Feature Request: prevent racial slurs from being posted

June 26, 2020 1:40 PM

I have noticed that certain egregious racial slurs are being deleted by mods with a note. I am requesting a feature to simply prevent users from posting these slurs in the first place.

This feature would prevent users from being triggered upon seeing these words on Metafilter before a mod can get to removing the comment and also lessen the burden on mods on spotting and removing these words.

I am asking for the simplest and most trivial implementation possible: if a user attempts to post a slur and clicks Post or Preview, they should encounter a warning that they are not allowed to post unless they remove certain words. I suggest this be a generic message, not tailored to any particular slur.

If preventing certain words from being posted has been proposed and shot down in the past, or if Metafilter staff decides not to implement this feature, could Metafilter staff be transparent as to why this decision was made?

posted by xtine to Feature Requests at 1:40 PM (235 comments total)

We've been talking about related ideas the last little while in the team slack, and there's been some discussion among site members in a couple threads recently. I think building in support for mods to be able to more actively monitor specific language usage on the site is a good idea, and we're looking at implementation ideas.

Where we're at currently is looking at an auto-alert process for a manually-maintained shortlist of words, so that any comment that gets posted that matches one of those words will create an automatic alert so the mod on duty will be aware of it quickly instead of depending on user flagging to make us aware.

I prefer this over an automatic deletion approach because (a) with an alert system we can consistently respond quickly and (b) it allows for the flexibility of cases where there is good justification for mentioning a problematic word or the match is a false positive (such as slurs that are substrings of longer unrelated words, URL cruft, etc).

How likely any word will be justifiable is going to vary from word to word—at this point it's vanishingly unlikely that putting the n-word in plain text on the site will make sense, but stuff with complex fixed-phrase referents like "gypsy jazz" may need contextual consideration. But us being actively aware of and able to evaluate those instances proactively would go a very long way to avoiding instances of something sitting around for a while because it wasn't flagged.
posted by cortex (staff) at 1:49 PM on June 26, 2020 [9 favorites]

Roughly how often are comments being deleted purely on word choice? 1/month? 100/month? 1000/day?
posted by Mitheral at 3:00 PM on June 26, 2020 [5 favorites]

It's infrequent enough that I don't have a good sense of an average rate, I think it probably varies a lot from month to month because it's such a small dataset. I'd gut check it at a handful of times a month we end up looking close at a comment specifically for weird or uncomfortable word/phrasing choice. By far most of those are more of a contextual issue (bad phrasing, reiterating a not-great trope, etc) than a given single word being specifically radioactive. So not a very common thing at this point.

But the usefulness of an alert system would be that it doesn't have to be human mods scouring 24/7 for a low-volume issue; we can just know that we're going to be made aware by the system if something needs a look, when it does ever come up, and then we can assess it promptly from there. That feels like a good match for a low-frequency issue that also can involve a more potent sense of harm when it does come up, and in line with how the flagging system works.
posted by cortex (staff) at 3:08 PM on June 26, 2020 [5 favorites]

cortex: Thanks for the explanation of what you're considering.

I hear what you're saying about how flag-immediately-for-mods (instead of blocking the post/comment from being made) makes sense because many of the problematic words, for instance, show up in URLs, or might be substrings of other words. If you want help starting that list, I know Darius Kazemi has a list of words he excludes from use in his bot projects (also see data/words/expletives.json in his corpora project).

But, I'd really like it if I knew the n-word just wouldn't show up at all, even if I happened across a comment or post before the on-duty mod(s) got to responding to the flag. (I think that most of the time on-duty mods look at flags within a few minutes; I don't know what the stats are.)
posted by brainwane at 3:13 PM on June 26, 2020 [8 favorites]

Perhaps a comment that trips that filter could be wrapped in spoiler tags until a mod got to it.
posted by bashing rocks together at 3:18 PM on June 26, 2020 [2 favorites]

Yeah, typical flag reaction time is 1-5 minutes though it can take longer at times if someone's trying to eat lunch or clean up cat puke or there's a very busy thread elsewhere on the site, etc. I take your point about wanting to narrow the window as much as possible; I do think the difference between "a few minutes" and as in the recent case "several days" is way more important than the gap between zero and a few minutes, but I agree that the notion of trying to implement a full-stop thing on it up front is the only thing that would guarantee that the number is zero.

Architecturally that's more complicated than alerting us about a comment, so I don't want to make any kind of prediction about how workable it is. An alert would probably be a pretty light code change; holding or hiding comments would be more ambitious and could have a lot more knock-on effects in terms of where we'd need to change site code to accommodate it correctly.
posted by cortex (staff) at 3:36 PM on June 26, 2020

I also should have said: big appreciation here for xtine for writing this up and proposing it!

Thanks for the further context, cortex. Totally understand about the level of software change involved being much easier if you hook things into the flagging system versus building a new holding/preventing/hiding comments feature. If there's some way to make a certain rare class of "this is extra-urgent" flag that notifies the on-call mod more "loudly," like with a pop-up, audible, or SMS notification, that might be helpful (to help override the concentration in case there is a very busy thread elsewhere that's currently occupying their attention).
posted by brainwane at 3:57 PM on June 26, 2020

I wonder though, I would kind of like to know who is out there posting racist or ethnic slurs. Deleting them seems a rather polite treatment.
posted by chavenet at 4:05 PM on June 26, 2020 [7 favorites]

What I am proposing is not a deletion, holding, or hiding of comments. Repeating what I wrote on the post, I am proposing that on the technical side that when a user hits "Preview" or "Post," the system will check (via a regex) if the slur is found (in this case, the n-word). If it is, then it will simply return a warning that their comment will need to be edited before posting. If they truly feel there is a need to post something that potentially triggering and offensive, they can contact a mod.

I feel that prevention will save time and deter this behavior rather than correction.
posted by xtine at 4:09 PM on June 26, 2020 [43 favorites]

I really like what xtine describes. Having a simple alert to remind a user to give pause and more thought to what they are posting is a useful tool and it could help Metafilter have a community that is slightly better at self-moderation, at least in instances of slurs. This would by no means reduce the need for mods, but serve as a simple self-regulating mechanism to help posts/comments reach a minimum standard that should be expected (i.e. the new norm) at Metafilter. It should be easy to implement, too.
posted by mayurasana at 4:26 PM on June 26, 2020

The details/summary (hidden spoiler sort of thing) evidently doesn't work everywhere anyways. Or at least that's what I was informed of when I proposed using it for other things.
posted by zengargoyle at 4:28 PM on June 26, 2020

My understanding is that xtine's proposal is trivial to implement client-side.

If you are deeply concerned about edge cases where a discussion would be materially limited by the inability of users to spell out the n-word, it should also be trivial to provide the mods a separate unchecked interface which you could use when you believe a comment really needs the word to be fully spelled out.
posted by meaty shoe puppet at 4:29 PM on June 26, 2020 [8 favorites]

Also remember, Metafilter isn't built on Ruby/Python/PHP/Perl/etc. It's not modern. It's built on some ancient sort of stuff from the 2000's and what we might now think of being a trivial little patch probably isn't. You need like a COBOL programmer or something. I do forget exactly what Metafilter is built on but remember that it's not something I'd even attempt to learn, much less want to work on.
posted by zengargoyle at 4:37 PM on June 26, 2020 [3 favorites]

It's ColdFusion on the server. That's why I pointed out the triviality of a client-side implementation.
posted by meaty shoe puppet at 4:45 PM on June 26, 2020 [2 favorites]

the system will check (via a regex) if the slur is found (in this case, the n-word). If it is, then it will simply return a warning that their comment will need to be edited before posting

Just include word-start and word-end markers in the regex to avoid the Scunthorpe problem.
posted by paper chromatographologist at 4:58 PM on June 26, 2020 [7 favorites]

If we are talking about back end difficulties, then we could still address this with literally dropping in a JavaScript script that will completely handle this on the front end. Sure, a user could technically bypass this if they turn JS off, but to go through all those steps just to post a slur would indicate more potential problematic behavior from that type of user.
posted by xtine at 5:06 PM on June 26, 2020 [7 favorites]

Giving the end user the code and list of words so you can do it client side is bad on the security side of things. That's how people find ways around the whole procedure. Plus it's slightly ironic that to keep from seeing bad words you send them a big list of bad words. Not being contrarian or anything, that's just my university network administrator / security / systems nerd hat saying that idea would never make it past the first meeting for being an unacceptable solution.
posted by zengargoyle at 5:21 PM on June 26, 2020

My technical solution would go something like this.

There's a flag somewhere for Deleted posts/comments (PC). If possible expand that to Deleted/Hidden, or add another Hidden flag.

Change the PC gathering code from 'PC.not_deleted' to something like 'PC.not_deleted or (PC.hidden and PC.author == User)'.

Use the bad word finder to trip the Hidden flag and notify mods. Only the posting user will see that content until the mods get around to investigating.

This would totally fit in Metafilter's long-standing general philosophy of mostly just deleting things without any notice/commentary/etc. just make it go POOF! But wouldn't tip off the bad actor by giving them a warning that their post was unacceptable or that it just insta-vanished on posting. It would look like mods doing mod stuff. Nobody would be any wiser except for a while the bad actor would be noticing that nobody is even reacting to their PC and then it vanished.
posted by zengargoyle at 5:46 PM on June 26, 2020 [2 favorites]

Why are we now arguing to over complicate a feature just to make a racist’s life easier? This certainly would go against the anti-racism principles that Metafilter is trying to work towards.
posted by xtine at 5:54 PM on June 26, 2020 [23 favorites]

I like the mod side pattern matching auto flagging. The chance anyone will see proscribed words in the few minutes they'd be live during events that happen so infrequently seems like a good trade off to completely avoid the Scunthorpe problem.

That's how people find ways around the whole procedure.

Keep in mind we aren't a bank or Facebook. People who are actively griefing in this manner get kicked and it costs them $5 come back. And the pay off is forcing people to read a word for a window of 10 minutes or less.
posted by Mitheral at 5:57 PM on June 26, 2020 [5 favorites]

I would kind of like to know who is out there posting racist or ethnic slurs.

The very few times that I've seen this recently, it's been that someone has either been using a quote awkwardly, accidentally (a word in their culture has a much worse meaning in American English), trying to talk about a word not use it against someone, or, rarely, for shock value. This doesn't make it ok! It would be pretty easy to implement a "hey wait a sec" filter so that people literally couldn't post if they were using words we don't want to see on this site. I play games on a goofy game site which has an implementation of this. That said, we aren't YouTube and the mods aren't incredibly busy. I do think it's a great question for the BIPOC advisory group that travelingthyme is working on because I have my own feelings about what words would go on this list, but I think it would be useful hearing from many other members who have to deal with microaggressions at a higher rate, about what would be better for them (not just BIPOC of course but that's what's been being discussed lately).
posted by jessamyn (retired) at 6:03 PM on June 26, 2020 [7 favorites]

Okay, finally, this is what needed to happen: someone with definitive information has answered positively about the feasibility of this feature request. jessamyn and xtine are right: this an anti-racism update that should be cross-posted with the BIPOC advisory group thread.

I am frustrated and angry and disappointed at all the bad faith nitpicking and attempted derails on this MeTa. MeFi staff are more than capable of answering "cold fusion" without requiring "defenders". All the needless speculation on the possible architectures, reminders that MeFi isn't FB, etc., is a ham-fisted ploy to shut down this discussion.

Now that we know it's possible and relatively straightforward, could implementation be given high priority? A new filter that tests against a one-word list is a quick and definitive first step commitment to making and enforcing change.

yes! quoting because this is so important:

I feel that prevention will save time and deter this behavior rather than correction.
posted by lemon_icing at 6:59 PM on June 26, 2020 [12 favorites]

My first thought is: why not?

My second thought is: who decides which words are on the list? Some are obvious. But I guarantee there's a word out there that someone would want on the list and someone else would be very upset were it to be there. For example: my city's mayor's office has just decided to eliminate the word 'chief' from all job descriptions because they consider it a slur. Reactions are mixed.

Ultimately, this is a proposed technical solution to a social problem; I'm wary of those as a matter of course. I do, however, agree that it would be nice to eliminate even that short window where something truly awful appears on Metafilter. What about the following: What if instead of being unable to post, the user is instead confronted with a big modal alert that says

YOU ARE PROBABLY ABOUT TO BREAK THE RULES
THINK TWICE BEFORE HITTING POST COMMENT

posted by dbx at 7:07 PM on June 26, 2020 [5 favorites]

I like the mod side pattern matching auto flagging.

I totally agree. And the bit about just POOFING that $5 account from orbit.

cortex had it from the start with 'we can probably do this without breaking things'.

I just think actually telling/warning posters before posting is a bad idea. It provides the opportunity to keep trying until they can get past the measures in place. Keeping the measures server side and mod monitored and acted upon quickly is a better option at the cost of a bad post/comment being up for a short while.

My overly technical blah was just shadow ban the fuckers until the mod comes around. I haven't seen any derailing here, just tossing about of ideas about how it could be done well.

Leave it in the lap of the mods (the lap of the mods).
posted by zengargoyle at 7:20 PM on June 26, 2020 [1 favorite]

Instead of slicing and dicing every possible scenario and pre-supposing the difficulty of arriving at a list of slurs, why not start from a simple and obvious premise: there is one specific word-- the n-word-- that almost all good-faith participants would agree need never be written out in full by non-Black people. Why not just see how Black members feel about hard-preventing usage of the word, and then just do it if there are no major objections?
posted by dusty potato at 7:26 PM on June 26, 2020 [30 favorites]

The pithy answer to your question, dusty potato, is that the post is not about one specific word - the n-word. It's about "these words", "a slur", "certain words".

