#dev 2022-10-31

2022-10-31 UTC
#
[tantek] do you have the domain / internet archive links for "that association"? perhaps we can document on /site-deaths
#
[tantek] [denschub] re: "have had to implement “whatever Mastodon does”, because AP leaves out a lot of details, and you kinda have to reverse-engineer existing implementations to end up with something that may be compatible" yes, that's exactly what happens to any semi-successful standard without a working & maintained test suite
#
[tantek] testsuites++
#
Loqi testsuites has 1 karma over the last year
#
[tantek] [denschub] it's probably the biggest weakness of AP frankly, the lack of a well maintained test suite with feature coverage
#
barnaby yeah everything involving private messaging or private content in general is a whole other can of worms
#
[denschub] > do you have the domain / internet archive links for “that association”? perhaps we can document on /site-deaths
#
[denschub] https://web.archive.org/web/20210420070552/https://feneas.org/activitypub-well-its-not-that-bad/ is their response-post. sadly includes less links than I remembered - sorry it’s been three years. :D https://github.com/friendica/friendica/issues?q=is%3Aissue+label%3AFederation+ is probably a better source of “concrete issues”
#
[tantek] yes, private messaging is still *hard* and in that regard it's been useful to see AP's latest attempt at it and seeing where the UX edges are still rough (possibly too rough)
#
[denschub] also, I feel like I have to make clear that I have no grudge against the folks in the SocialWG. I’ve had quite productive chats with them. They had a kinda-impossible task, and it shows, but that’s not their fault. they actually did amazing work.
#
[tantek] [denschub] you realized that many SocialWG participants are still here in #indieweb chat right?
#
[denschub] yes — which is why I said that.
#
[tantek] the AP work was only one small piece of SocialWG
#
[denschub] after the AP draft was published, I tried to get some traction into an effort to build a more-strict specification set on top of AP, like a “required set for interoperable implementations”. unfortuantely, at that time, the folks I talked to made it clear to me that they had no interest in pursuing that, as “we implement whatever we want and everyone else will maintain compatibility”
#
[tantek] when AP graduated from SocialWG, it at least had *a* test suite, even if it wasn't as thorough as we would have liked. problem is even that went offline 😕
#
[denschub] I think cwebber’s post https://dustycloud.org/blog/on-standards-divisions-collaboration/ is still golden :)
#
[tantek] not really tbh. it completely misframed it as a vs b, or a vs b vs c at best, when the reality was there were 15+ different camps and we somehow narrowed that *down* to 2-3.
#
[tantek] somewhere I have an unfinished response blog post to that, because there's a lot that is inaccurate there, and cwebber knows better because he was there and witnessed all the insanity
#
[denschub] but having even more different camps only makes parts of his points stronger
#
[denschub] but yeah, I understand
#
[tantek] he focuses on the differences at the end, which tells a very poor version of the story, which is that we actually got many many disparate things to converge
#
[tantek] ugh apologies, delete that "he" 😞
#
[denschub] oh. thanks for making me aware of that
#
[tantek] really there's a lot more research about all the background that went into all the specs in https://www.w3.org/wiki/Socialwg
#
Loqi [Social Web Working Group] Sixth Social Web F2F meeting
#
[tantek] I'm digging through to see if I can find the longer lists of "15"+ projects / proposals
#
[denschub] I think a big problem that I’m having is that I default to a very defensive position when it comes to those things. Primarily because diaspora and its team members frequently get attacked for having a “not-invented-here syndrome” or whatever every time we say we have no current interest in implementing AP.
#
[denschub] the sad thing is that the folks who are attacking other projects for not conforming to “the standard”™ are neither involved in any spec discussion, or any implementation of any spec. they’re just random bystanders who think they know the space better than the folks working in it, and that gets… tiring.
#
[denschub] I probably sound a lot more negative towards AP than I actually am because of that.
