342
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 28 Aug 2023
342 points (98.3% liked)
FediLore + Fedidrama
2159 readers
9 users here now
Chronicle the life and tale of the fediverse (+ matrix)
Largely a sublemmy about capturing drama, from fediverse spanning drama to just lemmy drama.
Includes lore like how a instance got it's name, how an instance got defederated, how an admin got doxxed, fedihistory etc
(New) This sub's intentions is to an archive/newspaper, as in preferably don't get into fights with each other or the ppl featured in the drama
Tags: fediverse news, lemmy news, lemmyverse
Partners:
founded 2 years ago
MODERATORS
These comments so far stink, yall are something else.
OK, I am going to take a minute away from the shit stirring and potentially provide some insight speaking as an admin who's had the misfortune of dealing with this so I can maybe shift this comment section into an actually meaningful discussion.
You can have your own opinion and feelings against lemmy.world but, this?
This is correct. Most lemmy admins likely agree as well, I don't speak for anyone but myself but I can say that I think it would be hard to find someone who disagreed. What happened today is a result of a catastrophic failure on lemmys end, with issues that should have been addressed over a month ago just being completely ignored. The lemmy devs shared a roadmap during their AMA & they essentially were more concerned with making shit go faster... that's about it.
Okay, honest question. What mod tools are lacking. If there's something needed, what is that thing or things?
I went over to the feature request page for Lemmy and I couldn't find anything massive in terms of requests for moderation tools that would have been sure fire ways to stop this particular event.
That said, there is over 400 open feature requests alone on Lemmy's github. I obviously couldn't go through every single one. But coming from the kbin side I'm just curious about our Lemmy brothers and sisters. It sounds dire and I'm woefully under informed on how bad it is.
There aren't enough roles. There's admin, moderator, and user, but it would be best to have tiers of user in between. Reports go to 4 categories of user when you file a report. Report a comment for violating a fun rule your community decided to implement (all post titles must contain "Jon Bois Rules!")? That report goes to: the community moderators (good), the community's host instance's admin (bad), your instance's admin (bad), the user who posted the "offending post"'s instance's admin (bad).
Only admins can permanently remove illegal content. If a mod "removes" it, it still sits visible to all in modlog, and for the purposes of CSAM specifically, that counts as distribution which is prosecuted as a worse crime than possession. Federation with other instances is effectively binary. You can or cannot federate, you cannot set traffic as unidirectional like you can on most other fediverse platforms. The modlogs make it hard to parse who the moderator performing an action is acting on the behalf of. Was it a community mod? An admin? Your admin?
There's more but my phone is getting low on battery
Agreed, I don't know what AutoMod did on Reddit but if what mods need is a rule-configurable post remover then I'd be happy to clobber together something in Python
There's this bot that is used in a couple of communities on feddit.de:
https://github.com/Dakkaron/SquareModBot
Oh great! This was literally how I envisioned my python script – JSON config file and all
If you've really got the time and energy I think you would see pretty heavy use of such a tool. I think the existing libraries are definitely mature enough. I've been surprised that nobody has done it already
Nice, do you happen to remember what the most popular moderation rules were? So far I can think of:
I have exams in September but if I get a free day it should be enough to get something working
Wait does lemmy have reporting functionality, so you could use some type of number of reports?
As in Python Lemmy libraries?
Yep. That's what I've been looking at at least
Here's some things Beehaw admind have been asking for from moderation since June: https://beehaw.org/comment/397674
See github issues #3255 and #3275
As an admin, how do kbin moderation tools compare?
Also does lemmy.world have the spare cash to offer cash for features?
Kbin moderation tools are worse. And potentially. I guess a bug bounty could be started up.
EDIT: my app attached my reply to the wrong comment. Please ignore in this context
I don't know this for sure, but I have a feeling that a hard fork is in Lemmy's future. I don't want to get super into it, but programming is a form of communication. What features you bake into a platform are reflective of the messages you want to propogate on that platform. Lemmy's devs vision for what the platform should be might not be reflective of what most of us might think it should be. The moderation tools might not be a focus for a while, even if most of us view that as the greatest need
It was worded a harshly but I'm happy to see you jump in here @gabe@literature.cafe <3
To users this might seem like it came out of the blue but instance admins know this is has been a big issue for months. The "roadmap" they shared was indeed, optimize the database queries to make things go brrrr, get more funding and update join-lemmy.org
100%, ultimately there might be disagreements amongst admins over many things but this is something that there is clear unity on and I felt important to establish it. Hell, I've disagreed with lemmy.world's decisions on numerous fronts as well which you already know. I think the harshness is understandable as well, given you know
Forking solves the problem of inactive maintainers, or the problem of maintainers who don't review and/or accept PRs, but Lemmy really doesn't have either of these problems at the moment.
With all forks of maintained projects it starts with saying several times "No, but seriously, you need to do something about this"
Forks are the enemy of open source. The goal is merges. When someone forks a project without plans to merge back, it's a sign that the project has failed them in some way
It's a shame it's not written in a PHP framework or something that's more common. Plenty of devs have been helping about contributing to kbin development, it sounds like it's a lack of manpower on Lemmy's end that's contributing to this
Yeah, Rust was a good technical choice but in practice it really narrows down the pool of potential volunteers
It was an incredibly poor technical choice.
Programming goes through fads where people will claim X can solve every problem. Eventually people realise a languages strengths/weaknesses and communities form.
Rust is the current fad language, its developed a strong following in C/C++ communities but they have nothing to do with middleware (the role Lemmy is using Rust).
It means lemmy devs will have to build everything themselves (instead of focussing on lemmy) and the pool of contributor's will remain small.
Rust is a great and fabulous language, but flexible it is not. If I were starting a Lemmy or Kbin type project from scratch I'd likely start with Python, TypeScript, Lua, or Go depending on what specifically I was worried about bogging me down in the future. And then later on if there were really heavy procedures or db calls that couldn't be simplified anyway else, do those in rust. I think Rust has some very interesting features for micro service development, but for a monolith like Lemmy, it's surely a nightmare
See my goto is Java/Spring Boot or Typescript/TSOA.
I avoid Python because Setuptools/Twine/FastAPI/\ docs conflict and seems to change so creating a good practice project layout is a huge time sink and none of the Python devs I meet seem to understand it.
I am doing GoLang atm, its ok but dev adoption is low where I am and no one has shown me a killer library/framework and being controlled by Google I am waiting for them to get bored and kill it.
Spring Boot takes longer to get going than TSOA/Express but hibernate makes SQL interactions trivial. I love typescript but types makes complex NoSQL queries far more convoluted than Java equivalents (its because Types can't inherit and client libraries don't use interfaces). So TSOA rocks in cases of speed or simplicity.
It is, there are currently discussions of attempting to do so but the issue lies that Rust is not only a really new programming language that really never was well suited for an application like this, forking means nothing if no one is going to contribute to the fork in the first place. I know that pawb.social is working on a fork iirc
Got a link to this AMA? Couldn't find it.
I agree with @Cube6392@beehaw.org, if modtools (one of the reasons for Reddit API protests in the first place) aren't being prioritized, a hard fork of Lemmy will be inevitable. I know the Lemmy devs are known for being strangely hardheaded about certain issues.
They have shifted gears recently and been pretty receptive to this major critique. Things are going in a much better direction now that 2 months have passed. If I can find the AMA I will link you.
OP never delivered. :'(
Yeah, unfortunately took a rapid shift away and my optimism is gone. A hard fork is being made from scratch in a new programming language, that I am actively involved in whatever way I can be.