I agree, we can probably get to universal agreement on that one. But please note that the original post referred to a hypothetical list of slurs, and so it's pretty reasonable to have that context in mind when thinking about a response.
posted by dbx at 7:32 PM on June 26, 2020 [1 favorite]

I doubt that the OP's intent was to insist on arriving at community-wide consensus on a comprehensive list of slurs before any progress on this issue can be made at all, but I won't speak for them.
posted by dusty potato at 7:36 PM on June 26, 2020 [16 favorites]

Fine, but the One True List of Unambiguous Slurs doesn't have to be completely hashed out to try this solution on the n-word. It might even be good to just start with one word and see how the new code works.

(on preview, yeah, what dusty potato said)
posted by Flannery Culp at 7:39 PM on June 26, 2020 [7 favorites]

I doubt that the OP's intent was to insist on arriving at community-wide consensus on a comprehensive list of slurs before any progress on this issue can be made at all, but I won't speak for them.

This is correct, I would be in favor with starting with preventing full use of the n-word first (the first quick and actionable item), and then see what other slurs or objectionable words/phases we can prevent as well if we come to agreement with those. And highly recommending that if a poc committee would be in place they would oversee this.
posted by xtine at 7:47 PM on June 26, 2020 [27 favorites]

Great idea.
posted by sallybrown at 7:59 PM on June 26, 2020

cortex: stuff with complex fixed-phrase referents

cortex, it would have been so easy for you to not actually type out the slur in your example and it's really frustrating that you didn't do it.
posted by capricorn at 8:57 PM on June 26, 2020 [16 favorites]

I am in complete agreement that there is zero need for anyone to use the n-word or other slurs in posting here, that this suggestion is excellent and should be implemented immediately, and that we don't need to be thinking about edge cases or the sad feelings of racists being shut down. And we DO want them to know that their racist bullshit is specifically unwelcome here! This is a way of being specifically ANTI-racist, which should be the goal here. Let's for once just get in and do the work and make something happen, instead of endless speculation about how it might not work or might make the racists sad, or talk about how hard it all is. This is not that big an ask here. It's really not a good look to be sitting here trying to pick this one apart.
posted by fairlynearlyready at 9:42 PM on June 26, 2020 [14 favorites]

I think cortex's initial proposal is probably the most workable -- as I read it, building a tool where they're alerted to words and phrases that need an immediate look from mods. Cortex says that flagged things take only a few minutes; this super-flag-thing will presumably be as fast or faster.

The thing about a hard automatic blacklist of unacceptable words/phrases is that either (1) it'll be nonpublic which means the mods get to decide what's deleted and what isn't with no accountability, or (2) it'll be public, resulting in endless heated argument about edge cases.

I do not think the main issues MeFi faces are about shitposting racists coming in and being shitposty racists who make obviously racist shitposts. Maybe I have missed these instances, but if so it's down to the fact that -- they already go away pretty quickly. Which is the point. The bigger and harder problem is building an anti-racist community, and that is not done using code fixes.
posted by tivalasvegas at 10:15 PM on June 26, 2020 [8 favorites]

Or to put it another way: for every use of an obvious slur, how many unthinking or even well-meaning but oppressive remarks or frames of references or assumptions get used on the site? Even if a perfect list of unacceptable words could exist, that wouldn't change the vast majority of the problem.

I feel like people are looking for a silver technological bullet and that's not going to be how things are fixed, either in society at large or on this site.
posted by tivalasvegas at 10:29 PM on June 26, 2020 [9 favorites]

Yknow this "Stop and take another look at your comment please" feature could just get implemented and hopefully none of us would ever find out because it wouldn't even occur to us to want to use those words. Seriously we all know what they are and anyone who pretends otherwise is pretending otherwise.
posted by bleep at 10:57 PM on June 26, 2020 [15 favorites]

Well, if we can't fix everything immediately, we definitely shouldn't try to fix anything. For heaven's sake.

Consulting a BIPOC (or whatever term is ultimately used) committee as to whether this should be a priority (given that there will always be resource constraints and so things must be prioritized or not) seems reasonable. The nitpicking does not. Mefi is a place where the mods make judgment calls specifically to shape a community atmosphere. It isn't necessary to be respectful of shit-stirrers looking to fight over edge cases, it really isn't.
posted by praemunire at 10:58 PM on June 26, 2020 [16 favorites]

bleep: Seriously we all know what they are and anyone who pretends otherwise is pretending otherwise.

We really don't. Some of us are from a very different culture than others, some of us use English as a second or third language; some of us might be talking about a certain word, some of us might be talking about hot club jazz.

I'm not saying that it's not worth trying to get this done, to me it's clear that it is. But it's not quite that simple.

The next part is not specifically meant for bleep:
Walking around an idea and kicking the tyres is the normal process for making changes to a technical solution. I can see that, if you're not used to that, it may feel like nitpicking or picking the idea apart. But it's how things work in many tech environments. It doesn't mean it's not a good idea. It often means people are trying to find the very best way to implement it because they feel it's a good idea.
posted by Too-Ticky at 11:22 PM on June 26, 2020 [20 favorites]

Remember when frimble built a whole new interface as an April fools joke?

It's really hard not to read this thread as Cortex (and some users) telling us exactly how much they care about the racism on this site.
posted by meaty shoe puppet at 3:23 AM on June 27, 2020 [13 favorites]

The "auto-alert and manual-delete" system that cortex sets out seems like a sensible and effective solution.

So, kudos to xtine for raising this issue, and kudos to the mods for having already considered it and having come up with a workable solution. Hopefully we can now move forward and implement it?
posted by matthewr at 4:25 AM on June 27, 2020

I feel like perfect is becoming the enemy of good here.

Block the most egregious and vile slurs immediately. There are no nuances that needs to be considered for those.

Continue working on something more comprehensive at the same time.

These things can happen simultaneously.

Stop dawdling.
posted by Kitchen Witch at 4:33 AM on June 27, 2020 [26 favorites]

What happens with...

Let’s say we start with an initial list of six works (I picked that number at random) that’s approved by the newly planned BIPOC board and the mods (all of whom are other human beings and not an AI system). One day I’m trying to write a post and my text gets flagged because it contains one of those words, which I happen to use in a non-slur manner. (This will already be a pretty rare situation given that most obvious slurs aren’t also everyday words.) The result is I stop for five seconds and think of a synonym or another way to phrase my post. If it’s a pull quote from a source in the post, I pick a different pull quote, use brackets, or another workaround. If that’s somehow impossible (and I can’t think of when that would ever be the case), I can post a MetaTalk seeking community input for getting the word removed from the list.

Both of the examples zengargoyle gave can easily be reworded. To act like this is some 1984 parallel is way over the top. This isn’t the government. Plenty of websites regulate profanity and harmful language.
posted by sallybrown at 4:35 AM on June 27, 2020 [12 favorites]

Mod note: One deleted. Zengargoyle, using offensive sexually explicit language in an argument about how difficult it is to ascertain what is a slur and what is not ... is definitely not helping the discussion. If you'd like to leave all that stuff out, give us your edits (via contact form is best, in case I'm not on duty when you answer) and we can make the changes. Let us know if you need your text.
posted by taz (staff) at 5:13 AM on June 27, 2020 [8 favorites]

Mod note: Another removed. Zengargoyle, are you trying to give proof to your thesis that if people want to be offensive they will just keep trying until something makes it through? Please stop and leave this thread alone now.
posted by taz (staff) at 5:39 AM on June 27, 2020 [3 favorites]

The Darius Kazemi project feels like a good example of why "auto-delete" is a less good solution for Metafilter than "auto-alert and manual-delete". He has made the explicit design choice to delete all content containing "Pakistan" or "homogeneous" because of the substrings those words contain — obviously not the right choices for Mefi. Yes, you could code a more sophisticated solution that handles those cases. But then you run into the next problem, and the next problem. There are good reasons why Metafilter chose to invest in a highly-skilled human mod team rather than the reddit-style automod route.
posted by matthewr at 5:43 AM on June 27, 2020 [6 favorites]

He has made the explicit design choice to delete all content containing "Pakistan" or "homogeneous" because of the substrings those words contain — obviously not the right choices for Mefi.

This is completely different, and not what is being proposed.

If you’re reaching so far beyond the proposal to find a reason to show why it won’t work, please consider what’s pushing you to do that and whether you understand what it’s like to see a slur targeted at you.
posted by sallybrown at 5:53 AM on June 27, 2020 [13 favorites]

Literally posted a comment supporting a solution which deals with the problem and encouraging it to be implemented quickly.

There are various different ways of achieving the goal of a zero-slur Metafilter and accusing people who are discussing those different ways, of somehow being pro-slur is .. not great?
posted by matthewr at 6:00 AM on June 27, 2020 [5 favorites]

Pointing out that you’re reaching beyond what anyone has proposed here to try and find a problem with xtine’s idea is not “accusing” you of “being pro-slur.”
posted by sallybrown at 6:09 AM on June 27, 2020 [8 favorites]

I like the proposed solution of a mod alert much better than an autodelete, since there are occasionally times those slurs can be spelled out without harm to the community (I'm thinking of a recent FB conversation where people were talking about how they reacted to having a homophobic slur shouted at them from cars, and that simply wouldn't have had the same power if people had had to talk around the slur without being able to name it), but absolutely can't be trusted in the hands of the general population.

Although it would probably cause unceasing argument, I'd also prefer if that list of slurs were available somewhere, for discussion and amendment. I know the BIPOC advisory committee was mentioned above, but would just add the reminder that there are plenty of slurs related to gender, sexuality and disability that have nothing to do with race, which should be considered too, and it would be nice if the affected members had some input on that.
posted by mittens at 6:56 AM on June 27, 2020 [5 favorites]

You could have both. You could have a warning when posting, that you had used a slur, please check before posting -- this would warn everyone that slurs are not ok but if I was using one that was ambiguous, perhaps a c word that is also a term for a small crack in armour (I do not believe it is offensive in that usage but please let me know if I am wrong) I could ignore it or click a checkbox saying I am sure I want to use that word. Then there could also be an auto flag system so the mods would immediately be able to check.

I do agree that the list of terms should be made public.
posted by jeather at 6:58 AM on June 27, 2020 [2 favorites]

It's incredibly frustrating to read all this and see all the fucking whataboutery.

Can we just immediately block the n-word from being posted, and then revisit around slurs where there's the slightest fucking degree of nuance?
posted by ambrosen at 7:08 AM on June 27, 2020 [44 favorites]

Explicitly adding my voice to the "can we please just do the simple proposed thing and stop arguing hypothetical situations?" pile. If "let's stop people from using 'the n-word'" is a controversial position on this site in 2020.. I don't even know.
posted by Alterscape at 7:39 AM on June 27, 2020 [37 favorites]

This is one of those times where you can spend ten million years discussing what is the ~~correct implementation~~ that will manage the greatest amount of edge cases nicely and never actually get around to implementing anything or you can just spend literally five minutes to get a minimum viable version that catches a lot while causing a minimal amount of actual problems. Just auto-delete clear-cut slurs and get it done. Christ.
posted by Soi-hah at 7:52 AM on June 27, 2020 [14 favorites]

I wouldn't expect a change to the posting/commenting interface to be implemented literally overnight on a weekend, particularly when the proposed solution is different from what the staff had originally been working on.

That said, I support the OP solution: if you have a choice between improving detection of a problem vs. acting to prevent a problem, you should implement the preventive action most of the time. "Cease dependence on inspection to achieve quality." And the nice thing about having a list is that you can adjust it as needed.
posted by Huffy Puffy at 8:18 AM on June 27, 2020 [7 favorites]

It's been less than 24 hours, much of it on the weekend, and Metafilter doesn't have a full time coder.

The n-word is obvious. But Darius Kazemi's list has many words that are perfectly fine in some contexts (say line 4, 9, 15, 16, 22, 38, 50 and 58) and obviously designed with a different outcome in mind.
posted by Mitheral at 8:25 AM on June 27, 2020 [5 favorites]

Or ya, what Huffy Puffy said. _must_ _preview_ _more_
posted by Mitheral at 8:26 AM on June 27, 2020

Yes, but that still seems like letting the perfect be the enemy of the good. The n-word and the k-slur for Jewish people are easy bans. Some of the less ambiguous ones on that list are too - if people have to come up with an alternate way to describe a gap in armor or a cigarette or an Italian sandwich they can learn to live with it.
posted by Flannery Culp at 8:33 AM on June 27, 2020 [16 favorites]

I support xtine's suggestion. This is a necessary feature that shouldn't have to be requested here in 2020.

It stops the comment from being posted in the first place rather issuing a mod alert that has to be evaluated and acted on while the word sits on a page until the end-users refresh it post-deletion. Use Kazemi's list and put the burden on the user to petition site management why they really need to use a specific word on that list and why no synonyms or phrases can possibly substitute for [word].

Because I'm willing to give up the few ambiguous ones on that list if it means the other words never see the light of day on MetaFilter again.
posted by kimberussell at 8:48 AM on June 27, 2020 [15 favorites]

Immediate implementation of something with the n-word seems like the way to go, as OP asks.

Going forward, Kazemi's list can serve as a starting point, though it is (as others have pointed out) flawed.

I have been in the position in my job of going through historical data being digitized to look for words that were once acceptable use, but are now not, and also words now considered extremely offensive, though once in common use. There are some words left off Kazemi's list that jumped out immediately to me. I won't post them here of course, but can pass them on to the appropriate person.
posted by gudrun at 9:02 AM on June 27, 2020 [3 favorites]

I am getting tired even reading this conversation. Start with the n-word and then work the other stuff out after. That word does not come up in any other conversation or as a part of any other word ever. It is unequivocally bad. We can sort the other things out later.
posted by corb at 9:47 AM on June 27, 2020 [35 favorites]

I'm horrified that we even have to tolerate someone coming into a thread like this and deliberately making highly offensive comments to prove their point. That isn't coming into the conversation in good faith and to me deserves more than a mod note. That made me sick to my stomach.
posted by fairlynearlyready at 9:50 AM on June 27, 2020 [11 favorites]

We didn't tolerate it. Zengargoyle was censured and their comments deleted.
posted by Mitheral at 10:14 AM on June 27, 2020 [8 favorites]

To be clear, my objection isn’t “why isn’t it implemented yet” but rather “why on earth do we need to argue about all the edge cases before doing the simple feature?” I know features take time and effort and it’s unrealistic to expect overnight implementation. That said, why is the feature worth such controversy and digging in over edge cases?