#
[tantek] I have seen exactly what you are talking about and it makes me sad
#
[tantek] the attacking that is
#
[tantek] and you're right, typically it's not people who were actually involved with any of the specs (or even implementations often) 😞
#
[tantek] ironically I see those attacks as coming from a "defensive" attitude by folks who have some idea of AP's weaknesses (including lack of a test suite, minimal or otherwise) and the UX awkwardnesses you speak of
#
[tantek] Every person I have spoken with who has ever worked directly on Diaspora has been very kind and a good listener. Even if we disagreed on things, it was always a very polite disagreement, and we often left with the understanding that we would still continue to consider other possibilities
#
[tantek] It was quite impactful really, my interactions with the original Diaspora four taught me a lot about the power of positive interactions in community, they were in stark contrast to so many other OSS / open standards "communities"
#
barnaby [continuing inbox/outbox model and private messages discussion from meta because it fits better here]: yeah, I can see the inbox/outbox method working well specifically for private message sending, but that’s not really a problem I’m interested in solving with social networking. hence my preference for h-entries at URLs connected with lightweight webmention pings
#
barnaby …at least, for (micro)blogging
#
[tantek] barnaby, maybe. that was true for me at the start of IndieWeb as well, though we always intended to "solve public first, then expand to solving private/limited-access use-cases"
#
[tantek] now I think there may be a stronger use-case need to solve "limited audience posts" to encourage smaller, more trustful interactions, rather than performative extremism that "public" social media seems to encourage & amplify
#
[denschub] that might be controversial, but: is it, actually? I’ve noticed a strong trend in people using more Chat-like applications (whatsapp, telegram, signal, matrix, …) for private things, and they exclusively use social media for “public” things
#
[denschub] I think if I’d start a project from fresh these days, I would exclusively focus on distributing and discovering public content
#
barnaby [denschub]: yeah that’s been my experience too, I use signal/whatsapp/email for private personal or small group chats, and tools like discord for small-medium semi-open chat-focused communities, which lets me focus on public stuff on my personal site and indieweb work
#
[tantek] [denschub] from a perspective of looking at the harms that "public" social media has caused society (disinformation amplification, threats to democracy, undercutting efforts at fighting climate crisis), yes, it is "controversial" as in worth thinking about deliberately in terms of the impacts of the tools we are creating
#
[tantek] [denschub], barnaby, I actually view *messaging* as a very different use-case / user-story from limited audience posts. One has an expectation of delivery, the other has an expectation of some people I know might see this and that's ok
#
[tantek] for example, IG friends-only stories
#
[tantek] that's the sort of thing I mean by "limited audience posts"
#
barnaby my site had support for limited-audience posts for a while, based on domain via indieauth login and email address via browser ID back when that was a thing. I just never ended up using it, because nobody was looking at my site while logged in
#
barnaby I can definitely see the value, it’s just not a practical priority for me at the moment
#
[denschub] [tantek] oh I do agree with the harm that public social media has done and is still doing. I am, however, not entirely sure that focussing on more “limited-audience” sharing is a big use-case these days, and I’m not even sure if that would resolve the issue of harm being done (I’m currently thinking about all those conspriacy-telegram-groups, where isolation and private comms only seem to make it even worse)
#
[tantek] barnaby, I've considered it because assuming I could make long living IndieAuth cookies work (similar to how our wiki works), I'm quite sure that my immediate family would login to my site and stay logged in, and check it for updates that way.
#
[tantek] if I got it "right" they would see the "family-only" or "friends-only" posts in my stream among others semi-transparently (maybe some indicator like a thin green border a-la IG friends-only stories)
#
[tantek] the other thing such a cookie/login could do is present a home/stream view of "here's all the new stuff I posted since you last visited"
#
[tantek] or a short message saying "You’re all caught up! I haven't posted anything new since you last checked."
#
barnaby [tantek]: do many of your immediate family members have personal domains where they could set up indieauth?
#
[tantek] lol they all do 😂
#
barnaby ha ha nice
#
[tantek] I've even purchased domains for nephews/nieces that I need to start getting setup for them
#
[tantek] I mean a personal domain is cheap enough to give someone as a birthday or holiday present
#
[tantek] and presumably you're willing to put in the time to help them auto-renew etc. so they hang onto it
#
barnaby I was going to suggest that setting up a passwordless emailed-sign-in-link system with long-lived auth cookies would be my method for this if I wanted my friends/family to be able to easily sign into my site
#
[tantek] "passwordless emailed-sign-in-link system" <- that's how you setup RelMeAuth via email 😄
#
[tantek] rel="me" href="mailto:..."
#
[tantek] so *you* can setup domains for them with RelMeAuth linking to their email address they prefer, and then *give them* their domain to use for sign-in
#
[tantek] your Web Sign-in UI could also accept plain email addresses which you'd support directly via "passwordless emailed-sign-in-link system"
#
barnaby lol it’s enough work just maintaining my own site, let alone taking on admin tax for my entire family in addition to the free tech support I already do for them xD
#
[tantek] and if/when they upgraded to using their personal domain with RelMeAuth, presto, your site would then know which domain to associate with which email they'd been using all this time!
