Social bookmarking is a novel use case for #ActivityPub and I’m super excited about it. I heckin’ love links and lists! I wanna use them for everything.
Things like #Bookwyrm are cool, but it’s not what I want. I just wanna link the thing. Books, films, podcasts, articles, songs.., they’re all just resource recommendations which can be encapsulated by links.
Thanks to @raffomania and @eb for the indirect prompts leading to this article mixing their ideas with my own.
As you may know I started delightful project of curated lists. Though decentralized via sub-lists with independent maintainers, its still burdensome to keep lists up-to-date.
Getting linked on the mastodon network (tm). A few notes.
A while back, somebody linked to the honk manual. Within one minute, 76 mastodons retrieved the link.
I replied some time later. Within one minute, an additional 58 mastodons fetched the link.
For the day, it amounted to 150 hits, some arriving a bit later. I think not so bad, since it takes me about 200µs to serve this file from cache. It would be more problematic for a larger or more expensive file, however.
There's been some discussion of having the origin provide the preview, and the consequences of fake previews, but I think a reasonable compromise would be to have the origin check if there's a reasonable preview to be had. If the target is not a web page, tune the link so recipients don't try to generate previews.
Also, over the day, 152 hits from Pleroma and 142 from Akkoma, although those numbers are double since every GET is preceded by a HEAD. I'm not sure what happens after the head request, but that could be another way to avoid inflicting unnecessary load.
Challenge for a UX designer: rethink the UI/UX around content warnings on the Fediverse (mastodon, pixelfed, etc) to allow capturing not just a text field, but also some predefined labels (e.g., 18+, sexual imagery, violent imagery, etc)
Technically these labels would come from "providers" which define what they mean and controls that should be implemented.
I'd like to give a heads-up to #SocialCoding movement where I'm involved and plan to revive the notion of #SX or Social experience design, esp. for the #SocialWeb.
So much chatter on #Fediverse is either deeply technical or app-specific. The field of SX may offer a more holistic and inclusive perspective, focus on methods, guidance and best-practices.
It's just a sketch in a different context (a Fediverse of fine-grained components & services instead of big siloed apps. #AppFreeComputing). It only depicts how a #Microblogging timeline might be richer.
One aspect of #SX may be to show what is possible in federated #UX, in order to inspire and envision a future #Peopleverse.
Mirlo is an artist-centered, collectively owned music distribution platform. On Mirlo, you can directly support musicians by purchasing their albums or providing monthly contributions.
I think either I or someone else had this idea before, maybe someone with more spoons than me can build it: a federated Thing directory, powered by #ActivityPub.
Some aspects of such an app:
No "content" would be hosted there. Instead it would point to where it should be, which could include federated platforms like Peertube, in which case it would "boost" instead of "link."
It wouldn't be built directly compatible with Mastodon (i.e. it wouldn't use Notes for published links). It might just be coincidental.
It probably wouldn't use WebFinger because there's not much of a point. (I guess it could use it to refer to users, but that sounds out of scope, and potentially confusing as link collections would probably exist too)
Each directory (instance) would search itself by default, and it could probably use a minified index to determine when it should try searching another directory, which would be done on that directory.
Metadata included, but not passed around.
Directories (and/or collections?) would have their own set of key-words which other directories would use to know when to search that directory. There would be a hard cap for how many are stored for each directory, although maybe it could be adjustable.
The intent is to make things like Fediverse instances and media streams discoverable in a decentralized way. It's not built for things like article sharing (like Lemmy) or bookmarks (like Postmarks), and that use should be discouraged.
Each entry would include a key or a hash to verify or decrypt the content with. Hashes aren't needed if the entry points to something content-addressable, i.e. DIDs or IPFS. Also they'd probably be "highly recommended," maybe penalized if it's not included or something, instead of outright required, since media streams likely won't be signed nor will they be hashable by its nature as perpetually changing.
Every time I post link aggregators, I wish the Fedi clients could decompose them. So you could more easily go from muh link to yer app without another website as a single point of failure in between the actual destination.
Hmm, the #Mastodon search works differently than I expected. A search for "fitness in:library" gives matches on the words "fit" and "first", so I get a ton of results.
For #ActivityPub the question of "Why use #LinkedData?" has never been answered. There should be clear merits to wade through all the complexity that this choice brings, right?
Yes, its ultra flexible, and you can define your own semantic #ontologies, and theoretically it could provide a robust extension mechanism to AP protocol. Except that right now it doesn't.
What's the vision of a Linked Data #Fediverse? What great innovative #SocialNetworking#UX would it bring, that makes it worthwhile?
I think more general than your case would be connecting various #OpenScience tools to the #Fediverse, related to open publishing. A field where there's lotta interest. CC @jonny
I'm interested in hearing from #ActivityPub developers who've successfully mapped #OpenGraph properties from <meta> tags in Web pages onto the Page type and its properties in AS2.
Just read about your intent to elaborate a bit on some of the #FediverseIdea issues. Would be wonderful, and most welcome.
The repository is under the #Fediverse org on Codeberg, where also the #FEP can be found, and fedi delightful lists I maintain (published to https://delightful.club ).
This is pretty terrifying and I can't believe I'm just hearing about it. Maybe something to do with not being on X/twitter and filtering out almost everything related to him.
Been thinking about how the #Fediverse and #ActivityPub could finally change online dating for the public benefit.
cc @evan
Frankly, most likely the ActivityPub integration is best for in-game communication. And for other interaction different network protocol can be used. Think e.g. Peertube for instance, that uses WebTorrent for videos and AP for comments, likes, following, etc.
But in theory a game-oriented AP extension might be created. Here typically inspiring folks in the ecosystem may help.
Sure, we have hashtags, but they rather work to put the posts into a pool of others also bearing that hashtag. I'm thinking of drawing a line between specified posts to say they're related.
There's a place where you can drop off these dreams so they may inspire others. It is part of #Fediverse repositories developer community where also the Fediverse Enhancement Proposals are kept. File a #FediverseIdea on the issue tracker of https://codeberg.org/fediverse/fediverse-ideas
I posed a #FediverseIdea the other day, about modeling the concept of "Community" in an #ActivityPub extension.
But really first of all there might be a #FEP on modeling relationships between actors in general (i.e. the diagram in the issue should connect to Actor, not to Group).
As you say there are different roles, and they aren't limited to admin, mod, guest, etc. which relate to very specific use cases (a subset of social media applications).
What if we had an #ActivityPub federated app to take care of the social aspects of serving packages from self-hosted registries? And used a mechanism like #WebTorrent which #Peertube also uses, to distribute the packages?
I filed a #FediverseIdea named "Community has no Boundary".
The idea is to define a standardized #ActivityPub vocabulary extension for "Community", based on as:Group where the members (as:Actor) have a collection of as:Relationship to other actors.