I’m a software developer. If I ever told my boss “I will not make any effort to implement this feature unless we can hash out every possible edge case in advance and solve it,” he’d tell me to get lost, especially if the main use-case delivered core business value.

(Note: I know some real-time systems in safety-critical applications are different, but nobody’s dying if the post filter somehow makes a mistake and someone has to send mod mail. Also it is non-Frimble people who are pushing back, but, I think the metaphor stands. Sometimes all you need is the MVP.)
posted by Alterscape at 11:01 AM on June 27, 2020 [16 favorites]

(I wish people would clarify what they mean by edge cases.)
posted by mittens at 11:08 AM on June 27, 2020

I would think there are two possible meanings for edge cases (though I can't speak for others). The first would be words that are offensive in one context, but not in another (like the garden implement/playing card suit that is no. 56 on the list that's been linked to - there is no other word that means that if you're talking about gardening and, say, have written an Ask looking for one that is high quality or if you're discussing a card game).
The other might be words that many people use casually, but that you could argue are offensive. For an example of that, I'd go with no. 31 on the list, though this is another one where context matters. It would be very offensive to use that word to describe someone who is mentally challenged. But it really isn't used that way anymore, so maybe it's OK in its newer meaning.
(I'm not bringing this up to discuss the offensiveness of these words - just as words that could arguably be called edge cases.)
posted by FencingGal at 11:54 AM on June 27, 2020 [5 favorites]

This is why actual user experience design never takes place and should never take place out in the public forum with the users. Gathering feedback and input, yes. Doing the actual work of "kicking the tires" means kicking the actual people who we're trying to stop from being kicked all the freaking god damned time.
I do not understand why we have to wade through this watery garbage room together every time we think of something that might help instead of help just arriving.
posted by bleep at 12:17 PM on June 27, 2020 [20 favorites]

Sometimes all you need is the MVP.

For people who aren't familiar with this bit of software jargon: Alterscape is referring to a "Minimum Viable Product", which is a starter version of an application or a feature that has the absolute minimum useful functionality, and that is good enough for users to start using for real. The idea is that creators start by making an MVP, and have users interact with it, because in real-life interaction, users and administrators will discover what further refinements are necessary.
posted by brainwane at 12:20 PM on June 27, 2020 [15 favorites]

And let's say there are people who don't know that the n word is the n word. Wouldn't it be a kindness to them to let them know that word is unacceptable in private before they unknowingly insult everyone else by making us look at it? If you had good intentions wouldn't you want to know you were about to make a mortifying mistake?
posted by bleep at 12:23 PM on June 27, 2020 [10 favorites]

I don't even like having to refer to it obliquely.
posted by bleep at 12:24 PM on June 27, 2020 [2 favorites]

bleep:
Walking around an idea and kicking the tyres is the normal process for making changes to a technical solution. I can see that, if you're not used to that, it may feel like nitpicking or picking the idea apart. But it's how things work in many tech environments. It doesn't mean it's not a good idea. It often means people are trying to find the very best way to implement it because they feel it's a good idea.
Also, thanks for assuming how little I'm familiar with, but I am a ux designer at a tech company, so I do know what a tech environment is, and this is not one. This is a community where people want to be able to communicate with each other without being referred to as sub-humans, and that's not too much for anyone to expect.
posted by bleep at 12:37 PM on June 27, 2020 [24 favorites]

Seriously this whole thread is disgusting and we should all be ashamed of ourselves that this is the environment we have allowed to grow. Where people think the right to be as hurtful as they want is more important than any one else's right to feel valued and safe. It's truly sickening.
posted by bleep at 12:47 PM on June 27, 2020 [13 favorites]

To add to my previous comment, I agree this change should be implemented now, and in particular for the n-word asap.

Since a lot of the users here are in the U.S., they may not be aware about this one: the first of the two words sometimes used as the name for this type of lime (the word that starts with "k".) This k-word, in South Africa, is very offensive, on the level with the n-word, but is not on Kazemi's list (it does get referred to as the k-word, in the same way we in the U.S. refer to the n-word; it is on that level of offensive). It should be included, even though that will require changing terminology when talking about that lime/cooking; there are alternate names for that lime and its leaves that people can use instead.

"Use of the word has been actionable in South African courts since at least 1976 under the offense of crimen injuria: "the unlawful, intentional and serious violation of the dignity of another"".

This is one of the two words that jumped out to me as being left off of Kazemi's list, and why I said it has some flaws.
posted by gudrun at 1:08 PM on June 27, 2020 [26 favorites]

brainwane: Thank you for translating developer-to-english for me. :)

mittens: In this case, what I meant by "edge cases" are all the unlikely-to-occur or maybe likely-but-requires-special-handling things, like "somebody wants to use a term that is racist when used in one context, but can be a valid term to refer to something else." My argument was "let's not let solving all those special cases that aren't as clear-cut as 'we should never permit a post/comment including the n-word' get in the way of solving the one clearly-defined case xtine presented." That's what I meant by minimum viable product -- a thing that refuses to accept posts containing one specific word that's uniformly agreed to be not wanted is the basic thing. Maybe Frimble and Cortex implement that and then we realize that actually we need some other behaviour too, but I have a hard time imagining the other behaviour being so onerous that it isn't worth just doing the first bit.
posted by Alterscape at 1:31 PM on June 27, 2020 [1 favorite]

Also I want to add that cortex's idea of just flagging instances of written out slurs isn't going to cut it. I resent the implication that seeing that word may or may not be acceptable to me depending on the context simply because it's never been aimed at me. I've never had a gun aimed at me either, that doesn't mean I'd be ok with seeing one aimed at someone else. It's simply unacceptable because of the ideas it represents and I don't like the implication that there are some people here who are fine with those ideas. There shouldn't be any.
posted by bleep at 2:03 PM on June 27, 2020 [12 favorites]

bleep: Also, thanks for assuming how little I'm familiar with

I assumed no such thing, but apparently it came across as if I did, and I'm sorry about that.
posted by Too-Ticky at 2:03 PM on June 27, 2020 [1 favorite]

I agree with this and it needs to include gender slurs, the most obvious being the c and b words.
posted by kinoeye at 2:15 PM on June 27, 2020 [2 favorites]

For real, if I need to describe a defect in some armor and the word I would most quickly refer to gets rejected by the filter, I can think of other words to use.

A systems approach that works to prevent errors is much better than working to fix them once they're made. No matter how swiftly we think we can do it. "Only" having one or two or ten community members see something offensive before a mod responds to it is not a good standard. We should be setting the bar higher and understanding that sometimes something might happen, and we'll work to prevent those things too.

The community could have done this over a decade ago to get rid of a lot of the boyzone and racism problems we had then. If I had known about regex then, I'd have suggested it.
posted by bilabial at 2:17 PM on June 27, 2020 [11 favorites]

This may be a "me not being as aware as I should be" thing. If the consensus is "those words that can refer to two or more things, and one of them is a slur" are to be avoided, then, that makes the feature even simpler. Still supporting it.
posted by Alterscape at 3:02 PM on June 27, 2020

The automated words ban list idea seems like it will not deal well with words that are being reclaimed or are being contested.
posted by Dip Flash at 3:24 PM on June 27, 2020 [1 favorite]

> The automated words ban list idea seems like it will not deal well with words that are being reclaimed or are being contested.

Eh, a list can change as language changes.

I think this is a great idea, and I look forward to seeing it happen but don't expect it to appear on the site overnight.
posted by The corpse in the library at 3:37 PM on June 27, 2020 [3 favorites]

I can vaguely imagine we might run into a situation or two where a fully automated process might turn out to catch something that didn't actually need to be caught, and couldn't be simply rewritten, and the mods would need to discuss how/if to update the list, once notified by the poster of what had happened.

I can also imagine some words being controversial - not gonna lie, I held my breath clicking on the list above to see if it was going to contain queer, because my god, I am so tired of having the "q-slur" discussion with people. We might need to have some nuanced and potentially difficult discussions around some words if there are sincere and reasonable disagreements on reclamation or other usage issues.

But I think those are better worst-case scenarios than the alternative where several people end up having to deal with seeing slurs about their race, orientation, etc. here. Even if it's just the people who happen into a thread in the five minutes before a mod can deal with it. So to whatever extent we're taking a vote here - my vote (as a white person but also a queer and disabled person, who has sometimes had to push back on language in various threads around those issues) is that I'd like to see an automated process to deal with the n-word as soon as it's technically feasible. And then once that's been in production for a little while to shake out any bugs or other issues, I'd like to have a separate discussion about how/if to expand that list.
posted by Stacey at 4:02 PM on June 27, 2020 [11 favorites]

For real, if I need to describe a defect in some armor and the word I would most quickly refer to gets rejected by the filter, I can think of other words to use.

Right, like "oh noe, Metafilter is forcing our hand to increase the prevalence of ways to say things that don't sound identical to racial slurs, this is a terrible loss of linguistic diversity".
posted by dusty potato at 5:19 PM on June 27, 2020 [12 favorites]

What the fuck are these comments? Just fucking make xtine's plan happen, right fucking now.
posted by thegreatfleecircus at 5:39 PM on June 27, 2020 [4 favorites]

I totally agree with that this is a good idea and should be done. I also note that I don't recall ever running into the n-word in posts I've read, and wouldn't use it myself, so it will not change my experience of the site in any way whatsoever. I suspect this is true for most users.
posted by paper chromatographologist at 5:50 PM on June 27, 2020

I feel like everyone here agrees that this first word does not belong on the site, and that all of the rest of the talk is about other words but not including that word. I don't believe anyone disagrees with immediately making sure this first word can no longer be added to the site.

So yes please let's do that, and then talk about other words after.
posted by Glinn at 6:37 PM on June 27, 2020 [8 favorites]

Couple notes: Our standard procedures sometimes have to change when trying to interrupt racist patterns. As a software developer and project manager, I'm aware that workshopping and bikeshedding are totally normal aspects of feature development. I get the instinct to do it with every feature request, especially on a site like MeFi that approaches UI decisions pretty carefully. I'd like other devs to pull themselves out of their standard mindset, though, because that's the sort of thing everyone will occasionally have to do in fighting racism. The way we're used to developing features isn't always best for every feature in every context.

I may be mistaken, but it sounds like a client-side solution that blocks just the n-word is both universally agreed-upon as a minimum first step and should be reasonably quick to implement. There isn't anything gained, as far as I know, by waiting on agreement on a more far-reaching solution.

The other thing to remember is that most of us have no idea when frimble will be available to implement this feature, so we might not get it right away. But it does seem reasonable to expect some sort of statement of commitment soon from a mod that says "yes, we will implement the mvp at our nearest convenience".

It may be the case that debate over the ultimate shape of the feature isn't going to be responsible for delaying the agreed-upon minimum. If that's so, that also would be a good thing for a mod to clarify.
posted by a snickering nuthatch at 7:13 PM on June 27, 2020 [14 favorites]

I am disgusted but not surprised that the initial mod response to this was to place priority on making sure edge cases are protected before implementing such an important feature that is so long overdue it's embarrassing.

A rebuttal to some above "obstacles"/"things to consider":

Cultural differences regarding the meaning of different slurs: Use the filter as a way to educate.

English as non primary language: Use the filter as a way to educate.

Possible technical limitations for implementation including frimble's availability: This is not a difficult feature to implement. It really, really isn't. If frimble isn't available, find someone who is.

Excuses, excuses, excuses. Such BS.

Address the core issue first. Deal with the edge cases on a case by case basis.

God.
posted by Kitchen Witch at 8:01 PM on June 27, 2020 [7 favorites]

As a stopgap if you install the diediedie script you can eliminate comments with any word(s) you choose.
posted by Mitheral at 8:24 PM on June 27, 2020

I don't think there's a mobile equivalent of that script, unfortunately.
posted by Kitchen Witch at 8:38 PM on June 27, 2020

My FB feed has been filled with BIPOC people (among other disadvantaged groups) talking about how they are constantly having their posts and comments or even links in FB messenger blocked/reported, and about getting consequences from FB for things they post. I want decisions to be made cognizant of how systems and rules often impact those who lack privilege, while those who possess it can circumvent them.

Also any post someone changes on their own to remove a slur is a post that doesn't need to get reviewed by mods. A post that gets auto hidden or auto deleted on the other hand should be reviewed by a mod. So a front end solution with a modal or countdown or whatever to post has merits on it's own, as a quick win and a step towards a more comprehensive solution, not as a final fix.

We don't have to do perfectly right from the get go. We just have to do better than things currently are, and then continue improving.

I've also often thought that the comment preview could show one of the Do or Don't items from the community guidelines with a link to the rest of them.

I don't think the site does as good a job as it could of setting cultural and behavioral expectations for the user, and surfacing them in the course of normal interaction with the site, prior to someone commenting or posting, would be beneficial. Tool tips for better behavior, kind of. We need to build tools and systems to enable and compliment moderators, not to try to replace them. But we also can build tools and systems to help problematic users themselves moderate their own behavior and self-educate.

Currently on comment preview I just see "Note: everyone needs a hug." Which maybe deserves a call out on it's own for being ableist and making cultural assumptions, but I digress.

Oh and Kitchen Witch you can use Firefox Mobile with Greasemonkey or Tampermonkey on Android to use userscripts. IOS I don't know about, but maybe worth an Ask post.
posted by gryftir at 10:19 PM on June 27, 2020 [1 favorite]

Kitchen Witch: A rebuttal to some above "obstacles"/"things to consider":
Cultural differences regarding the meaning of different slurs: Use the filter as a way to educate.
English as non primary language: Use the filter as a way to educate.

