#dev 2021-02-03

2021-02-03 UTC
[KevinMarks], AkyRhO_, rhiaro_ and [schmarty] joined the channel
#
[KevinMarks] the big users of AS were the silo monitoring apps, which ended up very B2B, yes.
#
[KevinMarks] you can still get AS1 out of Jira, for example https://developer.atlassian.com/server/framework/atlassian-sdk/consuming-an-activity-streams-feed/
#
jacky TIL
nertzy joined the channel
#
jacky wow I've spent like an hour trying to make a Web Component to render a face pile lol
#
jacky and I failed lol
#
aaronpk there is no "fail", there is only "not yet succeeding"
ShadowKyogre joined the channel
#
jacky lol at this rate, that might be my permanent state :P
btrem, beko, [KevinMarks], [schmarty], [chrisaldrich], [benatwork], [Emma_Humphries], [pfefferle] and [indieweb653] joined the channel
#
@nhoizey Whoops, @simonhearne noticed `h-entry` and `p-author` in the links pages made http://brid.gy believe that… https://twitter.com/i/web/status/1354697969522630663 (twitter.com/_/status/1354697969522630663)
deltab joined the channel
#
@GR36 Look like webmentions from comments is now working! Now to make sure they are working everywhere else (twitter.com/_/status/1356901695683493889)
#
@GR36 So. If I write a post on http://micro.blog that mentions my own blog will it work? Look like webmentions from comments is now working! Now to make sure they are working everywhere else - Me https://gr36.com/4693-2/ (twitter.com/_/status/1356903759272632320)
#
@AndreJaenisch ↩️ Funnily I thought about a GitHub alternative built on top of WebMention. Can I reuse some of your thoughts (with proper credits) (twitter.com/_/status/1356910185391022082)
#
@voxpelli ↩️ Cool, I’m very much behind on the development within this space for the last 4-5 years, so this is great info! So I guess I could do some things to my WebMention endpoint to push it to have a wider support as well? (twitter.com/_/status/1356920408268824576)
#
@pfefferle ↩️ The site of @aaronpk can also be seen as a node: https://the-federation.info/node/aaronparecki.com I think he was also the first to introduce #webmention as a federation-protocol: https://the-federation.info/protocol/webmention :) (twitter.com/_/status/1356919770067701760)
#
@pfefferle ↩️ Sure! @schnarfed also built a bridge that translates between #Webmentions and #ActivityPub or #OStatus https://fed.brid.gy/ (twitter.com/_/status/1356920990928957443)
#
@pfefferle ↩️ Sure! @schnarfed also built a bridge that translates between #Webmentions and #ActivityPub or #OStatus https://fed.brid.gy/ (twitter.com/_/status/1356920990928957443)
#
jeremycherfas Am I right to assume that timezone for this is CET? https://fosdem.org/2021/schedule/track/open_research_tools_and_technologies/ I can't actually see it explicitly anywhere.
#
sknebel yes
#
jeremycherfas Thanks.
#
jeremycherfas Some of it looks tangentially interesting to me.
#
sknebel yeah, fosdem often has interestng things. although I personally tend to go through the list afterwards, not much one for watching synchronously
paul3 joined the channel
#
jeremycherfas Especially if things have been prerecorded in the first place.
[pfefferle] and [KevinMarks] joined the channel
#
[KevinMarks] Is he being deliberately misleading, or is he actually unclear of the distinction? https://twitter.com/kevinmarks/status/1356930147207831552?s=20
#
@kevinmarks @balajis Turning products into protocols does make sense, but entangling them with cryptocurrency doesn't do that. A platform is easier to use than a product. A crypto entangled protocol is harder. You're adding friction, not removing it. (twitter.com/_/status/1356930147207831552)
Saphire and [schmarty] joined the channel
#
[schmarty] KevinMarks: I lol'd at "carrots of crypto". I feel his tweets are coming from some macroeconomic view where the only players are state-level actors.
#
[KevinMarks] he's a VC so he really like financializing things
beehiveth and gxt joined the channel; ShadowKyogre left the channel
#
Zegnat https://github.com/w3c/webappsec-change-password-url is one of the few .well-known URL specs I can totally get behind. Via https://adactio.com/journal/17794
#
Loqi [w3c] webappsec-change-password-url: A Well-Known URL for Changing Passwords
#
aaronpk wow iOS safari has already implemented it?
#
Saphire Ooh. Is there a list of .well-known things?
