The Perl Foundation is fragmenting over Code of Conduct enforcement
arstechnica.com
arstechnica.com
There's the crux of it. It's open source, and the developers usually doing the most work seem to be the ones that are less socially... adjusted, while the 'community managers' and people broadly more interested in DEI are carrying less.
Seen it time and time again. I don't have answer to the situation in general.
"Code of Conduct" of course only exists when people have an interest in working together - which GNUesque projects seem to be more focused on.
Relatedly I observe opposite behaviour the plan9 related community generally - where everyone basically maintains their own patch trees. If you don't want to deal with other people or your changes aren't getting upstreamed for some political reason - it's not a big deal. You can just mail your patches to some friends and they'll use it.
In some ways Code of Conducts is another way of enforcing 'business' culture onto programming and hacking.
At the same time, these 'hackers' and major contributors should stick to their own medicine. If they're all focused on the code, then stick to the code. There's no reason to say something offensive, if your goal is to get a better quality code out there. If you think someone is a @$#%(@% or a piece of @$#%#@ or whatever cartoonishly evil phrase you want to use, who cares? just focus on the code.
The primary purpose of CoCs is to enforce power relations.
For example, apple plainly committed misconduct when they unilaterally bullied the W3C into not adopting SPIR-V as the shader language for WebGPU. The implicit threat was that they would simply kill WebGPU by not making it available on iOS where they have the browser monopoly (which itself is a violation of good conduct.)
All other browser vendors, game engine vendors, and users supported the adoption of SPIR-V but in the end Apple got their way and WGSL was created as a bespoke shader language for WebGPU. (Naturally afterwards many post-hoc justifications were invented for this move.)
Users and game engine vendors were quite upset at this because they now have to learn and implement a new shader language. When they expressed their discontent, those complaints were quickly and forcefully shut down with reference to the CoC (which apparently does not apply to Apple.)
When the Linux OOM killer decides to eliminate memory hog processes, it prints to the kernel logs a message of the form "Kill process [...] or sacrifice child".
Currently there are panic memes spread by idiots on Facebook about their crashed smart TVs satanically demanding the sacrifice of their children (like the ATM in American Psycho demanding a dead cat).
The language that you use to describe things in computing matters because you don't know who might be reading, how they may interpret it, or how important that person is to the future health of your project. The religious right is still tremendously powerful in the USA, open source still smells like communism to them, and open source that demands the blood sacrifice of their children may just mobilize them to attempt the complete destruction or discrediting of Linux itself and any manufacturer that uses it.
The "parent", "child", and "kill" terminology in terms of Unix processes has been around since forever ago. And they still cause alarm! There's a strong case to be made here for changing the terminology. Think, then, how much more alarming more loaded terms like "master" and "slave" must be. That is the thing that angered Spek to the point of using the N-slur -- which is itself intolerable in a professional context.
We all must work together in order to make our software more accessible to all. That means that hackers working closely with the DEI people -- and listening to their concerns -- needs to be table stakes for any open source project if you want broad adoption.
No, I don't need to spend time and money rewriting or changing my code because some crappy TV company decided to freely use linux or other open source software and they didn't like the terms used in it.
How about educating, or else just ignoring, idiots? Seems simpler.
The problem with idiots is that they're idiots. Besides trying to educate them you can't do anything about it. If it's not this thing they don't understand it will be some other thing they don't understand or misunderstand. That's the nature of idiots.
We shouldn't try to create an "idiot-prove" world, as this world would be just idiotic in the end.
Sooner. Like, at least five years ago.
We need something more like a child lock.
> There's the crux of it. It's open source, and the developers usually doing the most work seem to be the ones that are less socially... adjusted, while the 'community managers' and people broadly more interested in DEI are carrying less.
Not quite true.
