@jalcine I'm planning on improving it over time, right now if you boost a post, instances that don't know about the server won't see it, but that's all coming soon as I learn more about AP
@jalcine also incrementally improving how the content layout is digested for Mastodon (just this morning I converted <li> tags to ascii bullets)
I can agree regarding targeted mediums doing targeted things.
But if it's something that's like doing conversion work like the ATOM/RSS to AP, shouldn't it conform more to the spec then? If not, maybe changing https://bots.tinysubversions.com/ to say "Mastodon-compliant" vs saying AP.
@nightpool I don't know what you mean by contentMap. (and why is it a bad idea to serve a subset of content if I detect a particular server on the other end?)
@darius like, broadly, having multiple content types/representation in a single Object
why is it a bad idea to care about software versions, rather then feature support? because that way lies Mozilla/5.0 (Linux; Android 8.1.0;) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.100 Mobile Safari/537.36
@nightpool having multiple content types in an object is a great idea. So like I'd provide an object of type: "Note" for maybe the plaintext and then type: "Article" for the full html? Do any popular AP clients look at that kind of thing and make decisions on what to use?
@puckipedia @nightpool Replying here since Mastodon doesn't seem to be rendering it, well, anywhere. I think my main thing is I want my tool to be useful to the users of today's most popular AP client/server.
Re: figuring out if an instance is Mastodon, I could always ping this:
and if it looks like a Mastodon-ish response, I assume it's more or less Mastodon. Not perfect but will mostly catch em.
@darius @nightpool Many pieces on software that do support articles (Pleroma, Kroeg) have compatibility with Mastodon in that way, and will return an equally valid response. Don't do UA Sniffing. also, you don't *want* to fit an entire article's worth of data in a small column, imo. BTW, everything worked as expected, it's just that Mastodon's code doesn't render articles inline. which is good. because it doesn't have support for rich text.
@puckipedia @nightpool so basically what everyone is telling me over and over is that I need to stop trying to chase a dream where Mastodon could eventually function as a generic feed reader. Which sucks because I am really enjoying using it as such in my own sandbox (see pic) but I'm running up against its limitations. (For longform I have a branch that automatically CWs anything over X chars as well.)
Anyway I guess I am gonna have to build my own thing sooner than I wanted to...
right now the way reading Weird content in mastodon works is that *mastodon* will try to convert it into a Note (in this case, that would work pretty well for RSS—it would show title and link).
i think it's also going to be useful if mastodon let the *sender* specify a "fallback" format, but right now that doesn't exist (and it's maybe hard to see what AS property we should use for that, we might want to invent one)
ultimately—i can't read anyone's mind. Maybe mastodon could be a good feed reader. so far https://bots.tinysubversions.com/u/nightpool has been pretty disappointing, although i think part of that is tumblr's lackluster RSS feeds. I don't really think feed readers are what people want out of it, but maybe they are. people seem to be annoyed enough at cross-posters that it doesn't seem obvious that "add more types of even less interactive cross-posters" is going to have a good reception.
@darius @puckipedia i don't think the reason people have problems with crossposters is because they think non-mastodon content is wrong. i think people have problems with crossposters because they're non-interactive—people turn them on and then go back to twitter.
they're not posts made by "real users", they're posts that have nothing to do with this network, reflections of people who half the time don't even know they're being cross-posted.
@nightpool @darius @puckipedia I see “discussion venue” and “feed reader” as completely different functions that just happen to overlap in many soc platforms. Most are really awful for discussion (probs w threading, notifications, local user presence) but pretty good as feeds. So IMO the challenge that needs attention is how to make a really good discussion venue that federates with (and draws from) the multitude of feed platforms out there already.