Discussion - Potential piefed instance in the future?
Discussion - Potential piefed instance in the future?
Hello m@teys,
I've been waiting for this a long time (half a year), personally. After seeing the surge in piefed instances, i.e. blahaj, .ca, .zip, quokk.au, and .world creating their own instances, and clients guaranteeing future piefed support, we've been thinking about potentially opening an instance in the future. Not a guarantee, just an idea. This isn't a voting thread, just a discussion. Later on we'll actually vote on this.
Do note that this thread will not guarantee an instance; we are discussing a hypothetical. Suggestions? Ideas? Criticisms? make your voice heard.
Have a good week!
pie picture comes from here
governance type: sense check
You're viewing a single thread.
A comment I came across a while ago puts it quite well:
I haven't used piefed myself but I wouldn't want to switch to it because I feel like it gives more power to downvotes and karma (or "reputation" as they call it):
- Comments with -10 score are collapsed by default.
- People who get downvoted a lot end up with a ‘low reputation’ indicator next to their name. You’ll know it when you see it.
- Upvotes in meme communities do not add to reputation.
Those are three of the twelve points listed as differences to lemmy on their features page.
I also don't agree with some of the points in their article on "PieFed features for growing healthy communities".
from: https://sopuli.xyz/comment/17119490
the comments' thread is also quite interestingPlus it's written in Python and has only 1,6k MAUs, so we can only guess how well it scales
Slightly edited from an older comment of mine:
EDIT: as such I believe that operating a Piefed instance would be a waste resources and actually harmful considering the fragmentation it would cause for little to no benefit at all (also it seems much better set up for a new Reddit with all the .world'ers flooding over there and the anti-features)
I think rimu has made the karma system to be optional (i.e. in the admin settings, not hardcoded) so we'd just disable it anyway. I think the second feature should only apply to mods/admins, since its a very useful tool. #1 #3 i agree, they are not good.
I'll address each point in the Piefed Features for growing healthy communities:
Find people who have low karma
This applies to admins, and while not really that effective, its okay to have.
Find people who downvote too much
This is a very useful tool, since this is a big problem on lemmy. Same thing; only for admins
Spot spam easily
Sounds like a good idea; but with feature #3 as you pointed out, this might cause some new-user shitposters to get falsely flagged, but we would disable it anyway.
Don’t award karma in low-quality communities
This is now opt-in, and we would disable it.
https://chat.piefed.social/#narrow/channel/5-admins/topic/Low.20reputation.20communities/near/1872
Warnings on unusual communities
We would add nothing to communities, anyway; users are free to do whatever they want, themselves. (maybe to our own instance communities like /0 and governance? But not much honestly.
Icons next to comments by low karma accounts
As i said before, probably should only be seen by admins/mods.
Icon for new accounts
No problem here; most clients do this anyway.
Automatically delete content based on phrases in user name
Sounds very useful, and which is why piefed didn't struggle at all with nicole.
Report accounts, not just posts
Good idea.
Instance-wide domain block
We have this on lemmy too
Automatic reporting of 4chan content
Not a good idea imo. We'd disable it.
Plus it's written in Python and has only 1,6k MAUs, so we can only guess how well it scales
We'll always have this lemmy instance anyway :) but afaik it scales fine, and has some optimizations coming along soon.
Genuine question: Lemmy provides a good, solid backbone and API, right? And a lot of 3rd party tools & clients make up for it's "deficits", even in admin tools. So why do we need a big monolith and risk fragmentation, even if the features are "native"?
Clients can only do so much sadly. For example, they can't make polls, user flairs, implement passkeys, following threads, disabling, proper blocking implementations, and etc.
Even then, the features that clients can somewhat implement (like post tags) rely on others using the same client, and aren't considered as a standard.
For the karma features, we'll be sure to disable it if we made our own instance. And as we said; the lemmy instance will always be active and the main instance.
True that, I suppose Piefed has it's advantages with a faster development pace.
Although I'm still sceptical, especially regarding fragmentationIs fragmentation really a concern when both can interoperate?
Skepticism makes sense, though if we were to make an instance we'd definitely start small at first (maybe invite-only?) and then see how it scales, just in case.
Example
- !schedule@lemmings.world used to be the way to schedule posts on Lemmy.
- However, over time the experience wasn't always the best, and quite regularly users would have issue getting their posts out
- Piefed now has an built-in scheduler feature, which makes it easier to manage
Don’t award karma in low-quality communities
Bad idea.
This is now opt-in, and configurable by admins
https://chat.piefed.social/#narrow/channel/5-admins/topic/Low.20reputation.20communities/near/1872
I know :) just addressing the albeit outdated post she linked.
Instagram is written in a much more heavy python framework, but it scaled fine.
I'd be willing to bet that the backend isn't
Its the backend. It uses django which is more resource intensive than flask, which piefed uses.
Huh, TIL. Though some websites say they made the performance intensive stuff like image processing in C++
On python those libaries are usually written in C anyway.