The crux of my argument with sawyer was that the things he wanted to be removed as cruft were, so far as I understood it, very different from the things the people doing the C level hackery (which he's just good enough to have been a bloody effective cat herder and I'm so terrible at that I couldn't even have done that so just have to trust the people who are good at it).
On top of that there'd already been most of a year of very passionate arguments leading up to that and he, and I, and almost everybody else involved, were risking burnout from them.
Then I misjudged a conversation horribly and pushed a dear friend over the edge into flat out burnout.
I didn't apologise for that because anybody made me to - in fact, the apology he deserved was delayed by months because the TPF CAT forbade me from contacting him even to apologise and told me that would be considered harassment if I did so.
Then I discovered he didn't know that, and my friend had spent the time in between getting steadily more upset because he thought the lack of apology was because I didn't care, and I decided the TPF CAT could go fuck itself with a rake and they could sanction me if they liked for caring more about my friend than their orders, and talked to him and apologised anyway.
We had a good conversation after that, and I only wish I'd done it sooner.
I could be wrong, but it seems now people feel obligated to make a public stance on social media or email lists. Some of it is genuine outrage, some of it is virtue signaling, etc... but the effect is that the perpetrator is backed into a corner in embarrassment at public shaming. It's difficult to come back from that.
I would dread being a young person now and live in terror of saying something stupid about trans issues or do some cultural appropriation I was unconscious of and be "called out" in a public manner.
That's not the way I remember it.
I think the pre-internet nostalgia of a proper “informal mediation process” is about as genuine as believing the US has an actual hardline between church and state.
That said, it’s funny you turn to “trans” issues as some powder keg when the article is about whether a committer should be allowed to push something along the lines that says “ni—-er get a job” to CPAN. I can’t think most professional communities would have caused a stink about that on their mailing list
Did I say "hardline"? No, that's your word. Of course there isn't a hardline. It's an ideal that we try to maintain. It's a stated goal that we can all read in black and white. We have lawyers and constitutional scholars to mediate disputes over this stated goal.
> it’s funny you turn to “trans”... “ni—-er get a job”
I don't know why this is funny. Or are you saying that you are offended I even brought it up?
And she has terminated her relationship with the committer.
The toxic element to be "purged" from the Perl community, has now returned and will make the Perl community more to his image . Which will only get more people to leave the Perl community. But that is no different from the past 15 years that I have had the pleasure of dealing with this toxic element.
I am, however, very glad that the Raku CAT has since corrected your mistake.
I am very well aware that I can be an asshole sometimes, but unlike you I don't defend outright anti-black racism because, to quote Patrick, "I call my black friends that" and "it's just a word".
Your repeated opposition to code of conduct enforcement is unfortunate but now you've resigned I hope we'll be able to move forwards on that matter without your personal dislike of me interfering.
(I know that part of this comment is referring to events that took place years ago; I don't have any firsthand knowledge about those events but want to correct the record about the present)
[0]: https://github.com/Raku/Raku-Steering-Council/blob/main/pape...
I appreciate that she's revised her opinion on this specific case and apologise for not making clearer that she's done so - I only wish she'd be more supportive of code of conduct enforcement in other cases.
Sam was absolutely doing her best, and if her initial decisions with regard to me were perhaps arguably a trifle draconian, I'd note that "a moderator getting out over their skis due to an excess of righteous anger" is absolutely somebody I've been significantly more times than Sam has.
The only person standing behind said committer's racism was lizmat, who forced me to reverse the ban I laid for the original fucked up cpan upload and has been attacking me for thinking he was a problem since.
However, lizmat has now resigned from TPF; I hope she isn't as much of a problem to raku conduct enforcement going forwards as she's been elsewhere, but I have faith in the raku team's ability to manage that issue and don't intend to interfere with their doing so.
He didn't push it to CPAN, his dotfiles repo is not on CPAN. Please do not spread misinformation.
Edit: The CPAN thing is different. It contains a directory whose name is "perl6-nigger", you can find it at http://backpan.perl.org/authors/id/T/TY/TYIL/Perl6/App-Cpan6...