Please note that while I mentioned cultural differences and people who aren't native speaker of English, I did not say this was an obstacle or even a thing to consider.
I just said that it was not true as was stated that 'we all know what [these words] are' and that it would hopefully never occur to anyone here to use them anyway.
I stand by that, but it's not something that I feel should be seen as a reason not to implement a filter of some kind.

We do not all know which words the American-English speaking users of this site consider to be slurs and we may sometimes try to use them in other meanings. But there is no reason why that should remain possible. Personally I would prefer not to have the possibility to inadvertedly hurt someone badly.
posted by Too-Ticky at 12:46 AM on June 28, 2020 [4 favorites]

Another "vote" (I know we're not really voting but whatever) for implementing xtine's suggestion exactly as described.
posted by equalpants at 12:54 AM on June 28, 2020

Oh my days, here is a piece of javascript that will remove the comment field if you type that word or anything like it, do with it what you will, I love you all. 💜

let mfr = /n....r/;var mfc=document.getElementById("comment");
mfc.addEventListener("input", function(){if(mfr.exec(mfc.value)){mfc.parentNode.remove(mfc)}});

posted by lucidium at 2:17 AM on June 28, 2020 [5 favorites]

Supposing I was on a forum in a language I don't speak natively, I'd be MORTIFIED if I found that I used some slur word because I didn't understand the implications. I'd be extremely grateful if the comment form would save me this embarrassment, and it would be a major bonus that nobody had to flag my comment, or read it and be upset by it, or moderate it or in any other way spend their time contemplating my situation.

Giving me a link so I can go learn the context around the word in question would be even better.
posted by quacks like a duck at 5:12 AM on June 28, 2020 [10 favorites]

Just for the record, the code posted above would prevent posting of the word in question, but would also prevent mentioning the capital of Madagascar (Antanarivo), or the phrase "on borrowed time", among many other words and phrases that happen to have an n followed by an r five characters later.

English is not a regular language (by the computer science definition) and trying to use regular expressions to parse it is one of those problems where it's very easy to write an algorithm that is making value judgments that you don't understand until after you put it into production.

(This is not to express any opinion on the broader question of what is the appropriate intervention here, except that the range of things that simply can't be posted should be much narrower than the range of things that would trigger mod review)
posted by firechicago at 6:11 AM on June 28, 2020 [3 favorites]

I thought the code posted above was just an example and didn't actually use the n-word for obvious reasons, so the problems you foresee wouldn't be an issue firechicago.

Add me to the list of people dismayed at the friction to a pretty straightforward idea and relatively (as discussed above) simple solution. You might see this as a technical issue (and it is a little) but don't let that overshadow the main issue: lots of people here don't want to see slurs, particularly the n-word, on this site. Why is this even a debate?
posted by kendrak at 6:32 AM on June 28, 2020 [14 favorites]

Arabidopsis and bleep and everyone else who made the point, thanks for pointing out the problems with a more delayed (even by a few minutes) mod intervention, and why in this context it would be better to have a more immediate block from posting a comment with the given slur.
posted by mittens at 7:12 AM on June 28, 2020 [3 favorites]

I think it's past time for a mod response.
posted by cooker girl at 7:22 AM on June 28, 2020 [5 favorites]

The system parses comments anyway, to close rogue html tags and such. If it rejects a comment, show a link to the contact page to ask for permission if you want to explain why you should be able to post the thing, and maybe keep the comment in the comment box, so they can edit it out or copy/paste? (I don't have strong opinion about this last part).

The hard-ish part is just blocking comments based on a list of words. The easy part, once you have a list, is adjusting the list of words on an ongoing basis depending on what the community wants. You can improve that part without having to do any programming.

(and for testing/debugging you probably want one or two test words in the list that aren't horrible slurs, so that if the blocker doesn't work right you accidentally post "chumpmobile" instead of something nasty.)
posted by Huffy Puffy at 7:27 AM on June 28, 2020

Chiming in to say I support the idea of a simple block on posting the most egregious slurs that would otherwise be deleted anyway. The site already blocks your comment if you try and fake a comment by-line, i.e.
<small>posted by</small>
will prevent a comment from being posted. So the technical solution literally already exists. If impersonating another user is completely impossible then I see no reason not to ban slurs too. There was no "what about this edge-case??" excuse with the "posted by" block and there shouldn't be one here either.
posted by EndsOfInvention at 7:28 AM on June 28, 2020 [21 favorites]

For reference, if you weren't aware of the "posted by" block, it takes you to a page that says

Please go back and remove the "<small>posted by..." line from your comment. With this line, the comment appears to be spoofed comments from other members so they're not allowed.

posted by EndsOfInvention at 7:30 AM on June 28, 2020 [7 favorites]

This isn't the first time there's been a request for technical changes on MeFi to prevent harm to site users. In 2006 it was pointed out the image tag could be used in a malicious manner. The ability to use it was switched off a few hours later, and some other fairly serious technical changes were made pretty quickly around the same time. I don't recall any of those flaws being exploited for actual malice, just that the potential was there.

And yet here we have many community members suffering actual harm from something that could be mitigated with a fairly simple technical change. No, it is not a technical flaw or security issue this time, but if you don't think the urgency is at least equivalent... I would suggest you reflect on why you think that, and maybe reconsider. Potential issues with implementation need to be weighed against actual and ongoing harm.
posted by FishBike at 7:30 AM on June 28, 2020 [16 favorites]

Morning, checking in on this. Where the mods are actually at: we're in support of both a mod-side alert tool and a pre-comment warning process, and I think the two of those together will work to reduce as much as possible the chance of problems (a) happening in the first place and (b) persisting unnoticed if someone shows bad judgement after all.

We're not getting hung up on perfecting a word list or anything like that. I agree that starting with the obvious basics, the n-word and some other unambiguous slurs, and just getting it working is the way to go and we can iron out the finer points in practice once we've got a satisfying basic version built.

The mod-side tool fits really naturally into our existing code; the pre-comment warning is according to frimble going to require some more work to make work well. We're aiming to do that work, but I want to ask folks to keep in mind that implementation and testing of even conceptually simple things takes some time and care and to stay grounded about the timeline for implementing a feature request. Hacking something out is not the same as making sure it works well for folks, on a variety of browsers and devices, doesn't break screen readers, etc.

I know it's frustrating for stuff to take time, but stuff takes a little time. We've got one part time dev who is full-time parenting a young high needs child. I appreciate folks' patience with getting stuff rolled out; it's something we'll be actively working on this week, and I'm hopeful frimble will have sufficient work time to get some stuff rolled out. We'll update as it comes.

I want folks who are feeling inclined to hash out the nitty gritty or worry about edge cases to be mindful of how that lands in this kind of context. It's one thing to brainstorm about what could go wrong; it's another to let that carry you into a perfect as enemy of good, "well but what if..." place that in response to a request about taking anti-racist action can end up reading hostile to that action. This is partly a problem of just the nature of a MetaTalk pony request thread as an open conversation, where folks who aren't actually doing the work end up arguing with each other about the theoretical shape of that work, and that's something we should talk about restructuring in the future. But it's on everybody to stop and think about whether what you're doing is helping to push a good process forward or just digging in on details for the sake of argument.

From the perspective of a mod team trying to do right on this stuff and make progress, it doesn't help us to have folks "help" by digging in on all the things that are wrong with a justice-minded suggestion. I appreciate folks trying to lay out potential pitfalls with any solution, but that's more frimble's job than yours and when that strays into being defeatist or worrying about censorship etc it gets actively unproductive and can understandably feel to folks like looking for a reason to say "no" when the mod team's already trying to commit to a "yes" about something intended to push against a racist status quo.
posted by cortex (staff) at 7:43 AM on June 28, 2020 [11 favorites]

So, let me get this straight. We are weighing:

1. the slight inconvenience of users having to think of other ways to express a thought (i.e., a small crack in armor)

versus

2. the pain of encountering slurs against one's race, gender, sexuality, etc?

I think the answer is clear. I see no reason we shouldn't implement a no-use policy for a list of words. If, as time goes on, we discover that it needs to be walked back a bit (I can't imagine why, but let's imagine for the sake of argument), we can do that.

Let's stop prioritizing the convenience of the majority over the pain of the minority.
posted by mcduff at 7:46 AM on June 28, 2020 [5 favorites]

Agree with 23skidoo. At this point, just route post and comment submissions containing that particular string to /dev/null. Several people have more eloquently than I could pointed out why this is a superior solution. I can’t imagine I would ever be inconvenienced by this implementation. If, somehow, I was, I would understand.
posted by Alterscape at 8:08 AM on June 28, 2020 [4 favorites]

Please correct me if I've misunderstood, but it sounds like what you're describing is a screen that warns users that they're about to post the n-word if a comment contains the n-word, but then still allows users with bad judgment to post comments containing the n-word if they want to anyway after reading the warning?

What I'm trying to describe is something that says along the lines of "hey, this really problematic word appears [here] in the text of your comment; using that word is generally a violation of the _site guidelines_ and a good way to get banned. If you'r quoting something, think twice about the visceral impact that'll have on readers and look into omitting or rewording your comment. If this feels like a complicated case, please _contact the mods_."

That builds in a warning that we don't have to worry about being perfect about so that it doesn't just eat people's comments without warning because of an edge case, but will appropriately and preemptively warn in all the actual use cases and be clear about what the problem is and why. Bad luck with substrings and URL content and typos can happen and stopping someone to say "hey, uh, take a closer look at this" should handle all of those good faith cases up front.

People intentionally posting with bad judgement and ignoring a clear warning about it won't be stopped by any technical solution; they'll also be promptly banned for their effort, and with a mod-site alert backing this warning up it can be prompt indeed.
posted by cortex (staff) at 8:32 AM on June 28, 2020 [4 favorites]

Absolutely unacceptable to just put a warning, it has to stop people posting.

And hopefully it's no harder to block than the code that stops people ending their comments with

p‍‍o‌st‍ed by someone else at 4:34 PM on June 28 [+] [!]
posted by ambrosen at 8:34 AM on June 28, 2020 [10 favorites]

I truly don’t understand that logic cortex. There is no reason for any user to post the n-word here. It sounds like it would be much quicker and easier to just prevent posts and comments from posting if they contain that word, not some “wait are you sure you want to use the n-word?” system.

Is the refusal to do this based on tech or something else?
posted by sallybrown at 8:35 AM on June 28, 2020 [18 favorites]

Agree that the onus in edge cases should be on the user to reach out and explain they just want to post a URL that happens to include a random slur, or whatever we're envisioning.
posted by the primroses were over at 8:35 AM on June 28, 2020 [5 favorites]

Sorry, we may be talking past each other on a detail of the workflow and I'm sorry for seeming stubborn about it. I was reacting to a couple comments that i read as saying literally nuke the comment, like detect string and zap it, but I'm hearing y'all that it's more a concern of the difference between "hey, think about this" and a flat "hey, rework this if you want the 'post comment' button to work", with a preference that it simply be the latter.

I think we can do that latter; it will mean some more work to figure out how to process actual edge cases when they come up, but per above we can try and get a simple version working and then visit those edge cases as we see it work in practice.
posted by cortex (staff) at 8:40 AM on June 28, 2020 [11 favorites]

flat "hey, rework this if you want the 'post comment' button to work"

This is how I was envisioning it. Something that just stops it from posting and alerts the commenter. “Error: this comment/post contains language that violates site policy” with a link to an explanatory page or something.
posted by sallybrown at 8:51 AM on June 28, 2020 [16 favorites]

Seriously, what the heck is going on, can you please address the actual request? Prevent any user from posting a slur, period. Not delete it later. Prevent them from doing so. Then no mod action will be required.
posted by tiny frying pan at 9:05 AM on June 28, 2020 [7 favorites]

I’m sorry sallybrown, others, and Cortex. I failed here by writing in a way that I can see sounded like I meant “fail to post a comment containing the string with no clarification or warning.” I was trying to support the idea that “allow the post to go through but notify a mod” is not acceptable to me, but did so in an overly snarky way that actually got in the way. Credit to sallybrown and others for bringing the light here.
posted by Alterscape at 9:05 AM on June 28, 2020

Even our discussion about what should happen when a comment can’t post is us getting mired in edge cases. The scenario in which someone well-meaning is trying to post the n-word unknowingly is going to be extremely, extremely rare.
posted by sallybrown at 9:14 AM on June 28, 2020 [14 favorites]

There should be a list somewhere of the words. Sometimes the words have to be discussed as words. I think the list should be quite short.

I suppose MeFi could have a way for flagged content to be held for moderation, but that might be a lot of work.
posted by theora55 at 9:17 AM on June 28, 2020 [2 favorites]

The scenario in which someone well-meaning is trying to post the n-word unknowingly is going to be extremely, extremely rare.

I'm sorry for my own muddling of things above, so to be clear I'll reiterate: hard-stopping someone from posting a comment with the n-word in it, and leaving it to them to contact the mods for the rare edge case false positive, is doable and what we'll aim for. That's my understanding of xtine's specific request, and we are going to do that. The rest of the details as far as larger word lists and the complications that come with them we can sort out from there.
posted by cortex (staff) at 9:19 AM on June 28, 2020 [14 favorites]

For the n-word specifically, WHAT are these supposed edge cases?! "But I really want to say it" is not an edge case! The closest I've seen to an actual edge case is if it was in a URL, but if that were to actually come up in the year of our Lord 2020, just like... get over it and don't post the link? That honestly seems less harmful than not implementing this.

Also to everyone talking about flaws in armor, please knock it the hell off. It was useful the first time to illustrate a true edge case, but now it just seems like y'all are enjoying it.

