ETA for fuck’s sake, even this post had to be censored down to go through. It took about 15 attempts.
A few times recently I’ve had posts refuse to go through with great big ‘you’ve been blocked!’ banners.
It seems to be very sensitive to mentions of being mean to people, not advocacy, just mentioning the existence of certain topics, and it seems to be way too hair-triggered.
I just now tried to post about the ethics of meat consumption, and why people see certain aspects as more troubling than others. Certainly nothing explicit or provocative, just the difference in perception between harvesting meat and deliberate unkindess - as abstract concepts.
I refuse to do the stupid zoomer thing of cens*ring words with numbers and punctuation, and frankly I shouldn’t have to.
I’m not sure how you’re meant to be able to have a sane conversation about ethics or politics if you’re not allowed to mention people being mean. God knows what would happen if someone tried to report actions perpetrated in an akka-akka-kablooey-competition zone.
Could someone maybe take a look at the settings, because jesus christ.
Lemmy.ml has extensive word censoring. It’s at the instance level, not community level.
Extensive? Do you have any more details or sources for that?
I’ve just checked the hard coded slur filter and it’s fairly limited and covers only the usual suspects.
To illustrate, some terms that should pass: cunt, fuck face, shit head.
You can say cunt but not a female dog?
Well I’d certainly say that the b word is more of a gendered pejorative than cunt, which would be the point. Thought, obviously the c word can be gendered too, depending on where you are and usage (I’m used to it being fairly neutral though blindly offensive).
I’m pretty sure the other instances don’t have a slur filter
Having a filter at all is rather different from having
extensive word censoring. Do you think having a filter at all is “extensive”? I can understand the reflex against it, but in the end, for indie social media with limited resources, putting a basic hardcoded prevention right in the software seems like a pretty natural and easy to implement measure.The other instances may or may not have the filter, I don’t know … but last I checked the filter is more or less baked into the software (as rust is compiled), which would require recompiling from source to remove, something which I bet a number of instances don’t bother with … though this could be very wrong.
https://github.com/LemmyNet/lemmy/issues/622
From the github discussion it’s an optional feature. See last message
The reason why I think word filters are dumb is evidenced by the reason this discussion started. They randomly censor stuff with no consideration for context or even if the meaning is the same. If widespread, any bad actor can remove or replace random letters to avoid them. And if you happen to use other languages, there’s a good chance you’ll run into random blocked messages for no good reason.
Ah right … thanks! … I’d missed that.
I’m aware of all the issues a simple filter can have, as you describe. My argument is that for an indie social media platform, where high quality and quick moderation is by no means guaranteed, such a filter would generally be worth it.