Riedel's complaint to the board points out that this isn't Spek's first racism rodeo. CPAN still offers downloads containing a 2017-era tarball from Spek with a folder named "perl6-n[word]". We downloaded the linked file and can verify that it's still available from CPAN, and it does in fact contain that folder.
There's a simple CoC that neatly solves all these problems, and it's small enough to put on a footer for every page of every document:
"If you're an asshole, I kick you out."
The answer is that you know the leader and you trust them. If you don't, you should leave.
That sounds like a recipe for a cult.
People like to know what is and isn't okay, and leaving it up to someone's feeling, even if you generally trust that person, makes the whole thing very fragile.
Monarchy in one form or another has lasted thousands of years. It does have the succession problem (or the Bus Factor), but requiring the monarch to designate a successor is the usual solution to that.
Here's the essay: https://archive.vcu.edu/english/engweb/transcendentalism/aut...
I was just involved in some local political drama that was, if anything, worse than what happened here. And ignoring the Perl community as a consequence. So I was quite surprised to see a news article about all people I know, all of them friends. I really think the months of isolation are getting to all of us psychologically. That and the fact that the color of the sunlight has been subtly wrong all summer long from the smoke. We are children of a dying planet.
I don’t know how to fix it, but I do think everyone everywhere needs to take some time and calm down if they haven’t already. I love involved in this, and hope that we can all have a beer together once again.
I think one reasonable way of "checking" statements about trans issues is to think about how similar statements about the obese would play out.
The Perl Foundation is not The Perl Community. Sebastian Riedel and Curtis Poe are still very much involved in the community and development of Perl. Elizabeth Mattijsen and Samantha McVey have not been involved in the Perl community for quite some time. Mattijsen is heavily involved in the Raku development and community, and likewise remains so after this resignation from TPF.
Sawyer X resigned from the Perl Steering Council, not TPF, and as such is the only one here that actually was related to the development of Perl. Regardless, Perl development continues apace under the newly redesigned governance described at https://perldoc.perl.org/perlgov, which has nothing to do with TPF.
When MacOS X was first released I asked a question about that on the MacPerl list and was told to "fuck off" because my question was "off topic".
I ended up requesting a new mailing list for MacOS X be created and it was. For the first year or two it grew very fast and it was a pleasure to be a part of it, but then some of the Perl "elite" started joining it. Randal Schwartz was among those who did and he was a complete jerk. He started trashing members on that list who'd been there helping others, including me. There were even developers who worked for Apple there, and Randal trashed them too.
I finally told Randal to "fuck off" and to my surprise there were a lot of members who stood up with me, and off Randal did fuck. It wasn't too long afterwards Randal screwed himself but good with his own arrogance.
It was around 2006 that I started using prototype.js to handle a lot of the things I was doing with Perl and not long after I unsubscribed from all the Perl mailing list I'd been on.
Nowadays all those Perl mailing lists are pretty much silent and I'm pretty sure the MacOS X perl list is dead.
I still admire Larry Wall though. He stayed far above that bullshit and remained friendly and humble when Perl was at it's pinnacle back around 2000-2005 and .cgi scripts were pretty much the backbone of web work.
Excellent article. Why does there have to be a singular "Perl community" and Foundation for it? Usually we get schisms and "Thrice Reformed" vs "New Revelation" groups sprout over differences of opinion that rise to intolerable levels.
Very similar to the "why shouldn't a project be forked" question; i get the feeling that forks are frowned on more severely now than in the past; but I can't see why. Recall Xemacs and ecgs, among others? Forks work or they die without bothering anyone.
> and the "yes" argument is explicitly racist and empty-headed.
If you are dismissing any possible "yes" argument before it is even made as "racist and empty-headed" I would say that you have withdrawn yourself from public discourse.
