It's interesting to me that federated search is one of the things that I've been told a lot that federated networks won't be useful without, but we're clearly seeing a lot of backlash here once it has appeared. Maybe opt-in to these things is the right approach, though you're still asking for voluntary complliance.
I wonder if public-by-default is also not what many people want (but since Twitter does it, that's assumed) and followers-only would be preferred?
@maiyannah It would still be federated in a public/subscribe sense but the public timelines would certainly be a lot quieter.
@cwebber What's this about federated search? Who's released it, and what's the backlash? (and who should I be following to have heard about this sooner?)
@sandhawke Well we obviously should have been following each other sooner, wtf how did that not happen
@cwebber We followed each other when I was on quitter.se but then I decided to try the mastodon UX more
@sandhawke that explains it, I remembered subscribing to you
@cwebber I often think people would be much more comfortable with systems where public messages take more manual effort than private messages.
@wilkie It could be that something like a web of trust type setting up of specific relationships is actually what people want (minus the traditional UI problems of a WoT system of course). That's also been proposed as a stronger anti-abuse direction.
Maybe there's a binary of stuff you definitely want public (like blogposts, etc) and stuff you definitely do not (which seems like more like followers-only type stuff)? But I'm not sure.
@cwebber public is simply a bad default. and that is frustrated by specs that define public modes first and punt on private modes.
WoT is confusing to people... reputation systems are difficult to get right. But you can have multiple forms of privacy that I think people can understand. 1:1 private message, group private message, followers-only private post, and public broadcast.
@cwebber so difficult!! private messages need their own inbox, basically, so they can be ignored. maybe ActivityPub needs to say that private messages are completely unspec'd or out-of-scope and that issue is closed with the notion it becomes an extension.
but yeah, this is an example where public is implied default and privacy gets stretched over it ambiguously. seems tough to get right.
@wilkie how about if we make all the user interfaces look like email clients
this will solve everything and has no downsides imagineable
@cwebber oh no
@cwebber do we have a real search now?!!
@cwebber Isn't the point of these types of sites to share? As long as users are given options which are easily accessible, why not aim the defaults to the common user. Who signs up for a site to publish 500 characters of whatever pops into your head...just to talk to yourself?
It's tiring always pandering to users who don't take the time to lookout for themselves. Outside of the internet, you take responsibility of your actions. Why do we have to change that when online?
@cwebber I think there's an important difference between public by default and full text search. It's one of context. I'm fine with people seeing my public posts in the context of my timeline or a public timeline. But when the only context is their keywords, that's what causes problems.
C.f. the discourse around bookmarking sites and tags, which is what hashtags grew out of, or read some of the posts from around the time where tumblr changed between tag search and full text search
@nightpool it's also the case that public posts were already searchable on many gnu social instances which they had federated with, but mastodon users might not have *realized* that too because that same functionality was not exposed on their own.
Don't get me wrong, I still think that maybe adding a flag may be useful; we're even talking about "should we add it to activitypub", but it might need to be an extension so that we don't rush and not understand the consequences? Unclear.
@cwebber does activitypub currently have extension mechanisms? How does it thread the needle between "as broken as XMPP" and "any meaningful addition will break existing implementations"?
@nightpool It has an extension mechanism. ActivityStreams, which is its vocabulary, is json-ld based which is extensible. One purpose of the Social Web Incubator Community Group's goal is to define extensions.
@cwebber I think there's an intent thing here too—theres a difference between something that already existed, and something that (at least appeared to) INTENTIONALLY violate community norms
@nightpool I think https://techn.ical.ist/users/vhf/updates/1028 indicates @vhf probably did not *intend* to violate community norms. (If there was a public expression of community norms before now I didn't see it myself.) It may be that the definition of these norms is happening right now, in real time.
But people definitely had their own expectations, and that at least was violated. Resolving those is norms-building, as well as allowing people to express intent.
@cwebber @vhf there have been pretty clear statements about it before—such as 2 weeks ago, the last time someone made a search engine (it turned out not to be full text, luckily). Apparently there was a very big to do about it around 4 or 5 months ago, but I wasn't around for that one.
Regardless, that's why I said "appeared to intentionally violate", rather then "did intentionally violate"
@cwebber A lot of networks seem to want everything public by default because when you have fewer users it makes more sense to show everything to everyone. As networks grow you see a tendency away from that model.
The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!