[eddie]ohhh, that’s true. After nginx, my website does hit my node.js server before being directed either to static files (jekyll) or dynamic files. So I could add my list in to the node.js side of things
tantekwhoever wrote https://commentpara.de/comment/169.htm (I'm guessing someone here?) I noticed after the fact that the links were not hyperlinked in my GitHub issue and updated the post :)
tanteklike these: https://adactio.com/journal/12585#comment56543 (literally that comment and every one after it is from the same GitHub thread that I just happen to be participating in, and that happened to have earlier linked to adactio's post)
tantekadactio is going to wake up to a lot of new comments on his Container Queries article and wondering *why* they are being sent to him (since none of them directly link to his post)
tantek!tell snarfed you may want to take a look at all these comments generated by Bridgy Backfeed from GitHub to adactio's post: https://adactio.com/journal/12585#comment56543 (note that adactio himself is not signed up on Bridgy GItHub, this is a side effect of me opening the issue and linking to adactio's article in the middle of the issue description)
tantekwhich sorta made sense for Twitter since it was usually just *a link* in a tweet, and it's likely the tweet / thread / replies are regarding that link
Loqisnarfed: tantek left you a message 10 minutes ago: you may want to take a look at all these comments generated by Bridgy Backfeed from GitHub to adactio's post: https://adactio.com/journal/12585#comment56543 (note that adactio himself is not signed up on Bridgy GItHub, this is a side effect of me opening the issue and linking to adactio's article in the middle of the issue description)
aaronpkFor example I would appreciate knowing if a GitHub issue or comment linked to one of my blog posts, but I don't want the whole GitHub thread to end up there
tanteksnarfed, re: text like "Originally published at", maybe if the last link in an issue is parenthesized ( http... ), and if it is a non-home-page link to the issue author's domain?
[markmhendricksoAre folks here fans of Docker? And are there any IndieWeb projects that use it? I'm considering it for mine after one request to do so, but I haven't explored it before
cweiskeon a normal server, you use the system package manager to install and update packages. with docker, you lose that advantage because you can't just "apt update && apt upgrade" to get security updates
cweiskeyou have to get new versions of each of the containers you're running, which is tedious. and unless you're building the containers yourself, you have to wait for the maintainer to update them
cweiskeaaronpk, indieauth.com tells me that "something went horribly wrong" when signing into webmention.io with http://cweiske.de/, with no more infos. could you check the logs?
ZegnatI am not sure how I completely missed that one, snarfed! Will have to add. Too bad they don’t use any words around it so it will not be helpful in narrowing down a possible property name.
ZegnatInteresting stuff snarfed! I have to take a moment’s break from wiki editing but will add that later! Do you happen to have any thoughts on how we could express this best in mf2?
ZegnatI don’t know what I think of the term generator. If I see parsed mf2, I would almost expect the generator to tell me which parser was used, rather than what software was used by the author of an h-entry to create said h-entry.
ZegnatYou might be right that generator as a name has been around so long that it makes sense, [kevinmarks]. I’m just always thinking: if I see the mf2 JSON of an h-entry and I read the properties will I instantly understand what a property expresses without going to the wiki?
[eddie]Between client and generator, client definitely feels more natural IMHO. Generator feels like it was created by an AI or algorithm. Client at least feels like it was the tool that was used to create it.
ZegnatI think “via” makes sense for things like OYG, but does it make sense for e.g. tanteks usecase of just “BBEdit”? How does that sound to the native English speakers?
aaronpkI was always unclear on how BBEdit relates to Tantek's posts but that's because I don't really know what BBEdit is. In 2010 when Tantek had a blackberry I thought it was a blackberry app.
sknebelvia sort of seems fitting for PESOS services, less so for direct clients. and has conflicts with "found via" (which admittedly isn't as common anymore, and also hasn't been marked up)
ZegnatI wonder if some of those app-generated posts using “via” aren’t more like a person saying “found via Person than “posted using Application”. E.g. “via Buzzfeed” may be expressing the former more than the latter, even if the Buzzfeed application was also responsible for the push.