Seriously, look in a mirror sometime; read what you've posted as if it were posted by somebody else. Are you excited for the argument that you've prepared for yourself?
The fact that this may or not be true, doesn't matter to him. Nor that he damages the Raku community with this (as well as the Perl community).
However, that this even applies to TPF / the Perl community, is mostly mr. Riedel work, using this trolling for his own agenda (which is reforming The Perl Foundation). The matter has been dealt with within the Raku community already. And Mr. Riedel should know that.
Now, there is that final 10%. Whether the committer knows this or not, he does show some other tendencies that could be considered racist. But in a country like the Netherlands, that is still trying to deal with "Zwarte Piet" (https://en.wikipedia.org/wiki/Zwarte_Piet) this is sadly not as uncommon as it should be.
Do you have some specific examples and had the Perl community informed him about these?
Not because I necessarily thought he was racist himself (I knew basically nothing about the dude back in ... 2018? I think it was), but because his behaviour was bloody stupid and going to hurt people and cause problems down the line.
He was, unfortunately, unable to understand that my goal was to avoid the "bloody stupid" continuing rather than to accuse him of any particular bigotry, so he got lizmat to force me to revoke the sanctions.
He's now in a far worse position as a result, and the "problems down the line" have, indeed, happened.
Frankly, as a relatively young developer, full of testosterone and even shorter on wisdom than I am, he deserved better friends.
How were you forced exactly?
Mr. Trout was made aware of the upload by Mr. Spek. An upload that had been superseded within half an hour by another one without the offending (empty) directory. With the original upload marked for removal, but since PAUSE doesn't let you remove modules immediately, it stayed up for a week. And *ANY* module ever uploaded to CPAN, stays on BackPAN for historical purposes. Which I think is a good thing, so we can see what we're talking about.
Mr. Trout, in his self-assumed role of moderator of the then #perl6 channels, kickbanned Mr. Spek without further ado from the #perl6 channels. Even though Mr. Spek was a regular contributor to the Perl 6 project and user of these channels for asking questions and giving answers. In an ironic way very similar to Mr. Trout. Who, by his own admission, has made *MANY* mistakes, yet until then, was *NEVER* banned from anything in the Perl community.
I considered the action of Mr. Spek more as trolling / stupidity / bad habit / young foolishness at the time. And the first offence to my knowledge, by no means grounds for a permaban on the channels as instated by Mr. Trout. I therefore, on behalf of the core team and after conversations with our de-facto BDFL Jonathan Worthington, asked Mr. Trout to raise that ban. Which in the end he reluctantly did. And since then has been interpreted by Mr. Trout as condoning racism.
Sadly, over time, it has become clear to many that Mr. Spek is either too stupid to realize what he's doing with his trolling (saying he is a proponent of free speech and using that), or indeed a racist at some level, or both. The commit in his personal repo (which he makes public as good examples of using dot files) was about using the "master" branch as the default for new git repos, rather than the current git default of "main". With as single comment "Get a job, N*". This was no longer a use of the word as just a word: this was adding context and meaning. For which the Raku CAT team has banned him from public Raku places and revoked all of his commit bits until Jan 1, 2022.
Meanwhile, it looks like Mr. Spek will not change.
As to Mr. Trout: I will no longer interact with Mr. Trout in any way, shape or form. With pain in my heart, I must say that I've given up on the Perl community. And I say this is a recipient of a White Camel award, awarded for services to the Perl *community*.
I'm not the only person who left the Perl community because of Mr. Trout's antics. I guess in the end, only people willing to put up with Mr. Trout, will be left in the Perl community. I sincerely hope they will be happy, as life is too short to be unhappy without good reason.
Could you elaborate what these were? I am curious. (I am talking about things other than the commit and the cpan thing)
As far as the Raku CAT is concerned, he is now serving his "sentence" of being banned, until Jan 1st, 2022. We will see what happens after that. I sincerely hope that the Raku CAT will not have to address any reports about him after that date.