#
Saphire aaronpk: I think it is one of more widely supported things due to various password managers rallying behind it?
#
aaronpk it's unusual for Apple to be the first one to implement things in safari :)
#
aaronpk tries to think of a place he has with a password dialog to try this out
#
Zegnat IESG/IANA I guess keep a list of .well-known things, as far as any official registration of them is concerned. But I would not have a link at hand
ShadowKyogre joined the channel
#
aaronpk has managed to successfully not have passwords for most sites he runs thanks to OAuth/IndieAuth
#
Zegnat Saphire: https://www.iana.org/assignments/well-known-uris/well-known-uris.xhtml is the official list, I think?
#
Zegnat I have always found these IANA registries a little magic. People add it to an RFC with a little IANA section, and then … submissions to age old mailing lists I guess?
#
sknebel yes, it says right at the top who to contact :D
#
Saphire Oh hey I know person who published sshfp
#
Zegnat Guess we did not bother for IndieAuth, which doesn't seem to be listed on https://www.iana.org/assignments/link-relations/link-relations.xhtml ... aaronpk is that something we should do?
#
Zegnat To actually list authorization_endpoint as a link relation, that is, and not just have it as an IANA Considerations section in the IndieAuth spec
ShadowKyogre left the channel
#
aaronpk is surprised to see micropub on that list
#
aaronpk i don't remember doing that
ShadowKyogre joined the channel
#
Saphire I am actually surprised how common rel="me" is
#
Saphire Because I have noticed it for years before finding indieweb >.>
#
aaronpk Zegnat: i think if we do that we're going to have to finally come to terms with this: https://github.com/indieweb/indieauth/issues/17
#
Loqi [Zegnat] #17 Link relations in HTTP Link header conflict with RFC 5988.
#
Zegnat Ah, right, forgetting about my own issues :P
#
aaronpk i do think my latest comment there might be the best way to handle it
ShadowKyogre left the channel
#
aaronpk especially since it means fewer things to register as link relations
ShadowKyogre joined the channel; ShadowKyogre left the channel
#
Zegnat aaronpk: any idea how far along pushed auth extension is? Saw they issues a new draft just yesterday. Is there any worth in tryin to do those two simultanious as part of a breaking-change modernisation, or no?
#
aaronpk i don't think there will be a last call any time soon on that
#
aaronpk also ugh i hate all this namespacing "urn:ietf:params:oauth:request_uri:..."
#
Zegnat Haha
#
Saphire GUID!
#
Zegnat What is that params supposed to mean? urn is easy, the protocol. Then ietf for the standards body. But surely you could follow that up with oauth immediately?
#
aaronpk IIRC it has to be a URI because it's actually building on this draft which uses UDIs https://tools.ietf.org/html/draft-ietf-oauth-jwsreq-30
#
aaronpk functionally the only significant part is the string at the end
#
aaronpk oh good, missed this before: "The format of the "request_uri" value is at the discretion of the authorization server"
#
Saphire ...for some reason I am now thinking of C's UB
#
Zegnat request_uri=mailto:martijn@zegnat.net - I love it when things are at my discretion
#
aaronpk lol well
#
aaronpk "...it MUST contain some part generated using a cryptographically strong pseudorandom algorithm..."
#
aaronpk but as far as i can tell, technically PAR doesn't say it *has* to be a URL or URN
#
Saphire Also, https://age-encryption.org/ looks neat
[Murray] joined the channel
#
Zegnat request_uri=mailto:martijn@zegnat.net?title=cc80a3e8be641076abff93e0f0ebcb82 - there you go, added some bytes :P
#
aaronpk oh right, but request_uri is actually defined in JAR
#
Zegnat ah, darn
#
Zegnat age is nice, Saphire. I think I would feel fine recommending it to people if they wanted to encrypt some files or other blobs.
#
aaronpk JAR says it has to be an https URI or a URN
#
aaronpk i still don't like it
#
aaronpk Zegnat: if you add a scheme like mailto: to your example i think it counts :)
#
Saphire The best part is that age encrypted things have a link to explain what the heck it is
#
aaronpk oh wait no it says specifically https or urn
#
Zegnat Hmm, alright, so either use an HTTP-resolvable-URL or use a completely-made-up-URL. Gotcha.
