[snarfed]I don't know why, but the ActivityPub developer forums and chats I watch have had a _lot_ of discussion recently around JSON-LD legalese and esoterica, abstract taxonomy philosophy, and other architecture astronomy. makes me count my blessings both that I don't have to care and that we don't do that much here, thank god
[dmitshur]It feels embarrassing to say, but the blog on my personal website had nothing one could "subscribe" to... I almost never blog so it didn't matter and I just never got around to it. But someone asked me a few days ago, and so I've fixed that pretty glaring missing piece—my blog has JSON Feed support now. \o/
[schmarty]Forgive my old-man-cloud-yell moment but: I know we're probably nowhere near a high water mark for ActivityPub discussions and I already find them exhausting. I don't particularly care to see it plumbed into every corner and I miss discussions about simpler building blocks.
[tantek][snarfed] when you focus on building user useful stuff, you have less time for that type of architecture astronomy. So I tend to use its strong existence somewhere as an indicator of folks not building user useful stuff participating or dominating conversations
[Ana_R], [marksuth]1, [tw2113]1, [dmitshur]1, [sebsel]1 and [jeremycherfas] joined the channel
srijanI was setting up bridgy fed and faced an issue: looks like bridgy delivers a mastodon reply as a "mention-of" instead of an "in-reply-to" webmention to the post URL. Is this expected? Or could I be missing something?
[tantek]GWG, you have asked about user stories. The diagram here is a very good example of how to use very little text to clearly show a series a powerful user actions combined into an overall story: https://indieweb.org/graphics
[tantek]For Ticketing, consider how would you draw and explain a series of user actions like that, that somehow depend on and make the reasons for Ticketing clear
ZegnatI would attend one but it has not my highest priority. I lack the usecase. I would need to hear about usecases again to be able to rethink which building blocks we actually need.
[tantek]Exactly. Not in my highest priorities as well, hence I was trying to help encourage how folks that are prioritizing it could make progress without those of us who are not prioritizing it
[tantek]E.g. I may propose a microformats dev session especially with Zegnat here to see both how many issues we can resolve (and make spec changes), and using those to improve mf2 change control
[tantek]I would be interested in discussing "small group" / limited audience use cases but tbh I have near zero interest in jumping into plumbing for it (for the moment)
[tantek]I still think there is insufficient documentation (clear, concise, plumbing-free) of use cases, user flows, user stories in that space. So much so that every protocol attempt (yes including AP) has done a poor job at solving those use cases
aaronpkthe user flow i would like to enable is enabling someone to add my feed to their feed reader, which is able to fetch the feed from my site that includes posts that I have allowed that person to see
[tantek]aaronpk that's a good user flow start and also a bit of a throwback to the 2000s feed/reader model. It's only part of a user story though IMO. add someone replying to such a limited audience post, and somehow preserving the audience, or further subsetting, perhaps only to you.