[kevinmarks]Nuzzel example "How Trump Conquered Facebook Without Russian Ads - WIRED http://nzzl.us/ju9rNhM via @nuzzel thanks @EmilyDreyfuss" - that's what they prompted me with when I share an article via twitter in their app.
tantekand Bridgy Publish is already adding one proprietary publish API at a time, if it added Micropub as a publish "target" then it reaches everything silo.pub already supports publishing to
tantekI think it was a testament to the reusability of Webmention as a building block, beyond the original use-cases that the designers/developers envisioned
snarfedthis bridgy->silo.pub idea may also be good in theory but not in practice, since hosted silo.pub itself is dead and maybe unlikely to return anytime soon
LoqiIt looks like we don't have a page for "Webmention overloading" yet. Would you like to create it? (Or just say "Webmention overloading is ____", a sentence describing the term)
sknebelsnarfed: you seem to have covered it pretty well ;) I had the same thought with reusing the the token, but Inoreader uses refresh tokens so it doesn't work for my specific scenario
[eddie]Ultimately I hope to get to the point of being able to syndicate via micropub rather than webmention (as I’m doing now). I enjoy bridgy, but for my workflow, definitely would rather use micropub. That said, silo.pub is down (so I’m glad I never started using the public instance). I’m thinking long-term the likely solution is a PHP version of silo.pub. So the question is: is silo.pub in another language silo.pub (for language) or is it some
KartikPrabhu[eddie]: this is one of the things I really like about indieweb; you can use Swift/PHP and I can use Python and we don't really have to fight over it :)
[cleverdevil]Hmm... having a lot of trouble figuring out how to use Together as my Micro.blog client. Primarily having issues determinding sending replies.
snarfeddiversity is good and all, but i think we as a community get less value and learn less from reimplementing n silo APIs than from implementing things on our own sites or protocols
snarfeda standalone proxy would accept requests, generate temporary html pages from them, send bridgy a webmention, and serve the temporary page as the wm source
[cleverdevil][manton] sure. The way I decide whether or not to syndicate a particular in-reply-to post to my Micro.blog JSON Feed is by looking at the in-reply-to URL. If the user is hosted at Micro.blog, that URL will be "https://micro.blog/username", and I can also use that to pull the username and inject the hyperlinked @-mention into the feed.
[manton]One of my concerns is that there are too many ways to do the same thing. So in Micro.blog I want to focus on just 2: being smart about detecting replies and threads in a feed, and Webmention. Both have some holes as [eddie]'s GitHub issue points out, but I think they can be improved.
[manton]Me too. That's where I want to go with Micro.blog eventually too. @manton.org instead of just [manton], so it works with people who haven't registered. That's the reason I created the test @nytimes.com Micro.blog account.
LoqiA nicknames cache is a way indieweb sites store information about people to improve the user experience of the site owner referring, mention, and/or linking to those people https://indieweb.org/nicknames-cache
[manton]To back up a bit... [cleverdevil], if you had an in-reply-to in your feed that pointed to an external site, what would you expect it to look like in the Micro.blog timeline?
aaronpkIf micro.blog were to consume my replies feed it would be able to match up my in-reply-to URLs whether they're micro.blog URLs or canonical URLs and id want it to ignore things it didn't match
Zegnattantek, I somehow missed your question about use case for marking up using/generator/etc for posts in between all the micropub and webmention talk! Woops.
ZegnatWith implementing Micropub I am taking the mf2 JSON it submits and storing that as the data for every post. I want to merge the details about the application I used into that same JSON. And would also like to be able to recreate said JSON (with an mf2 parser) from the outputed HTML.
ZegnatSo I was looking into how people (and silos) are already displaying said data, and whether there is some comonality in their markup, that I can copy.
