I thought Mastodon's quote post controls were extensive...these are "boundaries" that can be set in Bonfire for your posts.
General Discussion
2
Post
2
Autori
0
Visualizzazioni
-
I thought Mastodon's quote post controls were extensive...these are "boundaries" that can be set in Bonfire for your posts. Not to mention the roll your own "Custom Boundaries". Mix and Match!
I question how these boundaries federate, are they honored across platforms?
GoToSocial has some reply controls that basically just drop replies you don't want. So other platforms may not honor the request and still send them, but they are just ignored. I'm sure similar here.
-
I thought Mastodon's quote post controls were extensive...these are "boundaries" that can be set in Bonfire for your posts. Not to mention the roll your own "Custom Boundaries". Mix and Match!
I question how these boundaries federate, are they honored across platforms?
GoToSocial has some reply controls that basically just drop replies you don't want. So other platforms may not honor the request and still send them, but they are just ignored. I'm sure similar here.
@box464 we started drafting a section on "how do boundaries federate" in this document: https://docs.bonfirenetworks.org/boundaries.html
The short story is we interoperate with Mastodon's newly introduced canQuote and and GoToSocial's canReply/canLike/canAnnounce for general boundaries like only follows can reply or everyone can quote, and for everything else it's enforcement at the level of your instance (e.g. "basically just drop replies you don't want"). Feedback on that and on the docs as well is more than welcome!
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@jackwilliambell @silverpill or as the indieweb people will argue, you should just embed this information in the html which you are likely already publishing, because sidecar formats are bad (single source of truth).
-
@jackwilliambell @silverpill not entirely as it should be imo -- the problem is equivocation. people using the same terms but with different meanings. the expectation is that you agree ahead-of-time; the reality is that no one actually checks this. it's not about "one protocol to rule them all" but the protocol is undefined in many places and undefined behavior is bad. or people use their own interpretations not exactly matching the official definitions, because to them it's "close enough".
-
@jackwilliambell @silverpill i want to believe that good docs and clean design reign supreme, but the reality is that worse is often better. there are so many technologies that are widely used because their value is almost entirely in what consumes them. atom vs rss is the clearest example of this -- atom is far superior, but rss will live on because of iTunes-style podcast indexers. who consumes jsonfeed? not as many as atom or rss. but at least you can publish multiple formats.
-
> "fedi is ruled more by consensus and alignment than by spec."
Which is as it should be. Trying to tie everything down to a 'single protocol to rule them all' is both an impossible goal and a barrier to creativity.
But I am concerned that (a) Mastodon uses ActivityPub in a way not intended by the protocol designers and (b) is becoming the standard you have to follow – when the protocol itself isn't being modified to be more 'Mastodon friendly'; for lack of a better term.
-
@jackwilliambell @silverpill idk if activitypub has "failed", but if you want to talk to anyone, you are better suited going to their docs instead of the actual spec. or their codebase.
in practice, the biggest flavours are mastodon and lemmy. there's some overlap, and some other flavours that are more niche (funkwhale). fedi is ruled more by consensus and alignment than by spec.
-
I'd argue what drives adoption is ease of use first; even fitness to a use case comes second to good docs, a clean design, and a comprehensible reference implementation.
JSONfeed seems to be 'Atom in JSON'. Which works for me and the microblogging use cases I am envisioning.
-
@jackwilliambell @silverpill i think the "hard to grok" is often misattributed to json-ld and rdf, but those are entirely optional, so that's not the real problem. the real problem is the poor data modeling for anything that isn't an activity. you already identified things like Hashtag not making sense; the lack of sense-making extends to other types like Mention and Link. there's other ambiguities like Collection, which can be both paged and unpaged at the same time. it can be quite messy!
-
That's more background information I didn't have. And I've done quite a bit of research into ActivityPub, before I gave up trying to figure out a way to untangle all the ways extant implementations were using it that didn't match the docs.
So, what about the non-Mastodon (and Mastodon workalike) Fedi implementations? It seems to me they all have some weirdness to them, but have we got to the point Mastodon-compatibility is required to use the Fedi? Has ActivityPub failed?