Because it apparently hasn't been said enough, +1 to implementing this feature in a way that stops users from posting and asks them to rework their comment if the word is detected.
posted by sunset in snow country at 9:19 AM on June 28, 2020 [14 favorites]

Why on god's green fucking earth are we seriously prioritizing the "acceptable" reasons to use xyz slur over the option of not fucking having any slurs on the goddamn site??? Why is using a slur in the context of music so. damn. necessary. that we're okay with letting racism exist? The reality of the situation is that we either buck the fuck up and get rid of the racism, or metafilter wont exist anymore. Having no metafilter is better than having a racist metafilter. The house is burning to the ground and you all are standing around it simpering that if we put out the fire, then what happens in 18 months when someone shows up with marshmallows.
This was the plainest, most simple, reasonable request I've ever seen on metafilter. "Default the site to not allow slurs, and require mods to approve them in context", and it's met with derision.
posted by FirstMateKate at 9:25 AM on June 28, 2020 [16 favorites]

I apologise for fiddling with this as the proverbial burns, and I hope it doesn't come across as anything other than me trying to do what I can to help, but I think I improved the code above:

let mfr = /n[\w-]g{2}[\w-][rh]/;var mfc=document.getElementById("comment");var mfp=document.getElementById("postButton")
mfc.addEventListener("input", function(){if(mfr.exec(mfc.value)){mfp.parentNode.remove(mfp);
document.getElementById("lplabel").textContent = "Please contact the mods if you are unable to post.";}});

posted by lucidium at 9:31 AM on June 28, 2020

hard-stopping someone from posting a comment with the n-word in it, and leaving it to them to contact the mods for the rare edge case false positive
I'd love for all the edge case enthusiasts to enlighten me privately on what this could possibly consist of. You know it's not ok to sing it during karaoke right?
posted by bleep at 9:45 AM on June 28, 2020 [4 favorites]

Black people use the word for Reasons. I don't feel called on to tell them they may not use it at all.

I think it's healthy to shine light on difficult things; I'm not in favor of expanding any lists of prohibited word without lots of discussion.
posted by theora55 at 9:53 AM on June 28, 2020 [6 favorites]

Having no metafilter is better than having a racist metafilter. The house is burning to the ground and you all are standing around it simpering that if we put out the fire, then what happens in 18 months when someone shows up with marshmallows.

Quoted for truth.
posted by bleep at 9:54 AM on June 28, 2020 [11 favorites]

I'm pretty new here, and I dont postmortem. But xtine proposed a solution that is simple, and effective in protecting the community (and if, as some are positing, there are somehow innocent motives in posting these slurs, it protects the poster from inadvertently causing harm!). Cortex's solution still creates lag time between the post and the ability of a mod to act on it, which therefore doesn't work as effectively. It seems like the logical thing to do is implement xtine's request. This isn't something that, once implemented,, can never be modified! The fact that there is THIS much discussion about it is really alarming.
posted by TheFantasticNumberFour at 9:56 AM on June 28, 2020 [1 favorite]

Black people use the word for Reasons. I don't feel called on to tell them they may not use it at all.

I definitely don't think I'm allowed to regulate who's allowed to use that word but I just don't get the feeling that's what's being referred to although I'd love to be wrong.
posted by bleep at 9:57 AM on June 28, 2020 [5 favorites]

TheFantasticNumberFour: The fact that there is THIS much discussion about it is really alarming.

Welcome to Metafilter.
posted by Too-Ticky at 10:04 AM on June 28, 2020 [6 favorites]

I guess I have two thoughts to contribute here.

First, we know that Metafilter exists in the context of white supremacy and that BIPOC people have been saying they feel uncomfortable here, for basically ever. To me that means that whenever the site is faced with a decision that involves race, it needs to bias waaaaaay more in the direction of change, rather than preserving the status quo. Because the status quo isn't working. In this instance IMO that means yes, automate the banning of the n-word and other slurs, rather than farting around worrying about edge cases. "Worrying about edge cases" is how we got here, and it's time to take stronger actions so we get better outcomes.

Second, I really think the people who run the site need to reconsider how they approach these decisions. Metafilter continues to set itself up as the adjudicator on race stuff -- like, a BIPOC person proposes something, and Metafilter listens to everybody here, and picks the solution it thinks strikes the right balance among competing views. IMO that's not the way to do this. If a BIPOC person proposes something, and it seems uncontroversial among other BIPOC members, and it's something that is technically and financially feasible, then I think the site should just do it. If Metafilter wants this place to be welcoming and enjoyable for BIPOC members, and BIPOC members have done the work of telling Metafilter how to make that happen, why on earth wouldn't Metafilter just say yes?

Personally I am totally fine with Metafilter auto-deleting all slurs before they're posted. But I would support it even if I didn't personally like it, because it's what BIPOC people are saying they want.
posted by Susan PG at 10:06 AM on June 28, 2020 [22 favorites]

TheFantasticNumberFour: Cortex's solution still creates lag time between the post and the ability of a mod to act on it

That's not how I read the above:
"hard-stopping someone from posting a comment with the n-word in it, and leaving it to them to contact the mods for the rare edge case false positive, is doable and what we'll aim for. "
That sounds immediate to me, and like it does not require mod action.
posted by Too-Ticky at 10:07 AM on June 28, 2020 [1 favorite]

I think there's a lag between my initial "here's the variant we'd been already looking at" comment from when the post first went up Friday and what I've said this morning, is all. I hope my last couple comments have made the hard-stop position we're aiming for in line with xtine's request clear, but I get that people are coming to the whole thread and not just the last few comments.
posted by cortex (staff) at 10:10 AM on June 28, 2020 [2 favorites]

Why on god's green fucking earth are we seriously prioritizing the "acceptable" reasons to use xyz slur over the option of not fucking having any slurs on the goddamn site???

And

This “edge case” conversation is bullshit. I cannot think of a single slur that I would feel complicated about censoring. Even if you’re generous in your definition.

We obviously should have the N word filtered. I personally think the K word is also a no brainer. As is the H word and the I word but the latter two are more of an aboriginal Canadian slur so probably not as well known here even though they've been directed my way on uncountable occasions.

However this is the Racial Slur Database that got started around the same time as Metafilter and was built on similar lists distributed on USENET.

And this is Wikipedia's racial slur list. Not nearly as comprehensive as the RSDB but maybe mostly includes more mainstream American slurs.

I can't argue any of those words aren't hurtful oppressive slurs when used as a slur. It is obvious they are. Everyone or anyone seriously A-OK with just adding the entries of either or both of those to the banned slur filter? If not that is what I think of when people mention edge cases. Note that they both include includes things like the A word which again is a slur that has been pointed my way on numerous occasions but banning of which is going to put a crimp in a lot of the discussions here. I'm far from an A word fan boi but I like to see discussions about it here.
posted by Mitheral at 10:40 AM on June 28, 2020 [4 favorites]

What arabidopsis said.

cortex, as many racialized members of the site have already stated, all the discussion about "but what about edge cases" is a case of caring more about making sure white people will have leeway to use racial slurs and caring far less about protecting racialized and other marginalized community members from the harm of racial slurs. We don't need a #notall defense for racial slurs. White people are so worried about making sure that there are loopholes for them to be racist and get away with it.

To other white folks here who keep weighing in on how and why to make sure that "edge cases" can be permitted? I don't think this is the discussion for you. I hope you spend some serious time reflecting on why you care so much about exceptions to a rule about banning racial slurs. Concerning the specific case of the n-word and its usage by black identifying members of the community: It is for black-identifying people alone - not white people! not even white 'allies'! - to decide how the n-word is or is not permitted. There is a POC advisory board in development, in addition to the entire racialized community on this site. Trust that they will have these conversations, if they wish to do so.

There are subreddits and other online boards that have no problem taking a hard line on racial slurs.

Sometimes I think Metafilter is stuck in 2001 tech-bro hell because it wants to be stuck there.

Way too many excuses and multi-paragraph explanations about why things can't be done. And then wondering why folks are dissatisfied with the moderation on this site.

In addition to being a safer place online for marginalized peoples, Metafilter could likely attract way more new and active users if it implemented what folks are asking for here.
posted by nightrecordings at 10:47 AM on June 28, 2020 [21 favorites]

This feature request is the sort of interaction that poisons the mods' ability to collaborate with PoC MeFites. Collaboration requires vulnerability. Vulnerability requires trust. This thread is why the mods aren't going to get it.

A PoC MeFite came to the mods with a proposal to nudge the site toward less racism. It was a small, incremental change. It addressed a (obvious content warning on this next link) long-standing, ongoing, and systemic problem on this site. It used a method that is clearly available to us.

It was an opportunity for the mods to listen to people of color.

Instead, we got a immediate derail into the difficulty of implementing a completely different and clearly worse proposal that nobody wanted. After 143 comments we are still struggling to re-rail.

Here's what I learned from this thread: if I come to MeTa with a completely air-tight proposal, I should be prepared for hijacking, nitpicking, and trolling.

So I'd better not come with anything less than a completely air-tight proposal, and lots of reinforcements.

And maybe it makes sense to ask for something less good, if it's more demonstrably feasible.

That's a problem, because there are a lot of potential changes that can't be designed without good-faith cooperation from the mods.

I've been talking about a moderation log. To quote the lobst.rs policy, "While the individual actions of a moderator may cause debate, there should be no question about if an action happened".

There's no way to design that from the outside. If the only way to get a change into the site is to round up a couple dozen people to help ram an air-tight proposal down the mods' throats, we will never have any transparency or accountability for mod actions.

And that's increasingly looking like the case.
posted by meaty shoe puppet at 10:54 AM on June 28, 2020 [32 favorites]

Mitheral, I am going to give you the benefit of the doubt, and try and educate you rather than shutting you down. Your entire argument is 1) a derail, and 2) based on the slippery slope fallacy. Fallacies are ways in which our emotions make us believe an argument is factual, but it is not factual. Creating a script to stop the n-word from being posted is not going to cause the mods to start banning words like Apple. This is not hard to understand.

But, lets say that your fallacious argument did some how, improbably, come true. There is now a list of a few hundred words we can't use here. Here. Not ever, not in public, not on the internet. Just on one, single site. If the options are: 1) allow slurs to exist on the site, harm members, allow racism and B)disallow slurs on the site, make posters re-word comments - why is the 2nd one better? It isn't.
posted by FirstMateKate at 11:13 AM on June 28, 2020 [2 favorites]

My latest comment was a direct response to multiple people asking about edge cases in an attempt to better define terms. But hey, don't bother shutting me down; this BIPOC is taking a break.
posted by Mitheral at 11:22 AM on June 28, 2020 [11 favorites]

Definitely yes to preventing a list of slurs from being posted, even if we start with only one slur on that list.

I too would like to see a moderation log.
posted by maxwelton at 11:36 AM on June 28, 2020 [2 favorites]

White person "educating" a BIPOC on racism and causing them to button in disgust- now that's metafilter baby.
Ban the word. This is not hard. Once the implementation is in place we can have a good old fashion beanplate about what's next. Considering that the mods are now on board for this- but because of frimble's real world obligations it's gonna take a few days to put in place- maybe someone should close this MeTa up before more fighting starts? It's really instructive to see how many mefites want to argue and bicker with each other rather than push for change we can actually have- but considering we're bleeding members as is...
posted by Homo neanderthalensis at 11:43 AM on June 28, 2020 [15 favorites]

Cortex: honestly, your past few responses come across as prevarication and over-explanation to an issue that should be an absolute no-brainer. Perhaps you should know that over-articulation and over-explanation is a common response to issues of race that actually signals that someone disagrees or doesn't truly feel in solidarity.

I'd really appreciate a response that's like:

"Yes. This is important and the n-word should not be used under any circumstances. Thanks for your feedback. We'll get to work on implementing this and our timeline is by [specific date here]."

--

Honestly this discussion makes me deeply disappointed. I will mobilize my anger and frame a countdown clock for my own self:

If this feature isn't implemented by a month from now, July 28th, 2020, then I will stop participating in this community for one year, until July 28th, 2021.

This message is a gift to my future self, to help myself set boundaries and try to be part of healthy communities. The healthy communities I want to be in are the ones in which issues like these are brought up (thank you xtine). The healthy communities I want to be are actively anti-racist, listen to each other, and actively make and work on plans to make the site a better place. If a community that I'm part of doesn't want to do this work, then it's on me to either try to help the community change, or to take a pause from it. This comment is me trying to do the former before I do the latter.

I dearly hope this comment makes me look foolish and silly because the feature is implemented soon, that I am proven wrong.

--

I recognize that my statement above is difficult, like an ultimatum. I'm thinking a lot about Sarah Schulman's 'Conflict is Not Abuse, and especially this part, so am trying to make clear the conditions of repair.

I also think a month timeline is appropriate and generous considering that the client-side javascript has already been posted [1] [2], and that functionality already exists to block certain text from being posted. [3].

And because the healthy communities I want to be are actually also full of gratitude - Thanks indeed to cortex and many of the mods, despite your failings, for trying to hold this community here, and Thanks, to everyone who is listening and reading and talking about this and pointing out this what-about-ism/derailing. I'm angry and grateful at the same time, mind you; deeply angry that I have to mobilize this kind of statement and also optimistic and hopeful that this actually sends a STRONG MESSAGE across about how much these things are goddamn important, balanced with deeply disappointing it is to have elementary/simmple things like this to have SO much friction.
posted by suedehead at 12:02 PM on June 28, 2020 [51 favorites]

Oh and if you're reading this and if this kind of statement feels childish/unfair to you, 'all this fuss over a pony request', then I'd ask you to try to understand why someone would want to pause their relationship to this site over this. Perhaps you're actually underestimating how important of a symptom/indicator this request is; how much this indicates the problematic lack of an anti-racism stance here.
posted by suedehead at 12:06 PM on June 28, 2020 [15 favorites]

I said this in the BIPOC advisory board thread already: cortex, I honestly do think you want to do the right thing. In this case xtine has laid out a simple request and several users have offered up easy ways to do it. You have the ability to do it. I know you’ve mentioned paralysis and struggles with perfectionism—don’t let the perfect be the enemy of the good. You don’t have to have an exhaustive list right now. But there’s no need for anyone to post the n-word in full on this site, not even for two minutes while a mod is alerted.

