Hi, I am setting up a danish site. The word ‘sl*t’ means ‘to complete’ in danish. Whenever we make a post with that word, we get an error saying “no slurs”. And the post is not accepted.

It does not help that we have enabled NSFW in community and on the post.

Please help. Thank you :)

  • AgreeableLandscape
    91·
    4 years ago

    Unfortunately, multi-language support for the slur filter has not been implemented yet. But in the meantime, the slur filter regular expression is on this line. You could take out any offending words that are benign in Danish, and then recompiling the backend. I’d be happy to help you with editing the regex if you have troubles.

    • roastpotatothief
      4·
      4 years ago

      It’s a difficult and controversial topic, censorship. I’m sure you’ve already had the debate. You know about the everyday damage and about the risks, and banning certain words is the right approach for Lemmy.

      But nobody wants to go down the road of shadowbans, deleting off-brand subreddits, suppressing certain topics.

      Sometimes censorship really is justified. The only crucial thing is to be transparent about it. If you publish the forms of censorship in use or planned, then it would be much less problematic.

      Are you planning to publish a list of the forms of censorship currently used or planned?

      Ideally you would even publish the list of banned words. That would have prevented the Danish instance’s problem.

        • roastpotatothief
          3·
          4 years ago

          Thanks. I don’t even know half of those. Maybe they are common insults in America.

          I guess in the future each instance will chose its own list based on its region or language, it the culture it wants to build.

      • AgreeableLandscape
        3·
        4 years ago

        But nobody wants to go down the road of shadowbans

        Shadowbans aren’t supported by the Lemmy platform and there are no plans to my knowledge to implement them. In fact, the moderation logs are public, and will show what posts or comments have been removed and what users and communities have been banned. Only very problematic or illegal content are completely removed from the server.

        If you publish the forms of censorship in use or planned, then it would be much less problematic.

        The comment above this one links to the slur filter in the source code, which contains all the words being “censored”. You’ll notice we only have actual slurs in it, words whose primary or only purpose is to attack people, not less serious swear words like shit and fuck.

        The code is open source, and in fact, per terms of the license, everyone who modifies the code base for use on their own server needs to make their changes open source too.

        • roastpotatothief
          11·
          4 years ago

          Thanks that’s reassuring. So you can still have a discussion about precocious people (though i don’t know why you would) but just not using specific protected swear words. So the intention is not to suppress ideas or opinions or ways of thinking.

          But what about that time a post was deleted from lemmy.ml for criticising public health advice on mask usage?

          • AgreeableLandscape
            21·
            4 years ago

            The post wasn’t just mask usage, it was saying that the virus doesn’t matter and no one should be taking steps to prevent its spread. It’s provably wrong information (papers have been written about the subject) and is dangerous if someone actually decides to take it seriously. You can see it in the modlog still.

            • roastpotatothief
              3·
              4 years ago

              That modlog is very interesting. The censorship is transparent just as it should be.

              TBH I strongly disagree with your approach to freedom of speech. But you must be doing something right that Voat did wrong.

              • Star Wars Enjoyer @lemmygrad.ml
                12·
                4 years ago

                freedom of speech

                It’s not critiques of a government we remove, it’s hateful language, misinformation, trolling, and other such things. None of that is “freedom of speech”, it’s filth that has no place on Lemmy. The post you brought up, was one where someone was literally saying the virus wasn’t real, taking that post down wasn’t ‘censorship’, it was janitorially removing misinformation. If you have a problem with this, Lemmy might not be for you.

      • nutomicA
        7·
        4 years ago
        # build the image
docker build . -t docker/prod/Dockerfile -t your-tag
# store to file
docker save -o your-image.tar
# copy to server
scp your-image.tar server:
# load into docker on server
ssh server "cat your-image.tar | docker load"
# then just change the lemmy docker tag in your docker-compose.yml, and run docker-compose up -d

        From memory, there are likely some typos. You can also push to a docker registry instead (like docker hub), but this seems easier.

        Edit: maybe you could add this to the documentation :)

      • AgreeableLandscape
        2·
        4 years ago

        Just a reminder, please make your changes open source too. That’s required by the license.

  • Maya
    6·
    4 years ago

    in the github issue someone else has linked please note that it is eventually planned to make this smarter to avoid this case