#
aaronpk i just want to use a random string, i don't see why i have to namespace it into urn:
#
Zegnat Saphire: I first read about age back in 2019 when it was mentioned as an alternative to PGP for file encryption: https://latacora.singles/2019/07/16/the-pgp-problem.html#encrypting-files
#
Zegnat aaronpk: I guess forcing it to follow an URL/URI/URN standard means it can be more easily expanded to other things besides https and urn in the future without clashing with the random strings people have been picking?
#
aaronpk i guess
#
Zegnat Time to register urn:indieweb ? :P
#
Zegnat I kinda liked the tag URIs for naming. Rather than having to register a namespace with IANA the namespace is instead your email address or domain name combined with a date that you own that domain at.
[tantek] joined the channel
#
[tantek] Zegnat, the link registry for HTML rel values is specified by the HTML spec, which links to the microformats rel-registry and it's been that way for ages. The IANA one is only for HTTP headers
#
[tantek] so don't be "surprised" when things are missing from it
#
aaronpk this is for HTTP Link headers tho
#
[tantek] the IndieAuth ones are in HTML though right?
#
[tantek] aaronpk, you can make sure they're in https://microformats.org/wiki/existing-rel-values
#
Zegnat Not sure if they were added to the wiki registry?
#
Zegnat The aforementioned relation name issue is also specifically re HTTP Link headers. HTML rel values are more lax in their definition of allowed values, it turned out
#
[tantek] you or anyone with edit permissions can! no need to email someone and go through some ancient process
#
aaronpk yeah the indieauth ones have been on the microformats wiki for ages
#
Saphire Oh, curious, is there any kind of thing that adds HTTP header with a link to a "rich" content page, when looking at direct links to files (photos, etc)
#
aaronpk what's interesting to me is as far as i can tell, HTTP clients have no trouble with the technically disallowed underscore in header names
#
Zegnat micropub, webmention, and indieauth (except redirect_uri) are all on the wiki. microsub is not.
#
aaronpk which suggests that maybe the http spec should be updated to reflect reality
#
Saphire I guess HTTP spec is "minimum must support"?
#
Zegnat HTTP clients I think have no issue because none of them validate (or do anything, actually) with the contents of a Link header. It is a valid HTTP header.
#
Zegnat It is just not a valid value for a Link header per the spec on Link relations
#
Zegnat And who bothers to implement that spec to the tee?
#
aaronpk https://tools.ietf.org/html/rfc8288#section-3.3
#
Zegnat Yep. That is the Web Linking spec, not the HTTP spec. So I am guessing HTTP clients do not care ;)
#
aaronpk also "Note that extension relation types are REQUIRED to be absolute URIs..."
#
aaronpk sure, my point is clearly the specs do not reflect the reality of most implementations, so something is wrong with the specs
#
Zegnat yes. But in this case "extension relation types" = unregistered types.
#
@XabCorinti #100DaysOfCode R3|D4 Still struggling with comments. I have decided that I'll display the old ones in plain html (they are from 2013-2017, a Staticman and previous disqus bunch) so the discussion will always be part of the post For the future, I think Webmentions is the way (twitter.com/_/status/1356995865773297670)
#
Zegnat Was that not why webmention originally used a URL as rel, aaronpk?
#
Zegnat That is what I always thought ...
#
aaronpk it originally used a URL because sandeep decided to use a URL, i don't know why he chose that option
#
Zegnat Oh, I assumed it was because of that RFC, because that is what you do until you register
#
aaronpk could be
#
[tantek] I think because he had been exposed to some RDFisms and eventually we worked together to simplify it to a keyword because it reality that's all you need
#
aaronpk i never asked
#
[tantek] in* reality
#
Saphire Mmh. Kinda currently unsure where to go and what to implement on my site.. Kinda would love some tips?
#
aaronpk Saphire: you might find some ideas here! https://indieweb.org/IndieMark
#
Saphire Maybe try to compile a nice annotated master list of all the currently used microformats/etc. And then look over it to see what to implement..
#
Saphire aaronpk: oh yeah, that page ^^' Lack a blog so that might be a good thing to try and make in first place, hm.
#
[tantek] Saphire, best place to start with what to implement is to ask yourself what do you want to use your site to do? And if you don't know and want ideas, you can start with: How can I use my site instead of social media?
#
[tantek] interestingly enough we used to answer this question on our "Getting Started" page, that is, it was more implementer-centric, but we've subsequently edited for folks who want to primarily setup and use the IndieWeb rather than implement
#
[tantek] And then yes, IndieMark is a way to implement simpler steps first, along the paths that you want for your own site
#
Saphire Honestly, want to mess around with all the microformats and such. But lack content to use it with. Hm..
