@aredridel oh shit what have you seen??
@notspookypip Just seeing instances not federating with it — which means most people can't see you — and very little accountability process around moderation decisions, especially big and vastly affecting ones like that.
I'm somewhat sad that a significant part of the reason for this being considered bad (unless I'm misreading) is that it's popular enough that most people would expect to have a relationship with it. Using such reasons invites centralization by means of incentivizing the use of large entities, because they are "too large to fail^Hbe blocked".
@aredridel
Ah, I see: it's not about the number of users so much as the "typical user" finding this negatively surprising and not bearing any fault. I agree completely and am very sad that instance-level blocks (aka suspensions) are so widely used instead of silences (which do not prevent communication by individuals who explicitly wish to communicate).
@robryk @notspookypip Indeed. And it needs to be stated publicly!
But more clear handling of federated feed moderation would be great; make it obvious that there _is_ curation. (and this ties into my wishing that local-only posting was in Mastodon mainline—that's curating the local timeline)
Aside: do you know what's the reason people use defederation so very often instead of silences? I haven't really asked directly, mostly because the cases I noticed and I was most puzzled by were the ones where my instance was blocked in that way and I didn't want to appear argumentative.
> But more clear handling of federated feed moderation would be great
Do you mean "admins providing more clarity on what they're doing" or something else?
> that's curating the local timeline
You mean curating by causing more content that wouldn't be there otherwise to appear (because people wouldn't want to post it nonlocally)? Otherwise I don't see what curation you refer to.
@robryk @notspookypip Yeah, admins providing more clarity (and just in general I wish there were UI affordances for it. Show what's blocked, from where, in a sensibly integrated manner)
Mastodon still has an 'admin control' model of moderation (because those are easy to make).
One thing that I'm missing in terms of transparency is knowing whether an instance of a user I'm responding to is blocking my instance (I can still see their posts if they are boosted by someone not blocked by their instance, but then I respond into the void).
@robryk @notspookypip And the local timeline, if the instance has local posts, could be so much richer. And with the right UI affordances, could break into channels.
There's a concept of "groups". They are bot actors that do something like boost everything posted at them. This could create a stand-in for the channels you mean, which as an added benefit could be operated so that it's not tied to a particular instance (OTOH it can be easily tied to one instance by making the instance that has all the groups not federate with anyone else).
@robryk @notspookypip Yeah, kinda close, but more a layer on top. I just think the local community is criminally underused, and part of what makes the 'choosing an instance' conversation so strained. They act like communities sometimes, but not often enough
@aredridel
Yes, they're identity providers, moderation deciders, and more. It would be nicer if at least identity provider role was split from everything else.
@robryk @notspookypip Interestingly that part I don't mind so much — since my posts are more my identity than the card that is my profile pic.
@aredridel
Making instances providers of identity makes it harder to move between them and to be recognized as the same identity that's posting on multiple instances. This also makes your current instance's admin into gatekeepers of moving away from their instance (in particular, if they've blocked some other instance, that instance would normally never learn of the move and wouldn't start the whole following-update dance).
@robryk @notspookypip Yeah. I guess I have a strong enough way of being that that doesn't bug me much at all (and also I'd want more trust in the admins than that implies already).
IOW imagine a magical world where you can have the whole shebang without instances. In that world the point of instances would be purely to create communities and it would be natural and expected to be part of multiple instances. It would also be useful to be able to assert that it's the same user posting "on"/"at"/... all those instances they're a part of. (Do you think that last thing wouldn't be useful?)
Now, in our world, where instances also serve the function of identity management everyone has a single instance that can speak for them, so e.g. federation policies of that instance and about that instance affect whom they can speak to. In the ideal world above, a single non-blocked instance someone's a member of would suffice.
Now, you can obviously do something similar by having multiple fediverse accounts (and cross-linking them via posts), but there's no infrastructure for e.g. enabling others to easily follow this group of identities.
Now that I've written this down, I wonder if we shouldn't try to create support for such "cloud of accounts", so that one can (a) follow such a cloud (b) manage such a cloud for oneself.
@robryk @notspookypip No, not really useful most of the time. Consistent relationships across platforms do way more. "I am who I claim to be" is useful mostly in hostile situations where things like impersonation happen a lot. But mostly, it's the inverse proof (which can't exist) that would be helpful: That person is _not_ me.
But impersonation is way harder if people have relationships and community connections.
Which is why communityless systems (which most instanceless systems are trying to be) tend to be where impersonation problems crop up.
It's not inherent that instanceless = communityless, but I still firmly believe that communities should be one of the major units of social networking, not the account. And an individual can be in many communities. And may well keep them separate.
@robryk @notspookypip I find the idea of blocking 1:1 relationships at the group level somewhat abhorrent, without that group being involved. Like none of the relationships impacted were implicated in the problem.