[update: never mind, @dansup
Has gone and done it: https://mastodon.social/@dansup/113496597165906876]

I’m pondering how one would model starter packs on fedi.

I’m imagining an app you can log into with your fedi account that lets you a) create one or more lists of usernames with a theme b) browse all lists and follow lists made by others.

Would fedi culture suggest that accounts would want to opt-in before they can be added to a list, or would an opt-out be enough?

@janl Opt in plz.

@janl Some Mastodon people are thinking about it:

https://oisaur.com/@renchap/113443054222350463

@janl I think @dansup has built something for Loops which is intended for wider use. But I don’t know much about it. I would be curious myself.

@janl I recently had a discussion about this, and a very simple approach would be to allow making lists (the same lists used to manage your follows) public, optionally allowing a one-click import.

The question about opt-in vs opt-out is a bit hairy because the same mechanism could be used for shared blocklists, for example, and you obviously don't want people to be able to “opt-out” of being in a blocklist, neither to be able to refuse being added to one.

It looks like as usual, something like this already exists, it’s just not accessible to newcomers: https://mast.solarisfire.com/@solarisfire/113492440324722556

Back to the drawing board.

@janl that has lists sure but they are
not integrated into clients for simple usage. And starter packs are special because people can make them themselves easily. The challenge isn't "building lists" it is creating a safe but useful way for people to offer curated lists of people for others to follow.

@tante right, that’s the important point. These things need to be integrated at the first point of contact (an app, or the main server web interface).

@janl yeah. I keep being a bit annoyed that Mastodon has no plugin/API that allows to extend it without having to wade through byzantine Ruby cruft. Like this could be a great plugin thing that admins could add of they wanted

@tante yeah, the more we talk about this, the more a more pluggable architecture comes to mind. — For the longest time I wanted to be able to replace the main UI with, say, Elk, and only show the original UI when requested specifically

@janl It's one of the reasons why Mastodon isn't great for trying out things. You always have to fork the whole codebase and mess around with it. It's a big hindrance.
I am not sure if Mastodon's ever gonna get that though (and I am not experienced enough in Ruby to do it quickly and have absolutely zero desire to learn another dead language)

@tante yeah same. At the same time building an API compatible yet production ready drop-in replacement in something else is a big task that I’m not sure is worth it.

@janl I keep looking for alternative implementations (takahe for example) but they all lack resources to get there currently. So basically one would need a significant amount of money to fund it I guess. I don't see it happening in people's free time

@tante I’m also envisioning a more modular thing. One thing that will be needed is a reimplementation of the Mastodon API, but we should do this only once and then be able to compose it with whatever else we wanna iterate on.

@janl males sense. A very slim core with most features already implemented as "plugins" to ensure the API/architecture works and then iterate.

@tante yeah sounds like a perfect weekend project for me that I will never finish after proving the most basic bits.

Out of curiosity, what language would you’d like to see this in?

@janl personally I'd use PHP or Python

But that's cause I am most familiar with those ecosystems

@tante fair, it’d be JS for me and that’s then already part of the problem…

@janl your approach might get more support/traction. ;)

@janl I am kinda thinking about expanding @Edent's ActivityPub bot in one PHP file to mess around a bit

@tante sounds like a plan!

@janl don't expect anything in the next days/weeks. A lot of other work I committed my free time to :(

@tante @janl I've played around with it a bit to work for my quite specific use case and it showed to be a great starting point (https://github.com/kore/nordmann.name)

@tante @janl
Please do!

@janl @tante Have you looked into @bonfire? They're building a framework to more easily modularize a vaguely Mastodon-like platform for different purposes. Might be your best bet if you want to collaborate on something that already tangibly exists.

Re: starter packs, keep an eye on @dansup's spec discussion: https://github.com/pixelfed/starter-kits/issues/1 I still need to write a reply there today, I think I have a concept that we can make work without immediate client modifications.

@julian @tante @bonfire @dansup nice, this looks very promising. Sadly elixir :)

@julian @janl looks interesting but I'm not gonna pick up some esoteric language for that experiment. If I put effort into it I want more than 5 people to be able to run and work on it

@janl https://mastodon.social/@dansup/113495058111056864

@mrtoto @dansup hah timing

@janl The only thing even close to starter packs that I know of is https://fedi.directory/ (run by @FediFollows ), and I'm pretty sure they contact each person on the list for consent before including them (my wife is included on the Crochet pack). Would be cool to mass follow a group on there, but I also like just being able to browse through and pick individuals to follow.

@solarisfire nice thanks!

@janl fedi culture would probably go for only opt-in.

I'd argue that (for public accounts) opt-out (of certain lists or in general) should be sufficient.

@janl@narrativ.es It is already kind of possible to create starter packs if using correctly formatted CSVs with Mastodon's "Import followers" feature. I recall @stavvers@masto.ai actually tried to onboard some people this way a while back, so she might have some insights both on the UX of the process and how it was received from a "fediverse norms" POV.

@brecht @stavvers yup I know the primitives are there for this. You can also do it via the API, that’s how the find-your-twitter-friends things worked. This isn’t a technical question.