Suedehead, FWIW, count me as another POC user who will be starting the same countdown. And at the same time, as someone who is really hoping it doesn’t come to buttoning. I keep sticking around because I have often seen the good in Metafilter, but it’s getting harder and harder to ignore the blight of systemic racism that keeps it from reaching full potential. But cortex and the mods still have a chance to steer Metafilter towards being better, and this very clear, specific request from xtine is one small but important way. Start there.
posted by hurdy gurdy girl at 12:25 PM on June 28, 2020 [11 favorites]

I am a professor, and I work with faculty pretty much every day, and the extra analysis and fretting about edge cases and pretty much trying to get everything perfect the first time is very familiar. In some cases, it can be very useful. In a lot of cases, it’s frustrating noise when a decision needs to be made. This is the second kind of situation.

Rather than discussing variations and additional ideas, it would have been much better for cortex to say “yes, something like this should be doable, we will get back to you by Sunday with a plan.” And then come back with a clear and simple plan. “People will not be able to post FPPs and comments with this word. If they do, they will get this kind of message.“ At this point, I suppose there could be some discussion about the wording of the message, but the job would have been done. Expanding the list would have been an excellent job for the proposed BIPOC Advisory Board, and should be reviewed maybe twice a year for new inclusions (and possible if experience shows an issue).

Instead, we got a couple of days of prevarication and increasing anger that certainly hasn’t helped the community any. Sometimes, less discussion is more success. Say less, listen more, act more. If there was some technical barrier to implementation, that could’ve been raised for discussion, but that was really the only reason to say no, and it does not seem to be at all the case.

I really hope this can be done. I took a month and a half off from MetaFilter after the disasters of last summer, because I really couldn’t bear to interact here after so many of our members just gave up because they weren’t being listened to. Please listen more closely.
posted by GenjiandProust at 12:44 PM on June 28, 2020 [18 favorites]

Rather than discussing variations and additional ideas, it would have been much better for cortex to say “yes, something like this should be doable, we will get back to you by Sunday with a plan.”

I just want to say briefly that I agree. I'm still working on shifting when and how I response to stuff on the site, and when to get into details, and I recognize I get into the weeds too soon and too easily sometimes in a way that is legitimately frustrating for folks. I apologize for that, and will continue to try to be clearer and more focused in my responses to stuff here.

As above: we're going to make this hard-stop comment filter happen. I understand the request and the motivation and it's doable, and we're doing it. It'll be the main dev priority this week, and will probably still be a few days if that goes well but a week feels like a reasonable target to start with. We'll update folks as we make progress there.
posted by cortex (staff) at 12:51 PM on June 28, 2020 [23 favorites]

I have made Mitheral button. I did not know that mitheral is a person of color, there was no way for me to know. It is not my place to call-out derailing comments by users of color. I responded to mitheral because they quoted me in their comment, I thought they were talking to me specifically. I still agree with other users in here and in the other metatalk that the conversation about "edge cases" is a derail. that doesn't erase what is a huge misstep on my part. I'm not sure how to behave in these situations. I don't want to only support users of color who agree with me. I want to contribute to fighting the racism on the site. I want to default to users of color, but I don't want to put full burden on them, or have them to label themselves if they don't want to. I'm not the person who needs to decided how to navigate this in the most constructive manner.
posted by FirstMateKate at 1:07 PM on June 28, 2020 [4 favorites]

I can not imagine the ~behind the screen overhead~ that this place requires; and I fear that this so well intended; and necessary improvement; will add many, many more hours than ever imagined to site maintenance.

Plan on adding another mod folks. No way this is going to work without almost two people modding at a time in the future; and do well to realize that means more resource usage.
posted by Afghan Stan at 1:08 PM on June 28, 2020

FirstMateKate: I'm not sure how to behave in these situations.

I would try this:
- Apologise briefly, but wholeheartedly
- Then let it rest or take your feelings elsewhere. People on this thread have plenty to worry about, let's not make them do the work of worrying about you.

Keep in mind that we all make mistakes.
posted by Too-Ticky at 1:18 PM on June 28, 2020 [5 favorites]

I can not imagine the ~behind the screen overhead~ that this place requires; and I fear that this so well intended; and necessary improvement; will add many, many more hours than ever imagined to site maintenance.

The functionality to block a comment based on a specific string of characters already exists here, for what it's worth, so it's not completely new functionality. I don't see how a slur being blocked is going to require a second mod on duty.
posted by EndsOfInvention at 1:24 PM on June 28, 2020 [15 favorites]

To clarify - i'm not asking for anyone to mediate what just transpired. I don't need anyone to "worry" about me. I am opening up a place for users of color, if they have opinions, to voice how they would like white members in the future to balance fighting (perceived) derails, vs. standing back and only lifting up the voices of people of color. I recognize that this is labor I'm asking for. This might be better suited to the hopefully-soon-to-be-instituted advisory board.
posted by FirstMateKate at 1:34 PM on June 28, 2020 [3 favorites]

FirstMateKate, let's virtually go somewhere else, right now, and sit down together with a cup of tea. Please imagine me inviting you to do exactly that. This thread can do without your input, and mine.

Come on, let's go. I'm buying.
posted by Too-Ticky at 1:41 PM on June 28, 2020 [13 favorites]

We need room for people to make mistakes, and admit them and to reflect on how not to repeat them. I think FirstMateKate's comments are in that spirit.
posted by meaty shoe puppet at 2:01 PM on June 28, 2020 [12 favorites]

Have not been following this thread until today, because I saw the original request when it was first posted and assumed it would promptly be approved and addressed, especially in light of the recent anti-racist MeTas. That it took so many comments and derails (and emotional labor, and rehashing of gaslight-y/strawmen arguments that we've seen before on this site) to arrive at "we're going to make this hard-stop comment filter happen" - is... amazing.

Thank you to all who spent their precious time and energy arguing for this when they shouldn't have had to argue/convince people (esp the admin, who are supposed to be listening...) about this in the first place. Thank you to admin for finally listening... but next time could you please not make it so hard for people to ask for this and have to justify themselves to you? It's these kinds of arguments and these dynamics that make it so exhausting for marginalized groups to be here and speak here.

---

also - thank you gudrun for educating us on the k-word. I come from the region (SE Asia) where that fruit originated, and that fruit is commonly used in cuisine here. When we refer to it in English, that k-word is used here (not just in everyday speech, but in published recipes, English-language grocery listings/signage, etc).

When I first read your comment I was horrified that I might have inadvertently hurt someone through my usage of that word. (especially because I used to post recipes and talk about food online D: I went back and checked my posting history and thankfully this wasn't the case but it could well have been if I had chosen to share a beef rendang recipe instead of, say, a chicken curry recipe.)

Now that I know, I am definitely not going to use that word, and will also encourage friends and family not to use it. That word for the fruit is not used as a slur in our cultural context, and it's very commonly used here, but if people can find an alternative term (and of course we can! and there are!), we should learn to use an alternative. The choice is basically: minor inconvenience vs very hurtful slur with traumatic racist history. So... we should know what to choose.
Thank you gudrun for speaking up about it.
posted by aielen at 2:04 PM on June 28, 2020 [14 favorites]

I am a white user who has spoken up in this thread about my belief that this feature should be implemented as soon as possible. I'm joining suedehead and others in stating that if it isn't implemented in a month, I will take a year long break from the site.
posted by fairlynearlyready at 2:34 PM on June 28, 2020 [8 favorites]

I've not participated in this thread but I want to say this very loud for all to hear: ABOUT DAMN TIME. This kind of transparency and promise is what many people on this site have been asking for, for a long ass time.

I hope you follow through on this Cortex, mods, etc. People of this community are watching and this is just the first of many more things that should be followed through on.

I applaud everyone who pushed hard for this to happen. Keep up the work friends.
posted by Fizz at 2:52 PM on June 28, 2020 [15 favorites]

I can't believe this is still where we're at two days and 130+ comments in.

Sensible feature requests shouldn't be this hard. Here's what an iterative approach could look like:

1. Do Immediately: Change the 'Note: Everyone needs a hug.' at the bottom of the comment box to something like 'Note: we have now implemented a zero-tolerance policy for posting racial slurs and other hate speech on this site.' And update the thread as others have suggested.

2. Do in the next 5 days: Implement the request as suggested to block slurs from being posted; update or change the Note as necessary. Ask for help if needed.

3. Put on the roadmap: Implement the mod-side tools, other enhancements or anything not user-facing.
posted by iamkimiam at 3:58 PM on June 28, 2020 [7 favorites]

1. Frog.
2. Hop.

In this case, not the best scenario. For me the discussion, however cumbersome, tedious, and painful it has been, is worth the result, as long as the discussion is more than troll-food. I favor the move to eliminate certain words. They should not see the light of day on this board. I learned the term "BIPOC" today. I may have learned other things, too.

Dealing with the "edge" words and concepts is another hill to assault another day. It's not up to the mods to do all the heavy lifting, nor clever programs to decide what our prerogatives will be.

This thread has worn me out. I can't imagine how much energy it must have cost those to whom this discussion represents painful personal experiences.

[I'm sorry, I had to re-edit to remove the word "doesn't" in front of the word "represents."]
posted by mule98J at 4:28 PM on June 28, 2020 [2 favorites]

FirstMateKate, how can you write "I did not know that mitheral is a person of color, there was no way for me to know," when, in your comment, you spelled out the A word slur Mitheral referenced? After Mitheral wrote how often that slur, along with two others, had been lobbed in their direction? I don't understand how you read that comment, which mentions their own experiences on the receiving end of abuse; followed the provided links; then posted your response with the full slur (while attempting to take Mitheral to task for a "fallacious argument" derailment) -- only to assert later that there was no way to know.

Please, don't try to explain. This is me trying to explain to you, in an extremely limited way, how startlingly awful it was to read all that. Take Too-Ticky up on their lovely offer of virtual tea.
posted by Iris Gambol at 4:44 PM on June 28, 2020 [10 favorites]

I mean . . . Jeez-oh-man. . .

For everyone who is - in the most charitable interpretation - succumbing to your Nerd Urges to game out all possible ramifications and complications and considerations of edge cases;

Mississippi just voted to eliminate their state flag because Confederate, they're taking them all down as I write this, and they haven't yet even established a committee to begin to come up with a new design.

Does this maybe give a little perspective in how trifling and obnoxious it is to hem and haw about how to implement a method to eliminate a handful of obvious slurs on a website that's used by (maybe) a few thousand people?

Yeah, yeah, yeah, this is Metafilter where we can overthink a plate of beans, but some of you gotta figure out how to tamp that shit down when it's not appropriate.
posted by soundguy99 at 6:07 PM on June 28, 2020 [33 favorites]

Since there doesn't seem to be a lot of slurs thatnappear on the site, I can understand the admin(s), who actually has to deal deleting/managing those rare occurrences, not being in the "this must be done ASAP" frame of mind, because it doesn't happen the much. It can be difficult trying to manage the resources to maintain the site, while viewing dealing with sudden demands and needs from the readers.
posted by Brandon Blatcher at 8:03 PM on June 28, 2020 [17 favorites]

When it happens at all, it injures other humans who are right here with us. It is actively and deeply harmful to other people. I literally cannot believe the handwaving that is on display in this thread. And let's be real, this is not a "sudden demand," folks have been asking site leadership as well as the rest of us to pay attention to these needs for a hell of a long time now. This is not sudden, you might even say it's 400 years in the making. If I was the kind of person who says "smdh" that's what I'd be yelling from the rooftops right now, along with Black Lives Matter.
posted by fairlynearlyready at 9:40 PM on June 28, 2020 [4 favorites]

(I wish people would clarify what they mean by edge cases.)

There are words that are slurs in US English which a never a slur in other dialects - such that entire tv series names had to be changed for American audiences.

There are also words that are slurs in some contexts, but are reclaimed words in others, such as queer and dyke, as in "Queer Nation" or "Dyke March".

Obviously, this has nothing to do with the serious - and much more widely known - slur that is the main impetus.

But if other slurs were to be added, we should discuss them. I'd like to continue to post about my local Dyke March. The list that was developed for a Twitter bot is much too expansive and doesn't recognize the different uses of words - which doesn't matter, since it was just for a Twitter bot.
posted by jb at 10:33 PM on June 28, 2020 [3 favorites]

I like the idea.

Regarding the side discussion of which racial slurs are "obvious" -- as people talk around them in this thread, I know for certain what "n word" refers to but I'm only like 80% confident that I know what "c word" and "k word" refer to.

All the others mentioned? I didn't even know there *was* an "a word" or what it was until it was linked in this thread. And I'm still completely baffled as to what specifically the "h word" and "i word" refer to because there are multiple words that begin with those letters on the slurs lists linked. And I can't figure out what the "k word" that is sometimes a type of lime is either? But it seems like everyone else here knows?

So no, it's not at all obvious to those of us who grew up in community where racial slurs weren't common (my hometown was still hella racist but in a different way) and who in adulthood have been mostly successful in avoiding the type of mean-spirited people and media that routinely pepper their speech with slurs. From memory, I can only think of... like six? racial slurs that I even knew before reading this thread, and one of those I only just learned a week ago because of controversy over a local politician using it.

I'm assuming that there won't be a lot of words with double meanings on the slurs list, so hopefully this will be a non-issue because it's not like I'm going to be using words that I didn't even know existed. I just wanted to chime in to confirm that yeah, some of us actually *are* that ignorant of what most people might consider to be very obvious and well-known racial slurs.