#
barnaby yeah the latter solution is what I would do (and had back in the browser ID days(
#
[tantek] presumably one-page personal sites for IndieAuth would be quite easy for you to setup for your family on GitHub
#
[tantek] and if you do it right, zero maintenance
#
barnaby in addition to supporting indieauth, of course. just a single sign-in box where you can type a domain name or an email address and both work
#
[tantek] right, that's the work on your side
#
barnaby oh if they were using an external service for handling the indieauth stuff, sure, then only a static page is required
#
barnaby with taproot/indieauth, even setting up a self-hosted indieauth identity provider would be fairly trivial
#
[tantek] right, you're the one setting up the external service 🙂
#
barnaby but not something I want to offer other people quite yet
#
barnaby at least, not unless they ask me for it!
#
[tantek] this is a good technique for reducing the per-instance admintax of the code we write, make yourself setup&maintain minimum versions for your family 🙂
#
[tantek] this is why you only offer it as part of a gift that you would give to someone you *want* to do that for any way!
#
[tantek] a much smaller subset than "offer other people"
#
[tantek] [denschub] ok maybe not 15 but a lot still. I counted at least 8 dropped "options" here: https://www.w3.org/wiki/Socialwg/Social_API/Candidates
#
[tantek] that's a big part of the story that that post from cwebber completely ignores which sets a lot of important context for how we ended up where we did
#
[tantek] and that's only the set that WG members brought to the WG! there's a much better set of lists here: https://en.wikipedia.org/wiki/Comparison_of_software_and_protocols_for_distributed_social_networking
#
[denschub] I have the urge of linking a certain xkcd, but y’all know which one that would be anyway
#
[denschub] ah, thanks for that socialwg-wiki-link, I’ve never seen that. I guess that’s all a bit hard to summarize in a blog post targeted for people who aren’t necessarily interested in all the details, but I see your point.
#
barnaby !standards
#
Loqi http://xkcd.com/927/ http://imgs.xkcd.com/comics/standards.png
#
[tantek] what are standards
#
Loqi Specifications (or standards) are technical documentation for communicating between two or more implementations in an interoperable way; for the indieweb in particular, standards help reach greater levels of user functionality to publish, express meaning, notify, subscribe, and many other user actions https://indieweb.org/standards
#
[tantek] ^ also at the top of that page
#
[tantek] [denschub] it could have been a one-liner providing an estimate of the count of the various different prior approaches/projects that were proposed, and maybe listing/linking names of at least a half-dozen of them in a sentence. It changes the story dramatically
#
[denschub] that’s very fair, yeah
Evert1 joined the channel
#
[snarfed] I've been digging more into AT protocol. it's very federation world view, but if you grant it that, it definitely has some interesting ideas, notably around account migration
#
[snarfed] (to reirerate, interesting *federated plumbing* ideas)
#
[snarfed] for example, it attempts to store a full copy of your data on your local device as a backup, modulo disk space, which you can use to recover your account if your server/instance suddenly disappears
#
[snarfed] more in the weeds, under the covers, every user gets two private keys, one that their server has and uses for day to day operations, and another "meta" private key, which the server doesn't have access to, that can be used to revoke their server's access, migrate their account and data, etc
#
[snarfed] and the explicit distinction between individual peer to peer federation operations and global services (eg search) is refreshing, even if painfully underspecified: https://atproto.com/guides/overview#speech-reach-and-moderation . I discussed that a bit in my Indie Map talk back in the day, https://www.youtube.com/watch?v=bkc5afFohmI
#
[snarfed] granted, they also reinvent a few surprising wheels - RPC, schemas, social network data models and vocabs, etc - that don't seem worth the trouble. but they reuse a fair number of existing ones too - HTTP, did:web (and therefore DNS), NSIDs, IPLD, SHA256
slyduda_, oodani, [benatwork], Evert2, mro, barnaby, mro_, [kimberlyhirsh] and Mia1 joined the channel
#
sknebel https://mta.openssl.org/pipermail/openssl-announce/2022-October/000238.html
#
sknebel for the self-hosting crowd
mro joined the channel
#
sknebel important OpenSSL 3 patch tomorrow
#
sknebel (if your stuff is new enough to already use 3)
voxpelli, willnorris and AramZS joined the channel
#
@tkadlec ↩️ Honestly, the biggest outcome for me is giving me a big enough kick to finally sort out all the IndieWeb stuff I'd been putting off (webmentions, publish-once-syndicate-everywhere, etc) I've been saying I was going to do this for literally years, but I'm a slacker. (twitter.com/_/status/1587078079779930113)
geoffo and ben_thatmustbeme joined the channel
#
[tantek] whoa, am I only the second person here using Bridgy Fed?!?