aaronpkif we want to take a large step back, we could reconsider what it even means to reply to a post, like maybe the assumption that a reply is a post on the reply-author's website is not a good assumption in the first place
aaronpkwe could go back to the old-style comment forms where people wrote a comment on the original post website, but try to enable that in feed readers
Loqilocal comments are comments (sometimes called native comments) that are made directly (locally) on a post (like an article) in the UI of the site serving it, instead of posting a reply post on an indieweb site and sending a webmention https://indieweb.org/local_comments
[tantek]aaronpk, yes replying is more complicated, however, Mastodon via AP has demonstrated that limited audience replies, heck back & forths, are possible, *across* servers / instances and perhaps even running different software
[tantek]So we know it is possible. That should be the bar to beat. AP, despite its flaws, has set up a new baseline for limited audience posts & replies across servers and we should be able to do AT LEAST that in indieweb peer to peer use cases
omz13oh, and even if everybody hates the names I used, I will point out that my flow diagrams not only called the things by names, but also by their type: so "Bob" is a "Resource Server" is about as one can be
[tantek]omz13++ yes! Implementations, prototypes, even if they "only" solve a problem for one person, you, are 1000x more useful & valuable than purely theoretical discussions (AKA architecture astronomy)
[tantek]aaronpk, for static sites, I believe we have enough experience from how they implement webmention, Micropub, and IndieAuth (IdP & RP), that there's a "delegate to external service" model there that we can generalize and replicate to other things, such as limited audience viewing & replying
aaronpkthat doesn't work the same way if you want to have the static site actually host the limited-audience content, unless you're ok with the risk of using secret URLs
aaronpkmicropub is transient, the spec is only involved until the post is created and then it's out of the picture. indieauth doesn't host any content. webmention also doesn't host any content itself since it's just a notification. those all translate well to external services
[tantek]Please avoid reasoning of the form "unless you're on with [HUGE ASSUMPTION OF ONLY POSSIBILITY]" because not only does that discourage creativity (to explore possibilities) but it's also almost always wrong
Zegnatthe usecase aaronpk gave now, with the feed, was one of the main things autoauth wanted to solve. But I feel one of the factors that made it hard to iterate on was just how many people needed completely new tech in place to make even a first version work. Changes to the blog post authors side for keeping something secure, something in the feed reader to be able to fetch the content, and something for the person who wants to read to
[manton]Enabled Bridgy’s new Bluesky support to take it out for a spin. For a while, there has been a sort of disconnect in http://Micro.blog’s POSSE-ing because it doesn’t usually include a link back to your blog for short posts that fit in a tweet / Bluesky post / whatever. Maybe it’s time for me to add that as a preference for better Bridgy compatibility? Thoughts?
[manton]Ideally, silo platforms would have some kind of metadata for the canonical URL, but I don’t think Bluesky or anything (except maybe Medium) has that.
[manton]One of my originals goals in http://Micro.blog was for the cross-posting to Twitter (years ago) to feel like native tweets, no extra metadata or links. But I think it’s time for some kind of pref to control that. Really should be up to the user how they want to use it.
[tantek]I'd base the default on whether they are using their own domain or not. Own domain? Append original post permalinks by default with their domain. Otherwise yes I get it, http://micro.blog more like a posting client to the services, a la Hootsuite
aaronpkI think I will switch my syndication URLs to the bsky.app URLs, because while technically less correct, taht's the URL that's actually useful for people to click on
aaronpki like having them so that if someone lands on my post for some reason, they can find the icon that matches an account they have in order to go reply on that silo
aaronpk[manton]: if the original blog includes a link to the bluesky post, bridgy can find it. this is instead of changing the text of the bluesky post itself
[KevinMarks]Someone has just claimed to me that AP can't be parsed without JSON-LD. I suspect that this is a misunderstanding. Does AP put stuff in the @context that you have to make sense of to parse the content in the protocol (eg encryption keys) ? Given that several people here do parse AP, and AFAIK don't use LD to do it
[snarfed]right, that's incorrect. people get very picky about the language to use, I still have no clue what the "right" way to say it is, but LD processing is supported but not required
[snarfed]re needing @context to make sense of some content, that may technically be true if you're very strict and into spec legalese, but in practice I happily ignore @context and do HTTP Sigs and other extensions perfectly fine. afaik same w/[manton]
[manton]I’ve said this before, but I think AP would be better without JSON-LD. I’d go far as to say that LD is actually solving the least important problem with AP. AP doesn’t need to be more extensible, it needs to be better defined and maybe _less_ extensible until we have widespread interoperability.
