Say we have two different instances, none of which are blocked by or blocking other instances. Theoretically, shouldn’t their results be the same on the “All” page using the same sorting mechanic, e.g., Hot.
But apparently this isn’t so when I tested it on instances with no blocking or blocked by criteria. Can anyone explain to me why? I’m still relatively new to Lemmy, and I am genuinely curious.
PS: I’m using the “recommended” instances from this git: https://github.com/maltfield/awesome-lemmy-instances
Instance A doesn’t know about a community on instance B until somebody on instance A searches for it and (I think) subscribes to it. After that, the entire instance A can see the community on instance B.
So, with small instances and new communities there’s a delay until and unless somebody does a search for it and starts interacting. And that means that the ‘All’ on difference instances, especially smaller instances, is going to be different. As more and more users join various instances and search for communities and interact with communities in the rest of the fediverse these will be more synced.
Yeah, that behaviour is very weird. My decision to search for and follow a community shouldn’t affect the “all” of another person on my instance.
It’s an artifact of how the software works right now. Of course, this could be changed in the future in any number of ways. Things are likely to develop rapidly given the influx of users. Many of these folks are developers and can and will add and change things to make it work better. That’s the great thing about open source.
its a scaling thing and it makes sense in a lot of ways, discoverability however needs to be better. Inevitable instances will want curated default views while users need more power to see everything or block what they dont like easily.
Federation doesn’t happen immediately, the syncs could be happening at a little bit different times across the different instances. Not enough to make any major differences in the content you’re seeing or interacting with, but probably enough to make the 1:1 comparison not line up.
Each instance generates it based off of the subscriptions and the synced content that exists locally on the server. I think anyway…
