have serious ideological disagreements with.
That would be nice... unfortunately, filter bubbles are such that it's hard for most people just to locate a reasonable person who has serious ideological disagreements with them. The current polarisation didn't happen in a vacuum.
Right, and the suggestion is that this human being (and all the other ones) should be responsible for managing their own emotional state, instead of shifting the burden onto everyone else.
One of the biggest contributors to R's success over the past decade is folks having negative experiences with the Python community, particularly folks who are women, non-white, or come from non-CS background. The R community (and RStudio in particular) has worked hard to be much more inclusive and you can see this clearly reflected in the diversity of users and package authors.
Linux with Linus, who famously had to change his abusive tone? GCC with Richard Stallman, accused of various kinds of sexual misconduct? Python, which felt it necessary to impose a CoC eventually?
What random accusations am I "spreading so wildly"? The fact that Stallman is being accused of various kinds of sexual misconduct is a fact. I did not anywhere claim they were true. I'm pointing it out because it makes the points above stand on shakier ground.
The article you shared also fails to defend many of the accusations against Stallman and completely ignores them, for example the "Emacs virgin girl" situation.
Technically, what you did was to insinuate, not directly accuse. Which one is worse?
Since you seem to have a great interest in this, do have any concrete reference to the “Emacs virgin” situation? I have only a vague recollection that Stallman was referring to anyone who had not used Emacs yet as an “Emacs virgin”, and some people took it as meaning some kind of sex thing.
(Any other references to the “many of the accusations against Stallman” that wasn’t referenced in the linked article would also be interesting to see.)
Linux and Python (not aware of GCC) effectively have BDFLs that can just nix anything (hence the "dictator" in BDFL). So these aren't just really comparable.
And they absolutely turn people off. The thing is, as long as it doesn't turn everyone off, it allows the project to move forward, because even with burned bridges, it leaves ownership clear.
Communal decision making, however, does not have that advantage. If both sides of an issue, so to speak, become turned off of one another, you are more likely to have an abandoned project. There are, of course, other advantages (you don't miss generally accepted "good ideas" because of the particular vision of one person, and you can apply community standards to everyone, rather than having to weigh "continued participation in this project that is important to you" vs "dealing with -that- asshole again"), but that is definitely one con.
Yeah it's comical because Linux has so obviously driven a lot of people away from kernel development. Tech is male-skewed, and OSS more so, but Linux kernel dev is even then still at the far end of the gender disparity spectrum.
> Yeah it's comical because Linux has so obviously driven a lot of people away from kernel development.
I hear this "a lot" very often, but then it seems to be from people who have no real interest in technical work of kernel/OS core development. Linux is not the only way to scratch your itch for interest in low-level system dev. Like, this is just personal experience, but I have heard this on the order of 50-100 times: someone parroting how toxic Linux kernel dev is because of drama they heard -- but then you kind of dig a little bit and see what kinds of software stuff interests them, what do they work on -- probably only once or twice has it been anything embedded, hardware related, close to the metal. I would need compelling evidence to change my opinion that most of the complainers have no interest in the work being done by the community they are complaining about -- and I am fully aware that a number of people have departed Linux development, but we are talking about a tiny number of the thousands of contributors over the years -- you can't please everyone.
The hobby OS, emulation and demo scene is a pretty good indicator for "natural"* gender breakdown. These tend to be tight, tiny communities or often lone wolves working on projects. It is male dominated. This can't be explained by any systemic or community gatekeeping - because there is no system nor any mandatory community for participation or distribution. Nothing prevents anyone from putting their work out there.
* I am not discounting there may be other systemic reasons that set up this condition - but it has to be societal conditions that are in place in early childhood -- something that happens a bit before one considers contributing to the Linux kernel.
Agree completly. I would add that not just they are not interested...they have nowhere near enough skills to do anything in the kernel / os development space
Are you aware of how many people have been driven away from Rust because of their community? I've never seen people talking about avoiding Linux even 1/10th as much as they talk about avoiding Rust. It's entirely due to the culture; developers don't like a culture in which a programming language needs a multi-member moderation team (who resigns because they can't punish people 100% of the time).
It is simply demonstrably, factually, clearly not true that a growing community needs the kind of structures that Rust imposed on itself.
It really makes me sad that a certain kind of person these days sees some kind of censorious overlord as essential for the formation of healthy communities.
> There's no shortage of guidance to "stay away from X tech because the community is toxic and non-serious"
A disaffected and loud minority says things like that, and the rest of the world goes right on ignoring them. Zero people in the real world avoided using the Linux kernel because Linus was brusque.