#
[KevinMarks] Indiewebify.me will nudge you along too
#
Saphire Well, it suggests to check posts markup. Posts I need to make in first place? Hm..
[arush], [benatwork], ShadowKyogre, DanC and KartikPrabhu joined the channel; ShadowKyogre left the channel
#
jacky that's a bug imo
#
jacky but also expecting indiewebify.me to be a "wizard" of sorts is a bad assumption (the one we give for it)
[schmarty] joined the channel
#
aciccarello[m] Does anyone know of an easy way to be alerted of 404's with a netlify site?
#
aciccarello[m] Or at least be able to check 404s
#
jacky aciccarello[m]: I used to use https://updown.io/ for this
#
jacky which reminds me to bring it back lol
ShadowKyogre left the channel
#
aciccarello[m] Nice, that seems good for checking a server. I'm more-so thinking for when I link incorrectly or accidentally change a page URL.
[snarfed] joined the channel
#
Saphire https://en.m.wikipedia.org/wiki/Really_Simple_Discovery huh
[Ana_Rodrigues] joined the channel
#
Saphire People really like to nuke this from their WordPress head for some reason x3
[Rose] and ShadowKyogre joined the channel
#
jacky lmfao
#
Saphire Mmh?
#
jacky re: RSS removal
[calumryan], KartikPrabhu, ShadowKyogre, miklb_, lahacker and jamietanna joined the channel; ShadowKyogre left the channel
#
jamietanna hey aaronpk - is https://php.microformats.io/ working for you? seems to be timing out for me, wasn't sure if it was related to the bump to PHP 7.4 you mentioned?
#
aaronpk that one is actually on heroku
#
aaronpk huh not sure what's wrong
#
aaronpk same code lives at https://pin13.net/mf2/ on a different server tho, can't fix it right now
#
jacky re: those sites
#
jacky if I spun up something, could it be CNAME'd to something like https://elixir.microformats.io?
#
jacky or is there only a PHP one and more isn't needed
#
aaronpk yea for sure
#
jacky scribbles another note
#
aaronpk that's the goal with the domain, there's a bunch of them already: https://microformats.io/
#
aaronpk if it can be run on heroku i can launch it on our account there, otherwise you're welcome to run it and we can CNAME it
[KevinMarks] joined the channel
#
jacky oooh that's even better
#
jacky (and would be easier tbh)
#
jacky because then that's a subtle incentive for me to finish working on the Rust parser
[benatwork] and [tantek] joined the channel
#
[tantek] re: "expecting indiewebify.me to be a "wizard" of sorts" <-- perhaps file an issue to turn IndieWebify.me into a Micropub client? Then it could optionally post a "badge" for you to your site when you achieve a certain level of support 🙂
#
[tantek] Saphire, do you use any "social media" or any other content silos?
#
Saphire Uh, very much yes
#
@dnzm Nog even snel een #blogpost op de woensdagavond, want mijn blog ondersteunt nu ook #webmentions. https://doenietzomoeilijk.nl/archief/webmentions-en-isso (twitter.com/_/status/1357096207378710529)
#
[tantek] well that's a great place to start! pick one (maybe where you only post "notes") and start building the ability to do so on your own site instead!
#
jacky so on my walk with Stella, I was thinking about my site
#
jacky and how I wanted to test how different post types could render
#
jacky I was wondering if a tool could send a request to post content of a particular type and then conditionally delete it when needed
#
jacky the thing about it that I was thinking about the most was having it pull the list of known post types from the wiki and doing deeper introspection on what kind of markup to use
#
jacky (the latter half is a bit much and would probably read from #How_to_Markup)
#
jacky but I can imagine that being handy for people who want to not only test their site but also use the wiki as a source of test cases!
gRegorLove and [arush] joined the channel
#
[tantek] jacky, does your system allow for mutability of "post types" based on the presence or absence of some number of properties?
alex11 and ShadowKyogre joined the channel
#
jacky I can't see why not
#
jacky tbh it would be as simple as picking a post type, picking a set of properties and then sending it to a site
#
jacky tbh it'd be _amazing_ if it was like how StackOverflow allows you to run code snippets in the browser to test answers
gxt and [snarfed] joined the channel; ShadowKyogre left the channel
#
[snarfed] [Zegnat] did that for authorship, maybe he’d do it for PTD too