cendawanita

joined 2 years ago
[–] [email protected] 4 points 2 years ago

@be_excellent_to_each_other in principle as that's the point of federation, it's value-neutral to participate in such comms.

But say you're concerned about downtimes or culture or you're anticipating that instance will not be reachable or even it increases traffic load, then maybe you'll find it sensible to support a local community or magazine

[–] [email protected] 2 points 2 years ago (8 children)

@crossmr can you confirm that he's posting direct to microblog or is the microblog slurping his content because it matches a tracked tag? If the first it's a bug, if the second it's a built-in flaw.

@NOOBMASTER

[–] [email protected] 1 points 2 years ago* (last edited 2 years ago) (8 children)

@NotTheOnlyGamer it's still worth worrying. Like, in the examples i gave in the comments, i can't login to kbin.social with my other account, and i can't delete my posts. (ETA: but the larger problem is the flood of traffic and kbin hasn't had time to sort out proper instance blocks yet. Spam is already an issue)

[–] [email protected] 6 points 2 years ago* (last edited 2 years ago)

@LollerCorleone it's not online yet, but Chris Trottier is developing spacehost with that purpose iirc.

@haubles @vvuksan @renchap @piotrsikora @ernest @Nougat @bourbonmakesitbetter

[–] [email protected] 1 points 2 years ago* (last edited 2 years ago)

@asjmcguire i was talking about kbin. The AP programming language (eta: yes it's called protocol or standard as well iirc) is definitely old :)

The fact that kbin does fetch posts from originating instances without instances having the setting to allow or disallow is a problem once something with the computing power of meta and its userbase comes on board. Because mastodon for example doesn't have whitelists, only blocklists.

[–] [email protected] 3 points 2 years ago

And just to provide an example, copying straight from my comment here https://mefi.social/@cendawanita/110585975153683699:

Yup that's happening rn. It really got driven home for me when my kbin account gave me a comment alert... For this account. It went to the correct person because the usernick is the same. Also the comment is to a post that is uh untagged 🙃 https://kbin.social/m/random/p/498351/I-m-thinking-once-there-s-a-protocol-i-really-like-just

[–] [email protected] 0 points 2 years ago (7 children)

@imaginelizard wah lau buka sembang right at midnight ka

[–] [email protected] 1 points 2 years ago

@neonfire scheduled posting would also be good, to stagger the cross-posting imo

[–] [email protected] 1 points 2 years ago

@LegendofDragoon do you mean to ask how to follow a whole instance or specific accounts in that instance? For the second, you just search the same way (@ user @ instance) or click their names from the posts that end up on your microblog and add from their profile. For the whole instance, the kbin functionality doesn't have it yet (and the typical fediverse clients i know are optimized for pure microblogging protocols for now, or else I'd recommend client-side solutions)

[–] [email protected] 2 points 2 years ago (2 children)

@Otome-chan@kbin.social oh yeah, definitely some kind of copy primacy/conflict issue for sure. I have no idea how to describe it well enough for the codeberg ticket tho - definitely worth raising.

[–] [email protected] 4 points 2 years ago (5 children)

@Otome-chan@kbin.social yeah, that's what needs to be tested:

  1. Is the originating instance the true copy? Or is every copy of equal status? Kbin's handling of having local copies is probably the issue here because other fedi protocols don't choose to do it this way (and it's probably because this reduces the load from all the fetching; big fedi accounts have been known to induce DDOS-like results when a post gets popular)

  2. If that can be confirmed/clarified, is the mod log a universal log only but specific instance changes not propagating? I think Q1 is going to be the pickle here - because it implies a lot in terms of coordinating copies across the kbin infra at least (not sure with Lemmy I've not poked around at all). Copy conflict is going to be inevitable at this point....

[–] [email protected] 1 points 2 years ago* (last edited 2 years ago) (1 children)

@Otome-chan@kbin.social what's probably happening is that the mod action, like edits to a post do get federated along as well. The only time you won't see the changes propagate is if a federated mag is on an instance that gets defed, then no future changes will occur. Not sure if you've had a fedi account before this but if you do, it's like when you boost a post and the post owner edited it and you get a notification that your boosted post has been edited and you see the new text. So the kbin copy (this is where the threadiverse differ, in having local copies. One reason why 'copy url' has two versions, the instance copy and the original copy) may have the instance owner as the delegated owner of the mag but the original mod actions can still travel. So the mod is still the mod of the magazine in all federated copies of their community/mag.

ETA: I'm letting my comment stand but just noting, i didn't answer your question at all, i think. OTL

@Shortcake

view more: ‹ prev next ›