@janl is https://communitywiki.org/trunk still a thing? That’s part of how I got started — I thought of a few interests and then randomly followed folks.

Give that a way better UI and you basically have a starter pack.

@janl is it "how do I find my friends" or is it more "who could I follow who is sharing interest y"? I feel one would profit from lists of people around a social graph (I look through the followees of somebody I know to find the rest) or is it about people of topic specific instances or people who hashtagged topic Y? (I scroll through the hashtag Y or the public timeline of topic-Y instance) - then I profit from easier accessible lists. (either way I'm in, I actually still don't like mastodon 😂)

@sushee this is “I’m new here how do I get a timeline”

@janl hmm how about the thing a bunch of platforms did back in the day with choosing what interests you based on hashtags - you can follow a hashtag technically - and you have a nice set of your choosing where you can later follow people (or never)?

@sushee I don’t see hashtags being used a lot here outside of a few specific ones. I like the curation idea of putting folks together who know about a thing.

@sushee I mean, I’d follow #DistSys or whatever but I’d miss all your great posts on the topic ;)

@janl I feel like the concept is diametrically opposed to fedi idealism, and it’s why Bluesky will continue to grow at a much faster rate until it implodes (tho that might take another 15 years)

@benschwarz this sums up my motivation: what can we do now so that we can harbour bsky folks when that happens.

@janl gonna write something about pile on driven dev as if the fedi does grow it will be because things are built despite the most "passionate" users.

This does seem like a case where asking users to confirm being added provides benefit with little downside though.

@dale initially I was thinking opt-in to be listable, but now opt-in to be on a specific list is a lot more sensible, *except* now you have list invite spam.

@janl a very paranoid part of me does not like being on any kind of list although I find them convenient over there on bluesky

@JohannaMakesGames this is me right there ;D

@janl @JohannaMakesGames is there maybe an element of public Vs private space here to some extent.

For me at least, this is my instance of Mastodon and I control it - also who sees it. After 17 years on Twitter, I'm happy to have this control.

But BS is a more public space - I've seen others say they'll create accounts there, but act differently 🤷🏼‍♀️ (I know I'm tired of all these personas online and in quite happy with the quietness of this place)

@tanepiper that’s a good observation in the sense that fedi is a lot of things to a lot of people and what’s on my mind now is what do we have to do here to provide a home for bsky folks when that implodes.

A valid stance is “we do nothing”, but I dont think that’s what I want.

Part of me is frustrated by the “we don’t do innovative things because of safety” without actually putting any technical barriers in, so adversaries can already do all the abuse.

@janl that's an absolutely valid stance to take on it - Fedi was not ready for the Twitter implosion and has still left a bad taste for people.

Yes the safety aspect is trotted out a bit too much to counter innovation - it's a fine balance always

@tanepiper yeah, I absolutely do not want to downplay safety, but if we do trot it out, let’s make it count.

Bearing in mind that any adversaries will have already built this privately for themselves and there are no special protections aside from what’s available already (approve followers, post visibility, search toggle in account).

@janl this. Consent doesn’t work if it’s optional for bad actors. The only viable protection in fediverse-terms is when people gate the following behind approval. Controlling someone else’s lists doesn’t work if following the profile is completely public anyway.

@freddy yeah, but we also can’t run roughshod over expectations. I learned that the hard way with Searchtodon (https://searchtodon.social/Adventures-in-Mastoland.html).

I strongly believe that we have to back those expectations with safe mechanics, or disabusenfolks of the notion that something can be kept culturally safe at scale.

(Then don’t scale is a valid option, but I don’t think it’s viable any more even now)

@janl Right, I almost forgot about your search experiment 🤦. You got a fair point. So sad, but so true. - Writing out loud (without too much thinking): Do you think it would be OK if anyone can list anyone else but if the listed person blocks the list author (or the list?), then it stops working? -- I suppose blocking is also a mechanism that doesn't really work in terms of "threat models" if the blocked can still read when using public anonymous access and that seems widely accepted?

@freddy yeah, plus the “let me invite you to 1000 lists” spam

@janl

I agree and always expect opt-in. If your excuse becomes, it's ok because bad actors won't do this, you are just a bad actor in my eyes and I'll call it out and block and whatever else I can do to discourage it.

All we have that makes Mastodon better is the culture so if we are finding excuses to undermine that, or adopt a capitalist (eg scale is good) consciously or unconsciously, things will slide.

I expect that to happen eventually because that mindset lives in most of us.

@freddy

@markhughes thanks for the threat.

@markhughes my (sub-) point here is that there is nothing technical that can be done for various things that the culture keeps promising. I don’t particularly care which way this goes, but it needs fixing.

@janl
I guess you were offended by me sharing my views on this issue and saying how I deal with it, but I didn't personalise it.

I agree that technology can't solve social issues. It can though help us do so.

Sharing how we feel about things and what we currently do about them is part of how we solve these issues, and that in turn influences the tools we build. You might for example have had a response about this that I hadn't considered. 🤷‍♂️