So, when implementing the idea, I like the suggestion to not just disable the "Post Comment" button when the text contains a slur but to also include an error message with a link to the list of the prohibited words and explanations as to *why* they're prohibited. While the priority should be implementing the block ASAP, I urge you please do follow up with adding those explanations later because some of us really do need to be told!
posted by Jacqueline at 11:20 PM on June 28, 2020 [10 favorites]

I’d like as few words as possible to be on the list that prevents posting. A longer list that alerts mods behind the scenes makes sense. (I’d actually thought this existed for some reason.)
posted by michaelh at 11:26 PM on June 28, 2020 [2 favorites]

Why, michaelh?
posted by Kitchen Witch at 11:44 PM on June 28, 2020 [1 favorite]

> From memory, I can only think of... like six? racial slurs that I even knew before reading this thread

Must be nice.
posted by lucidium at 3:26 AM on June 29, 2020 [9 favorites]

I would like as many racial slurs as possible to be on the list that prevents posting. Edge cases should be automatically decided against to prevent long, drawn-out, nit-picky arguments (such as these) about why that's not 'really that bad'. I have had many decades experience of arguments on the internet with self-described 'logical' people who want to tell me "But we call English people Englishmen, French people are called Frenchmen so why shouldn't I be able to call you (racial slur)?"

I cannot help but draw parallels between "but if there's a ban list of racial slurs there is a possibility my precious freedom of speech will be infringed upon" despite any damage that may be caused to BIPOC and "wearing a mask infringes on my liberty" despite any viral transmission that may occur.
posted by Comrade_robot at 5:06 AM on June 29, 2020 [8 favorites]

Zengargoyle, if you happen to read this at some point: I’m one of the people who flagged your first deleted comment (thanks taz) and I’m still angry. I hope you noticed what the OP wanted is going to happen just fine without much problem, and the solution you claimed to care about didn’t actually require brainstorming various offensive sexual terms that I barely tolerate on porn sites and didn’t need to see tossed into conversation here as random examples. If you could devote even a little more effort away from solving the problem tech something something into remembering words have meaning, can hurt people, and are not interchangeable parts in a thought exercise that would be great.
posted by sacchan at 5:58 AM on June 29, 2020 [7 favorites]

Freedom FROM harm is more important than freedom TO harm. Slurs hurt, whether it was the intent of the use or not. And speaking of hurt, it hurts to see this handwringing performance art over a small change that would have immediate positive effects for the diverse community of Metafilter. We do want to keep Metafilter diverse, right? Because the days of getting away with doublespeak are over. We excoriate Facebook and other companies for their refusal to make substantive changes, but here on our own website we are debating—debating!—the implementation of a small but substantive change.

We need to do better. Metafilter has given me a lot over the years, and it hurts me to see this and other threads pay lip service to small structural changes while spending millions of words talking about making radical structural changes in society. Just shut up and listen to yourself. It is embarrassing. My iPhone doesn’t have a problem changing fuck to duck. Why does Metafilter have a problem with blocking the n word? It is like the recent Pro Publica article about zoning changes in Connecticut: outwardly a liberal bastion but only for white people.

I need this community to step up and make these changes.
posted by apex_ at 6:11 AM on June 29, 2020 [11 favorites]

> From memory, I can only think of... like six? racial slurs that I even knew before reading this thread

I'll believe you that this is the case in your life, but frankly, this genuinely boggles my mind. Ignoring edge cases totally (like, ambiguous words with multiple meanings depending on context), plain and simple unambiguous slurs are used all the time in fiction and non-fiction literature, in movies and TV shows, in songs and other creative media, and these days are repeated on the news by leading politicians. We get exposed to them simply through routine media consumption, even assuming one is fortunate enough to never be around people who use them in speech.

And, if you have a friend/social group that is at all diverse, plenty of those people have had numerous slurs directed at them, whether or not they choose to talk about that with you.
posted by Dip Flash at 6:41 AM on June 29, 2020 [2 favorites]

I would like as many racial slurs as possible to be on the list that prevents posting. Edge cases should be automatically decided against to prevent long, drawn-out, nit-picky arguments (such as these) about why that's not 'really that bad'.

for slurs that are, in practical terms, solely slurs (i.e. the word has no non-slur alternate meaning or is very unlikely to be mentioned on metafilter in the context of that alternate meaning or as a reclaimed term), i would like all of them to be included on that list, and for that list to be implemented as soon as humanly possible. that should not have been a decision that required a thread this long.

where a word is a slur and also has a non-slur alternate meaning (within or across cultures), or where at least some people targeted by the slur would like to be able to use it here as a reclaimed term, i think there should be further discussion about how to deal with that word - transparent discussion, guided by input from those affected. some words should probably be added to the no posting list. others may be better dealt with through mod pre-approval for posting, auto alerts, or some other means falling between 'impossible to post' and 'leave untouched'. that will involve complicated discussions. auto-banning every single edge case is also going to result in complicated discussions. there is no approach to this that is going to avoid complicated discussions.

we don't need to have any of those discussions right now, and the prospect of them shouldn't distract from the extremely basic step of preventing the n-word being used on metafilter.
posted by inire at 7:22 AM on June 29, 2020 [5 favorites]

that will involve complicated discussions. auto-banning every single edge case is also going to result in complicated discussions. there is no approach to this that is going to avoid complicated discussions.

It's not really a complicated discussion -- don't use ethnic slurs, and use a word that's not also an ethnic slur if you feel like you need to express an idea. Again, I have witnessed decades of arguing over 'edge cases' on the internet ("How can this possibly be slur?! It's just an abbreviation of a term describing somebody from the country of Japan! Let's have a long nit-picky discussion!") I have no confidence that the end of the 'long, complicated discussion' is one that ends on the side of BIPOC.
posted by Comrade_robot at 8:18 AM on June 29, 2020 [5 favorites]

Allowing this slur unchecked on the site is a bug. If you think of this as a bug fix instead of an opportunity to reply guy yourselves into a sprint planning, does that help?
posted by fluttering hellfire at 8:52 AM on June 29, 2020 [2 favorites]

It's not really a complicated discussion -- don't use ethnic slurs, and use a word that's not also an ethnic slur if you feel like you need to express an idea.

in the interests of trying to explain why complicated discussions would end up happening without actually having them in this thread, i would only suggest that you look through the wikipedia list of ethnic slurs linked upthread (which is not even the most comprehensive such list), and think about whether you're happy to have every word on that list auto-banned from metafilter.

including, for example, the english word starting with 'i' that is commonly used to refer to people from india (including by those people to refer to themselves), or the common US / UK name beginning with 'j' that is also an anti-german epithet and the name of the cartoon mouse, or the english name (beginning with 'l') of the ravenous insect that is currently swarming across kenya and other parts of africa and threatening local food supplies, in what should be a bigger news story than it has been. (to emphasise, we don't need to hammer out how to deal with these words in this thread, i'm just flagging them as examples of the kind of 'edge cases' that might come up in future discussions)

if your reaction is that some of those words should be dealt with in a way other than auto-banning, or that this is a badly put together list and we should use a different one... well, that's where the 'edge cases' and discussion come in, and i don't see how that will be short or simple. after the last few metatalks, i share your lack of confidence in long complicated discussions having good outcomes for bipoc mefites, so i wish this was something that could be fully dealt with easily and straightforwardly. bits of it can be. other bits can't.
posted by inire at 9:10 AM on June 29, 2020 [4 favorites]

I am a religious minority who has had slurs thrown at me. I'm also part of a gender & sexual identity minority of which some terms are in contention and some of which I personally reclaim.

I offer that to say I understand a part of "these words cannot ever be used" and "some words are legitimately debatable" but I also think that's a completely different discussion from this simple and absolutely necessary request. Maybe the banned language proposal could be its own MetaTalk, and we keep this one only for updates on the specific roll-out?

If this feature isn't implemented by a month from now, July 28th, 2020, then I will stop participating in this community for one year, until July 28th, 2021.

I cannot in good conscious financially support a website that allows the n-word to be posted. I'm using this deadline for the end of that monthly donation, if I may.
posted by hapaxes.legomenon at 9:53 AM on June 29, 2020 [6 favorites]

White people on this site will wargame out elaborate scenarios ("What if aliens land and Metafilter is the only means of communication and that slur is a compliment in their language...?!") rather than listening to BIPOC that tell them the word hurts them, because somehow the second one is the unbelievable part.
posted by Ghostride The Whip at 10:03 AM on June 29, 2020 [30 favorites]

I also just want to note that if you want people to give you their money, time, and ideas, then the answer to when the site will become more usable can't be "We'll get to it when we get to it, we're busy" even if that's the truth. I know it seems like a chicken & egg problem but it's not. You could just say "Here's when we're going to do it". Nobody thinks you're all just sitting around with nothing better to do or expects this just to get turned around in an hour but we do expect that changes to the site to make them usable are a fair bargain in exchange for the donations & the time spent on this free design work. This isn't a favor we're asking for.
posted by bleep at 10:11 AM on June 29, 2020 [5 favorites]

Mod note: hi all, here to encourage folks to keep this thread focused on the topic at hand, which is as many folks have mentioned, the specific feature request made by xtine. Discussing the future of potential banned language is a necessary conversation that could happen in a thread of its own.
posted by travelingthyme (staff) at 10:42 AM on June 29, 2020 [12 favorites]

^ This message would have been great several days ago.
posted by bleep at 11:45 AM on June 29, 2020 [7 favorites]

Okay, this isn't user friendly yet, but here's a Chrome extension that basically packages the above code up. To install it in Chrome:
- Download and unzip the files somewhere on your computer,
- Go to Window / Extensions,
- toggle "Developer mode" on at the top right,
- click "Load unpacked", and select the folder called "mefiplusplus".

It just looks for certain strings as you type, and if it sees them it removes the Post Comment and Preview buttons and appends a message to your comment. You can get the buttons back be refreshing the page. I've tried to make it do no harm, but I've had about four hours of sleep since Saturday, so I'd appreciate it if anyone savvy could double check me.

I'd like to follow up with some longer thoughts, but if I don't it's probably because I decided better and went to sit down.
posted by lucidium at 12:22 PM on June 29, 2020 [1 favorite]

Since I originally pointed to Kazemi's list in this thread (in the spirit of "If you want help starting that list" for MetaFilter's use), I followed up with gudrun privately regarding this comment:

This is one of the two words that jumped out to me as being left off of Kazemi's list, and why I said it has some flaws.

and helped get both of those two words added to Kazemi's wordfilter list. (See "pull request" items #40 and #41.) I'm mentioning this here so future visitors to this thread don't get completely confused about what gudrun was talking about! Also, if you are not familiar with GitHub (the platform where Kazemi's project is hosted) and you want help suggesting a word for addition to that list, you can MeMail me.
posted by brainwane at 12:39 PM on June 29, 2020 [8 favorites]

So, I try to think of a few things when I do things: better is better than nothing, now is better than later, most things are reversible, everything harms someone, try to design for the most vulnerable first, then build from there.

I know how hard it is being on a small team and carrying the workload of a huge company, but there is also an army of caring volunteers here. I think we need a team of techy people on call 24/7 for small things like this, and we also need a site-culture where small, hacky, temporary fixes can be deployed without oversight and with understanding that they'll break things for some people.

I think we can find a team of people like that, I would like to be one of them for free. I'm on BST, I have a full time but flexible job.
posted by lucidium at 12:54 PM on June 29, 2020 [9 favorites]

For reference, my desired response to this request would have been to immediately append some javascript that wraps certain words in spoiler tags to every page, add a (experiment in progress, see metatalk) message up top, then see what breaks.
posted by lucidium at 12:58 PM on June 29, 2020 [1 favorite]

I hope this is not a derail, but the plan can’t be for all MeFites to individually install something on their own browsers that prevents them from posting slurs, right? I’m genuinely confused here about how this can be done effectively in a client side way when the people we need to mute aren’t all that eager to be muted, but it’s come up several times so I want to check my understanding.
posted by eirias at 4:17 PM on June 29, 2020 [2 favorites]

Eirias, that is not the plan, it was just a tech savy Mefite trying to help out.
posted by Brandon Blatcher at 4:21 PM on June 29, 2020 [2 favorites]

That's my bad for being the "why don't we just" guy. I have spoons, I'm just offering the ones I have spare.
posted by lucidium at 4:32 PM on June 29, 2020 [3 favorites]

The effort and thought is no doubt appreciated, but maybe just let it go for now.
posted by Brandon Blatcher at 4:35 PM on June 29, 2020 [2 favorites]

So I guess I'm doing this. I'm visibly Chinese, I don't generally experience much aggression in the street because I'm mean looking, but my loved ones do. I won't type what they hear, but assume it begins with something that rhymes with "suck cough" and ends with "China".

I've given *checks notes* $245 to metafilter recently, which I share because apparently I need bonafides. It's fine, my therapist costs more.

See you next week.
posted by lucidium at 4:41 PM on June 29, 2020 [12 favorites]

FFS, can white people stop posting stupid objections and just implement the banned words filter already?

Also mods, this is going to sound harsh, but if you have time to do a podcast, you have time to implement some basic code.
posted by daybeforetheday at 2:29 AM on June 30, 2020 [14 favorites]

Xtine's proposal treats all posters equally, preserves the poster's dignity by giving them a chance to educate themselves before linking hate speech to their profile and thus possibly their real-world identity, and even provides a way for posters to still post the word if a mod says it's OK. There is no better way to implement something like this than what has been suggested. I support it totally.
posted by mdonley at 3:18 AM on June 30, 2020 [9 favorites]

There’s no solution yet and it’s Tuesday.