#
[snarfed] no there are dozens
#
[snarfed] let's see
#
[snarfed] (but yes it's a small user base)
#
[snarfed] fluffy, unicyclic, [chrisaldrich], unrelenting.technology, [Khurt] are example recently active users who are here
#
[snarfed] looks like 132 total users (ever)
#
barnaby I hope to add myself to that list soon
mro joined the channel
#
[tantek] well I added myself to the list, I'll let others add themselves 🙂
#
[snarfed] UX recommendations are also welcome! eg I'm currently debating whether to auto-publish any posts, like we discussed briefly. not the right idea for separate silo accounts, which is why Bridgy Publish doesn't, but maybe it is here? 🤷
#
[tantek] Yes still thinking about that, to better fit expectations of how "feeds" work
#
[tantek] I'll try to file a feature request today if you haven't already
#
[tantek] [snarfed] speaking of which (the "empty" profile experience), while I have you here, could I upload https://snarfed.org/tantek_mastodon.png (or you could) to the wiki? Or would you prefer hotlinking to it 😈
#
[snarfed] Yes! Feel free
#
[tantek] Thank you!
geoffo joined the channel
#
@jrtashjian #Protocols > #Platforms. Platforms create their walled gardens on top of protocols. There's a reason IRC, RSS, SMTP, Webmention, Micropub, ActivityPub, Matrix, etc last. #OpenSource plays #TheInfiniteGame and wins in the end. (twitter.com/_/status/1587121846868185090)
#
ash[m] XMPP notably absent from that list
janboddez, AramZS, jonnybarnes and ben_thatmustbeme joined the channel
#
@TheGreenGreek ↩️ Same. I did make a Mastadon account and plan to get a bit more active there. Luckily I already have my blog though my webmentions are primarily pulling in Twitter atm. Also have a newsletter. I'm feeling keep this account and publish less on it while trying other platforms more. (twitter.com/_/status/1587138152115961865)
gRegor and boek joined the channel
#
@TheGreenGreek ↩️ Also, webmentions are enabled on my site https://sia.codes/posts/webmentions-eleventy-in-depth/ (twitter.com/_/status/1587144244782833664)
[aciccarello] and janboddez1 joined the channel
#
gRegor The Twitter archive is pretty impressive so far. Nice that they made it a HTML/JS browsable page. Probably js;dr though.
jonnybarnes and ben_thatmustbeme joined the channel
#
@TheGreenGreek ↩️ Lol nevermind. Apparently 1444 webmentions are cake to request and cache (twitter.com/_/status/1587172082022055936)
#
barnaby gRegor: I could be wrong but I think benward might have worked on the twitter backup when he worked there? He definitely worked on their embeds at some point
#
[tantek] !tell [manton] searching for the DNS setup page for micro.blog I found this help thread and your post https://help.micro.blog/t/help-setting-up-dns-for-micro-blog/182/2 which links to a "How To: Create Domain Shortcuts" article which currently 404s. Might be good to put in a redirect for that!
#
Loqi Ok, I'll tell them that when I see them next
#
[tantek] !tell [manton] like maybe redirect that 404 link to https://help.micro.blog/t/custom-domains/53 (hoping I found the right page)
#
Loqi Ok, I'll tell them that when I see them next
[jeremycherfas], Evert3, [Murray], [benatwork], gRegor, lagash, mcint and [jeremycherfas]1 joined the channel
#
@_alivealife_ ↩️ Sounds different. Webmentions is the way people talk to each other across platforms on the indieweb. (twitter.com/_/status/1587219096151670790)
#
@_alivealife_ ↩️ That's cool! Does it work with webmentions? (twitter.com/_/status/1587218370096730114)
#
@evertp Porting Curveball to Bun: https://evertpot.com/bun-curveball-framework/ #curveball #node #API #http #Bun (does webmention just work out of the box?) (twitter.com/_/status/1587229486554152966)