I banned him for responding to it being pointed out that it was an incredibly stupid upload by explaining it wasn't a problem because the slur in question was "just a word" and anyway, he called his black friends that and they didn't mind.
> Mr. Trout, in his self-assumed role of moderator of the then #perl6 channels
I took responsibility for the perl6 IRC infrastructure with Larry's active approval.
Jonathan Worthington was in no way involved in the unbanning process, only Liz being angry at me in private messages on behalf of her 'friend' until I gave up and lifted the ban.
> permaban on the channels as instated by Mr. Trout
The ban was intended to last while he thought about his mistake and considered whether that was the person he wanted to be going forwards.
I still think he would be in a better place today had he had that opportunity then.
> I'm not the only person who left the Perl community because of Mr. Trout's antics.
On a long enough timeline, anybody with an opinion and a spine will get into personality clashes that result in people leaving.
I know of several that were because of me, some of which I regret. I know of several that were because of you, some of which I hope you regret.
So it goes.
TL;DR; - my modules had been up for adoption for *YEARS* already, but nobody was willing to take over maintenance - you can *NOT* just remove modules from CPAN. You can only schedule them for deletion after a week - during that period, no alarm bells started ringing that modules that had downstream dependencies, were being completely removed - I knew that my modules would always be available on BackPAN, and that it would be just a matter of someone re-uploading them and adjusting the master list. Annoying at best, because it would not be breaking anything that was already in production. - And I still had my own copies, so if anybody would have asked to please upload them again, I would have. Nobody did.
What this showed, is that CPAN has a definite vulnerability in this area: - someone (hopefully the owner) can schedule all modules for removal - if these modules have downstream dependencies, then that's too bad: they will stop working
Now, almost two years later, this vulnerability still exists.
As to the reason of me removing my Perl modules: it signalled my leaving the Perl community for good. Before that, I had still hope that some day, I actually would enjoy being part of the Perl community again.
Should I not have removed my modules like that? Well, I don't know how I could further indicate that I did not have any desire to keep maintaining them. The ADOPT-ME status was clearly not enough.
I'm not sure what you tried with ADOPTME, but transferring first-come permissions to ADOPTME would remove your ability to maintain them and allow anyone to adopt the permissions. Permissions for future uploads are a rather separate concern from the currently indexed latest version of the module that dependency chains rely on. The index does not include distributions on BackPAN, so such distributions are only available via manual intervention.
Poor kid deserved better friends.
Mattijsen's idea of friendship helped get him ejected from the raku core team (as to whether his being given an opportunity to reconsider his life choices and maybe start acting like an adult would've helped avoid him ending up suddenly single is a rather more open question).
If thinking that was a bad thing as well as the racist comments makes me toxic in your eyes, then so be it.
There, FTFY.
The Perl Foundation is a non-profit that handles grants and conferences (and I'm sad to see their work on the latter has regressed since I helped get a code of conduct in place for american perl conferences close to a decade ago).
It has no authority over core development - that's the Perl Steering Council.
It has no authority over conduct in core development spaces - that's the perl5-porters code of conduct which has its own moderation team attached (of which I am not a member but respect the people who are as effective moderators and also nicer people than I am).
It has no authority over irc.perl.org in general or libera.chat #perl - those are handled by the irc.perl.org code of conduct with a moderation team that I helped recruit after taking the lead on the community consultation part of the code of conduct adoption process, but is far from only me and intentionally includes more than one person who will disagree with me quite vocally on the (sadly more common that I would like) occasions when I fuck up.
What we need to do going forwards is to get conduct enforcement at the TPF/CAT level up to the standard the rest of the community has been reaching for years.
This was not stated by McVey, but by me:
> "I'm fresh out of ideas with regards to handling the toxicity in the Perl community," Mattijsen told The Register, "I'm doing my best to make sure that the Raku community will not allow individuals to create a toxic atmosphere without being challenged for what is now decades."