Here’s how I’m looking at this situation: If I worked somewhere that had that word written on the wall in graffiti, and management didn’t do everything necessary to remove it immediately, I would no longer work for that company. This isn’t complicated and it’s frankly embarrassing.
posted by Drumhellz at 10:27 AM on June 30, 2020 [4 favorites]

At the very least, maybe a note that it is being worked on and a date when we could expect to see it implemented. This would be an excellent time for some transparency and advanced info. Even if there is some glitch in the process, and you have to say “we are working on this, and we hope to have it done by Friday, but it may be the 10th.” That wouldn’t be good news, but it would be news.
posted by GenjiandProust at 11:17 AM on June 30, 2020 [3 favorites]

Instead what we're going to get is yet another "Yeah, sorry for the lack of updates, I will do better" and a bunch of valued community members who we will lose for no good reason. This is absurd.
posted by bleep at 11:26 AM on June 30, 2020 [7 favorites]

Well, maybe, but we might be surprised. I live in Hope.
posted by GenjiandProust at 11:37 AM on June 30, 2020 [1 favorite]

The note you're requesting is already posted upthread: it's the main dev priority this week, and the ETA is Friday. frimble is not a full-time dev for MetaFilter, and the feature being discussed is probably significantly more complicated than anyone here is giving it credit for--there is a world of difference between a client-side Javascript plugin and making changes to a 20-year-old ColdFusion app.
posted by Mayor West at 11:52 AM on June 30, 2020 [18 favorites]

Thanks, I'd missed that. Or read it and lost track.
posted by GenjiandProust at 12:04 PM on June 30, 2020

I didn't miss it I just found it insufficient.
posted by bleep at 12:09 PM on June 30, 2020 [6 favorites]

I have now blocked the use of the n-word in comments across all subsites. I will also do the same for posts.

The message that appears is the following:

Your comment includes the n-word, the history of which as a racial slur makes it unacceptable to post on MetaFilter, regardless of intent. If you are a quoting a passage or title, please rewrite to obscure the actual word, if it needs referencing at all. The post button will be disabled until it is removed. If you have a question, please contact the moderators.

posted by frimble (staff) at 12:32 PM on June 30, 2020 [113 favorites]

Thanks frimble!
posted by Brandon Blatcher at 12:38 PM on June 30, 2020 [4 favorites]

It's a good start; thanks.
posted by GenjiandProust at 12:39 PM on June 30, 2020 [3 favorites]

Flagged.
posted by meaty shoe puppet at 12:56 PM on June 30, 2020

Frimble, would it be hard to extend the filter to check for the spelling that ends with - GA?
posted by meaty shoe puppet at 12:57 PM on June 30, 2020

(in case people are confused by the "flagged" comment, I was trying out the new filter and briefly posted another spelling of the n word that the filter did not catch.)
posted by meaty shoe puppet at 1:06 PM on June 30, 2020 [3 favorites]

Thank you frimble
posted by sallybrown at 1:08 PM on June 30, 2020

Frimble, would it be hard to extend the filter to check for the spelling that ends with - GA?

No. Have done.

(in case people are confused by the "flagged" comment, I was trying out the new filter and briefly posted another spelling of the n word that the filter did not catch.)

If people are looking for edge cases, the same filter is also used on comment preview.
posted by frimble (staff) at 1:13 PM on June 30, 2020 [22 favorites]

God, thank you, frimble.
posted by Kitchen Witch at 2:05 PM on June 30, 2020 [3 favorites]

Thanks to everyone who made this happen. It's long overdue and hopefully (I'm nothing if not an optimist and yes, sometimes that's not good) it's part of (again long overdue) real, actual change.
posted by cooker girl at 2:08 PM on June 30, 2020 [1 favorite]

If anyone is interested in taking this filter further, Unicode provides a list of confusable characters for use in profanity filters. For instance - n and 𝗇 are near-identical, as are 𝗋 and r.
posted by saeculorum at 2:35 PM on June 30, 2020 [2 favorites]

Thank you, frimble, and mods, for doing this with all speed.
posted by blurker at 2:47 PM on June 30, 2020

Wow, that’s awesome thank you Frimble :)
posted by Drumhellz at 2:52 PM on June 30, 2020

This is great! Thank you to everyone who helped make it happen! ❤️
posted by Susan PG at 3:09 PM on June 30, 2020

Thank you frimble, I have never been so happy to be wrong!
posted by bleep at 3:10 PM on June 30, 2020 [2 favorites]

If anyone is interested in taking this filter further, Unicode provides a list of confusable characters for use in profanity filters. For instance - n and 𝗇 are near-identical, as are 𝗋 and r.

You know, I say, if somebody really wants to try to circumvent this filter, they probably will be able to, and then the mods should ban them. No second chances, no messing around, just immediate banning.
posted by GenjiandProust at 3:22 PM on June 30, 2020 [33 favorites]

Yeah. And to be clear, it's not a should thing for us: that kind of malicious behavior is something we would ban someone for, period, end of.
posted by cortex (staff) at 3:28 PM on June 30, 2020 [29 favorites]

Excellent news! Thank you frimble and everyone who made this happen!
posted by hurdy gurdy girl at 6:08 PM on June 30, 2020

Thank you Frimble, Cortex, and everyone who pushed for this.
posted by Alterscape at 6:19 PM on June 30, 2020 [1 favorite]

Thank you.
posted by xtine at 7:09 PM on June 30, 2020 [9 favorites]

Thank you xtine for voicing this and being the catalyst.
posted by fairlynearlyready at 9:14 PM on June 30, 2020 [1 favorite]

Agreed. Thank you xtine!
posted by hurdy gurdy girl at 9:21 PM on June 30, 2020 [1 favorite]

Thank you, xtine and frimble!
posted by Pandora Kouti at 11:02 PM on June 30, 2020 [1 favorite]

Thank you so much, xtine!
posted by Kitchen Witch at 1:08 AM on July 1, 2020

Thank you xtine!
posted by daybeforetheday at 1:17 AM on July 1, 2020

Thank you frimble and xtine!
posted by capricorn at 7:49 AM on July 1, 2020

Thanks to xtine and frimble!!!
posted by kimberussell at 9:46 AM on July 1, 2020 [1 favorite]

Thank you!

I’m very heartened to see this change.

I also feel a little foolish and embarrassed about my message above. (I almost wrote a comment taking it back). And at the same time I’m so so glad to be, because it indicates that my comment was an underestimation, that I’m proven wrong. Thanks cortex and frimble for making it happen, and xtine for bringing this up and keeping us focused on the central request.

I hope that we can have a sense of urgency and importance about other issues regarding race and whiteness here, similarly, and move to change things. A great many ideas regarding site features was floated last year in the race Metatalks, and I think it will be important to incorporate new site features alongside moderation practices. For example, I recall that some ideas that people floated was making offensive comments hidden-but-viewable or grayed out, so that the comment wasn’t visible or clearly understood to be problematic, yet comments weren’t deleted outright to provide a sense of community accountability / prevent against erasure. Facilitation/moderation practices need helpful methods and tools, and I think the site itself actually changing will be an important aspect of making Metafilter an actively anti-racist community.

As a whole I am really looking forward to change and growth; this is a great place to start and I’m looking forward and excited about the other things that we could change together, here.
posted by suedehead at 10:40 PM on July 1, 2020 [21 favorites]

Very much no reason to feel foolish; there was (and is) a lot of precedent of well-meaning-sounding prevarication preceding months delays with excuses afterwards. The very firm 'ultimatum' limits-setting of what is personally tolerable probably helped drive the (comparatively) swift implementation of this no-brainer filter.

The very tastiest crow to eat will be if future incidents of no-brainer changes are implemented without the need for hundreds of comments including initial non-committal wall-of-qualified-frets mod notes. I very much look forward to eating that crow myself!
posted by Drastic at 7:18 AM on July 2, 2020 [5 favorites]

I also feel a little foolish and embarrassed about my message above. (I almost wrote a comment taking it back).

You should not feel anything of the sort about that message. I'm embarrassed for Metafilter that you thought, even for a moment, that it was necessary to put such measures in place for your own sanity/well-being. You are a valuable member of this community.

cortex, this is definitely a good step, but my god, that was a low bar to clear. Stop worrying so much about alienating white people and just do the right thing because it's the right thing to do.

When you feel stuck on this sort of thing, you can borrow my yardstick for progressive change if you like: What would Jessamyn do?
posted by heyho at 7:20 AM on July 2, 2020 [10 favorites]

Yardstick, playbook, whatever. Words are hard some days.
posted by heyho at 8:34 AM on July 2, 2020 [1 favorite]

Speaking of words; were the words part of a digital list; or did somebody have to actually type them in as part of the code?
posted by buzzman at 7:10 PM on July 2, 2020

If I'm understanding correctly (I'm not about to test) the only words controlled at this time are the original request and variations.

Maybe this should be going to a new meta (and I'll write one up if that is the way we'd like to go) but in the meantime cortex is there currently a plan to expand the list? Are we going to have a discussion about words that should be on it or is the initial expansion (if one is planned) going to be determined by staff?
posted by Mitheral at 7:56 PM on July 2, 2020 [2 favorites]

Mitheral's correct, we're only handling one word with two variations in the current implementation. frimble did some extra plumbing while getting this set up to make it a little less onerous to make further changes, but we were also able to work faster in service of this specific case than if we were trying to handle a broader set of possibilities from the get go.

We plan to work out a more expanded list, and ways of handling a broader set of cases well; those are both just at the discussion phase at this point, though there's a short list of common (in racist/homophobic/etc reality, though not common in practice on MeFi) slurs that are obvious candidates for inclusion if we're going to flag stuff down more widely. My feeling is that a core list would be easy to populate but necessarily not be complete, so even if we do roll something out prior to further community discussion I think that discussion would be helpful in iterating on the list.

There are two major design challenges that I've been considering as we look at a broader list: where to draw a line on "you cannot post this word/phrase" vs. "you should be careful when posting this word/phrase", and how to handle false positive substrings for a broader variety of words where some of them will exist within other unrelated words with greater frequency than is the case for the n-word variants. We could skip dealing with those for this first bit but will have to tackle them when looking at expansion. We'll also have to have more flexible messaging when capturing a wider variety of terms so that the warning/stop messages aren't unhelpfully vague or thoughtlessly lumping different cultural and language issues together. Documenting and referring to a list of what's blocked would need to be part of that as well.

I don't have a specific timeline for all of that right now. It's something I expect us to review as a team in our Sunday meeting and firm up a bit.

A MetaTalk to discuss that broader set of word list and messaging challenges seems like a totally reasonable idea to me, so if you're interested in tackling that and/or collaborate with the mod team on drafting something up, I'd welcome that Mitheral.
posted by cortex (staff) at 8:34 PM on July 2, 2020

OK, I work on it over the weekend and have it ready for Monday (I don't think anyone wants to have this conversation over the US Independence Day weekend.)
posted by Mitheral at 10:28 PM on July 2, 2020

When the topic is posted you might want to consider how we can have this discussion without actually mentioning the words.
posted by DanSachs at 3:08 AM on July 3, 2020 [2 favorites]

Well that de-escalated quickly! Good work, and it is really refreshing to see one of these threads end up with progress and positivity.
posted by snofoam at 8:34 AM on July 3, 2020

Otherchaz : Let's move all humans to White, and automation to Black. Paint them Vanta Black.
"The robot. It's magnificently fit to build your car, but would you let your daughter marry one?"
Then we get to post "How the Africans became White"

You sure about that, snofoam? (._.)
posted by meaty shoe puppet at 8:37 AM on July 3, 2020 [1 favorite]

DanSachs said: When the topic is posted you might want to consider how we can have this discussion without actually mentioning the words.

I have a suggestion about that. At least when it comes to words that are on Darius Kazemi's wordfilter list, they have numbers next to them. (The words I said were not on the list previously have been added now, thanks to brainwave.) So, it would be possible to reference the Kazemi list by number when we talk about what should be on the expanded list for metafilter, for words that are on his list at least. For example, what I was calling the k-word is now # 46-50 on Kazemi's list. #84 now on Kazemi's list was the other word that I saw had been left off, and now is included.

Also, previously when I was talking about the k-word in this metatalk, when it was not yet on the Kazemi list, I linked to an article that referenced the issue with the word for the lime, so that I would not have to include the actual word in my post.
posted by gudrun at 8:53 AM on July 4, 2020

Making people figure out what you mean by going to a list that forces them to read not just the slur you had in mind but a bunch more besides seems... suboptimal.
posted by Flannery Culp at 9:05 AM on July 4, 2020 [12 favorites]

Flannery Culp - if someone has a specific slur to propose be excluded, then my suggestion is indeed not the way to go about it. Then doing what I did with the k-word is my suggestion.

If, on the other hand, you want to (as I do) grit your teeth and make an attempt at a working list of agreed upon words to exclude without having to resort to something like google (I've had to do some of this googling for work related reasons, would not recommend, plus it contaminates your google search history), then using as a source a list or lists that others have already compiled as a starting point is actually a way to avoid some of the nasty stuff you can land on otherwise. I find this all pretty painful to contemplate, personally, and would hope something could get worked out for metafilter without a long painful dive into specifics.
posted by gudrun at 12:28 PM on July 4, 2020

Followup to note that Mitheral's new MetaTalk is now up for folks who want to discuss specific further additions to the comment filter. n.b. it's necessarily a thread discussing hurtful language, so no one is expected or obligated to dig in on it; do what makes the most sense for you and if you want to add thoughts or suggestions without digging in there you're welcome to reach out to me or the contact form directly instead.
posted by cortex (staff) at 9:49 AM on July 7, 2020 [1 favorite]

There are two major design challenges that I've been considering as we look at a broader list...

Another occurred to me:
In my experience, people who want to use these words often have terrible spelling. It's not possible to catch all the potential variations of these, but are there some common ones that should be added? This is already kind of on the list with hyphenated variations

Which brings up - #29 is multiple words. Are multi-word phrases in general eligible to be added to the list, or is this a special case?
posted by bashing rocks together at 11:52 AM on July 7, 2020

Probably best to port further along these lines to the new thread, but: trying to catch all misspellings would be a great deal more difficult; I think common attested spellings is where we need to aim in practical terms. Most of the attested mentions of these words on MeFi historically have been in quotation or folks referencing being on the receiving end of them, so the dicey spelling skills of addled driveby internet racists isn't really what we're designing for here.

Multi-word phrases should be workable in general though it may be harder to avoid ballooning the list when taking that approach so I'd want to size up some examples and think about the impact. Frequency of use is probably the more important consideration, rather than single word vs. multiword phrases.
posted by cortex (staff) at 4:16 PM on July 7, 2020 [1 favorite]