aaronpkif an AP implementation aliased all their dictionary keys with a @context then a "regular JSON" consumer wouldn't recognize the keys, but a consumer aware of JSON-LD would
[snarfed]one big disappointment I have with AP right now is that the main test suite development push seems to have ratholed so far down into JSON-LD esoterica that I've kinda lost hope of it ever being useful 😢
[KevinMarks]I'd go further and say that imaginary LD extensibility is distracting people from the real possibilities of using more of AS2 within AP, given that AS2 has a lot of fairly well-defined structures based on actual published data in other places
[KevinMarks]The design centre of AS1 was trying to be a union of existing types, and was pushed that way by Gnip and the other stream aggregators, so I can see the sense of the choice. There was some enterprisy stuff wiht Jira and Oracle Beehive too iirc
Loqi[preview] [smallcircles (Humane Tech Now)] @hrefna @KevinMarks @jenniferplusplus I proposed on #SocialHub that going forward we (W3C #SocialCG) *might* consider #ActivityPub to be a JSON-first spec, instead of #LinkedData-first. It should be a change of 'stance' and need not break backwards-c...
Loqi[preview] [smallcircles (Humane Tech Now)] @gugurumbe @hrefna @KevinMarks @jenniferplusplus True. Something we were discussing today on the @fedidevs matrix chatroom.Currently the OWA does not hold in our mostly json-only fedi. Or rather it holds if you do the per-project codebase deep-dive i...
[tantek]Tumblr << ^ “Once my backend counterpart did some initial investigation and scoping, and it became obvious it [ActivityPub] was a complex task and that it would become a cost center without any chance to generate any cashflow, the project was quietly put on long-term hold.”
Loqiok, I added "^ “Once my backend counterpart did some initial investigation and scoping, and it became obvious it [ActivityPub] was a complex task and that it would become a cost center without any chance to generate any cashflow, the project was quietly put on long-term hold.”" to the "See Also" section of /Tumblrhttps://indieweb.org/wiki/index.php?diff=90007&oldid=90006
GWGLooking for someone to develop sharing private stuff with by implementing ticket based proposals and iterating on them... interested parties welcome to apply.
[snarfed] I don't know why, but the ActivityPub developer forums and chats I watch have had a _lot_ of discussion recently around JSON-LD legalese and esoterica, abstract taxonomy philosophy, and other architecture astronomy. makes me count my blessings both that I don't have to care and that we don't do that much here, thank god
[dmitshur] It feels embarrassing to say, but the blog on my personal website had nothing one could "subscribe" to... I almost never blog so it didn't matter and I just never got around to it. But someone asked me a few days ago, and so I've fixed that pretty glaring missing piece—my blog has JSON Feed support now. \o/
[schmarty] dmitshur++ hooray!
srijan I saw https://github.com/snarfed/bridgy-fed/issues/686 , but it might be talking about an additional "mention-of" to the author page. I'm seeing the webmention to the post URL also as a "mention-of".
[tantek] GWG, you have asked about user stories. The diagram here is a very good example of how to use very little text to clearly show a series a powerful user actions combined into an overall story: https://indieweb.org/graphics
rubenwardy what are local comments
Loqi local comments are comments (sometimes called native comments) that are made directly (locally) on a post (like an article) in the UI of the site serving it, instead of posting a reply post on an indieweb site and sending a webmention https://indieweb.org/local_comments
Zegnat the usecase aaronpk gave now, with the feed, was one of the main things autoauth wanted to solve. But I feel one of the factors that made it hard to iterate on was just how many people needed completely new tech in place to make even a first version work. Changes to the blog post authors side for keeping something secure, something in the feed reader to be able to fetch the content, and something for the person who wants to read to
[manton] Enabled Bridgy’s new Bluesky support to take it out for a spin. For a while, there has been a sort of disconnect in http://Micro.blog’s POSSE-ing because it doesn’t usually include a link back to your blog for short posts that fit in a tweet / Bluesky post / whatever. Maybe it’s time for me to add that as a preference for better Bridgy compatibility? Thoughts?
[KevinMarks] are they publicly visible now?
aaronpk tumblr << 2023-10-25 Plans to support ActivityPub are likely put on hold indefinitely https://notes.ghed.in/posts/2023/tumblr-activitypub-integration/
[aciccarello] That's unfortunate, though not all that surprising considering Tumblr's history