More to the point, the fact that basically every online community eventually develops these behaviours makes it look like everyone on the internet is an asshole.
> In March of this year, Patrick Spek—a former member of the Perl 6 (Raku) steering council—committed to .config/git/config with a commit message saying "Get a job" followed by the n-word. Sebastian Riedel lodged a complaint with TPF board, which he says was railroaded by a board member—therefore, Riedel quit.
Is a commit message a programming issue? If that kind of conduct is accepted, I don't think I'd want to be associated either.
> committed to .config/git/config
In his personal dotfiles repo.
IMHO, this is the source of quite a bit of division of late.
> The entire point of using words is that they are well-known abstractions used when communicating to other people
I will disagree with this, and so will every diary-owner and note-taker. In addition, even if this was true there are various languages (like tut) which are intended to be used within a specific community, just because something is written it does not mean that it needs to be understood by everyone, although it is desirable in public projects. Not to mention that this statement assumes that he knew that they would interpret it in a way which he did not intent.
Anyway, he made a new commit earlier today without it: https://git.tyil.nl/dotfiles/commit/?id=891b7a7a31941dfe253c...
Yes, words have meanings. Those meanings may shift, language evolves after all. This is not what happens currently though when words are forcibly redefined to fit a certain narrative. This is a language revolution by means of revolutionary language.
The so-called 'n-word' and the absolute ban on its use by 'whites' where the version ending with 'a' seems to be used in every other line of most rap lyrics is in this comparable to the Dutch word for a homosexual man ('flikker') which was banned from common use but taken up as a nom-de-guerre by homosexual men. This habit of seeding the field of language with linguistic landmines has done nothing to ease tensions, the opposite is true. This is a relatively recent development as can be seen by watching earlier commercials: bout a week ago I happened to play a video of the first two hours of MTV [1] in the corner of my screen when my ears and eyes were drawn to a commercial about 'The United Negro College Fund' [2] using its full name. The fund still exists but I'm hard-pressed seeing them use their full name in any advertising campaign now due to the stigma attached to the word 'negro', according to an entry in the talk page [3] they are in the process of changing their name because of this. This commercial was 40 years old.
[1] https://www.youtube.com/watch?v=PJtiPRDIqtI
That doesn't look accurate to me? The article has bullet points listing five people who left, all of whom are programmers.
The other perl coder resignations are people resigning from the non-profit, not stepping away from the code.
Sam has actually been a pretty damn effective community cat herder (and I say this as somebody for whom she was the lead inquisatrix, but I have enough moderation experience myself to not take that part personally) and I hope once she decompresses she'll come back.
The only 'programming member of the community' we'll likely lose permanently is lizmat, and she'd already become a raku-only programming member of the community and deleted her entire set of perl5 cpan uploads without warning (breaking a number of users' installations) in a fit of pique because she's convinced the perl5 community is trying to persecute her, so between that and her active opposition to enforcement of basic standards of civility she won't realistically be missed.
Much though I love perl, terrible perl really is truly terrible to work with.
On behalf of those of us who're aware that if perl was a tool it would have black and yellow stripes down the side ... we've all also encountered the sort of perl that would make you think such a thing.
With great power comes great capacity to screw all of the pooches.
1. Absence of an informal Code of Conduct
2. Presence of a formal Code of Conduct.
Also, I feel like characterising this as a CoC enforcement issue is a bit misleading - after reading the article this isn't really about the same thing as the backlash to CoCs (like not using preferred pronouns or whatever). This is just straight up abuse and racism. You don't need a CoC for that.
In the Perl community there is only a notional pumpkin but similar rules apply. Only the person who holds the "patch pumpkin" could merge submitted patches into the mainline Perl tree. The duty of pumpkin holder, or "pumpking", rotates periodically, I guess to prevent the perceived issues with a single BDFL approving changes for all time.