[eddie]I think aaronpk hit the nail on the head. The goal is to consume all the replies that are for posts within the micro.blog system, regardless of their canonical URL
[manton]I'd lean toward wanting to process all the replies and improve how they look for external sites. Piggybacking on the domain name discussion above, maybe even adding @someone.com for replies to external sites that Micro.blog has no idea about.
[cleverdevil](That said, that was always my vision with Together: let people send everything they have, and let the user decide if they want to see it or not).
[manton]Related [eddie], I think there's another choice in your GitHub choices... I'd call it 2b: pull the source for a permalink in a feed and check it for Microformats. That is a lot more work for the server to do, though.
[eddie]Yeah, that’s a valid concern. Definitely something to think about and look into. If you can swing it, seems like the most complete. If it’s not tenable, then I guess it’s back to the list with the question of what is “second best” but still performant
sknebelsince you recently talked about reporting back syndication URLs, if those are there a posting interface could automatically decide to syndicate the reply to micro.blog as well
rainbow-zshjeremycherfas messaged me re: client attribution. I maintain Macchiato for 10Centuries AKA 10C. The API automatically attaches client info based on the token used to create a post. Docs are at: https://docs.10centuries.org/content
rainbow-zshThe docs are not terribly complete, but the creator is responsive to questions, and the first-party Web clients are unminified, so they’ve been sufficient for me.
[manton]Quick correction from the earlier discussion about in-reply-to in feeds: I meant "3b", not "2b". "2" is Micropub syndication, "3" is MF2 in feeds.
AngeloGladdingaaronpk thankfully just a duplicate of Zegnat's from Dec -- rel="authorization_endpoint" set to a relative path -- experienced the same with micropub.rocks not too long ago
aaronpkso <div class="e-content"><b>foo</b></div> will result in {"html":"<b>foo</b>","value":"foo"} and <div class="p-content"><b>foo</b></div> will result in just "foo"
tantek is looking at https://github.com/snarfed/bridgy/issues/794 and now going to dig through his webmentions to see if anything special happened
Loqi snarfed: tantek left you a message 10 minutes ago: you may want to take a look at all these comments generated by Bridgy Backfeed from GitHub to adactio's post: https://adactio.com/journal/12585#comment56543 (note that adactio himself is not signed up on Bridgy GItHub, this is a side effect of me opening the issue and linking to adactio's article in the middle of the issue description)
jeremycherfas.net edited /Indieweb_Podcast_Club (+200) "/* Ask the Indieweb about the Indieweb */ Add thoughts for first episode" (view diff)
[markmhendrickso Are folks here fans of Docker? And are there any IndieWeb projects that use it? I'm considering it for mine after one request to do so, but I haven't explored it before
petermolnar I still share the view that docker is a very complicated package management system
cweiske aaronpk, indieauth.com tells me that "something went horribly wrong" when signing into webmention.io with http://cweiske.de/, with no more infos. could you check the logs?
[kevinmarks] generator in Atom+RSS, text name +URL
KartikPrabhu I use "via Bridgy" for webmention replies that come through Bridgy
[eddie] Ultimately I hope to get to the point of being able to syndicate via micropub rather than webmention (as I’m doing now). I enjoy bridgy, but for my workflow, definitely would rather use micropub. That said, silo.pub is down (so I’m glad I never started using the public instance). I’m thinking long-term the likely solution is a PHP version of silo.pub. So the question is: is silo.pub in another language silo.pub (for language) or is it some
[cleverdevil] Hmm... having a lot of trouble figuring out how to use Together as my Micro.blog client. Primarily having issues determinding sending replies.
rainbow-zsh jeremycherfas messaged me re: client attribution. I maintain Macchiato for 10Centuries AKA 10C. The API automatically attaches client info based on the token used to create a post. Docs are at: https://docs.10centuries.org/content
Zegnat 10Centuries documentation on client attribution I ended up with: https://indieweb.org/User:Vanderven.se_martijn/h-entry-using-property#10Centuries
[jeremycherfas] Rainbow-zsh++ for responding to my call for help.