2013-04-24 UTC
caseorganic, tantek, mxuribe and mxuribe1 joined the channel
# 01:10 aaronpk !tell tantek no there isn't, which will be funny when someone named eschou shows up in the channel in a couple years
# 01:10 Loqi Ok, I'll tell him that when I see him next
scor joined the channel
duckbillp and tantek joined the channel
# 01:49 Loqi tantek: aaronpk left you a message 39 minutes ago: no there isn't, which will be funny when someone named eschou shows up in the channel in a couple years
spinnerin joined the channel
caseorganic joined the channel
duckbillp joined the channel
spinnerin, tantek, b0bg0d and duckbillp joined the channel
spinnerin joined the channel
# 05:39 tantek or is p3k going to be the second automatic indieweb comment display supporting tool?
# 05:46 aaronpk this code and design will also be used to display comments below as the next step
# 05:53 aaronpk i have to say, the new microformats2 JSON spec is quite nice
andreypopp and morrocco_mole joined the channel
b0bg0d joined the channel
andreypopp and seyz joined the channel
# 07:15 Loqi Ok, I'll tell him that when I see him next
b0bg0d joined the channel
eschnou joined the channel
# 07:38 Loqi eschnou: tantek left you a message 8 hours, 12 minutes ago: I wrote up a brief list of ways to determine the author of a post: http://indiewebcamp.com/authorship - please review and let me know how what you think.
b0bg0d joined the channel
# 07:56 eschnou !tell aaronpk Catching up on the surprise of the night, great work on the replies context!
# 07:56 Loqi Ok, I'll tell them that when I see them next
duckbillp joined the channel
lindas, friedcell and b0bg0d joined the channel
andreypopp joined the channel
barnabywalters joined the channel
b0bg0d joined the channel
# 09:21 barnabywalters aaronpk: loving the reply contexts — the layout for multiple replies is fascinating and rather bizzare! my comment on it would be that it doesn’t effectively communicate the order in which the replies happened
# 09:22 barnabywalters I rather like twitter’s approach where they show the entire context in sequence, but you have to scroll up to see the whole thing
# 09:22 barnabywalters I’m lagging behind a bit with all of this ;) Might get reply contexts working today
# 09:24 barnabywalters might be an idea to put those context replies in a child h-feed on the parent h-card
# 09:26 barnabywalters e.g. just the link to the replied-to content, or it embedded, or a feed of context
# 09:28 barnabywalters aaronpk: might also be an idea to collapse the multiple-reply grid into a single column on thin screens :)
b0bg0d, xtof, andreypopp and donpdonp joined the channel
mxuribe and qirtaiba joined the channel
# 12:20 qirtaiba Hello, I'm new here
# 12:20 qirtaiba Wondering why I can't get indieauth to work
# 12:20 qirtaiba I added the "rel=me" tags to the links on my home page
# 12:21 qirtaiba And I added the links back there from Facebook, Twitter and Google+
# 12:21 qirtaiba But it won't log me in
# 12:21 qirtaiba What did I do wrong?
# 12:21 qirtaiba My page is jere.my
# 12:22 tommorris you might want to change the Twitter link to remove the hashbang part. ;)
# 12:22 tommorris but it's probably going to the openid rather than the rel-me links
# 12:23 qirtaiba Done, and trying again
# 12:23 qirtaiba Only for Twitter, but that's oK
# 12:24 tommorris facebook doesn't link back with rel-me, and neither does Google
# 12:24 qirtaiba I'm going to be working on my blog software to add all the cool features that you guys are working on
# 12:24 qirtaiba Then I'll add it to the wiki
# 12:24 tommorris I've got loads I want to add to mine, when I can find time to work on code outside of work. ;)
# 12:25 qirtaiba I'll change it again
# 12:26 tommorris yeah, that's a known problem we argue about on the github issues page. ;)
duckbillp and b0bg0d joined the channel
duckbillp and cweiske joined the channel
andreypopp and matthias_ joined the channel
# 13:11 cweiske about indieweb comments: how do you extract the comment content from the webpage?
# 13:11 cweiske do I have to use a special class for the main element?
scor joined the channel
# 13:14 cweiske and you guys really don't support openid providers for indieauth? shame
# 13:15 cweiske and I have to rely on indieauth.com to be available, which is bad
b0bg0d joined the channel
duckbillp joined the channel
# 13:18 barnabywalters earlier on it did support openID as well as relmeauth, but there were many problems so aaronpk turned support for it off
# 13:20 Loqi cweiske meant to say: h-entry it is then
andreypopp, pfefferle and tantek joined the channel
# 13:52 tantek and of course just displaying the context of replies period! :)
# 13:52 tantek time order makes sense. I like how they're essentially laid out left to right in a 2 column grid.
b0bg0d joined the channel
# 13:54 tantek and that is silly about app.net failing to even publish microformats when it's trivial to do so
# 13:55 tantek is alpha.app.net somewhere on github so you can submit a 2 min pull request to fix it?
# 13:57 tantek perhaps it's a hint to not bother replying to people on app.net until alpha.app.net actually has support for some of the simplest standards like microformats ;)
pfefferle and texburgher joined the channel
# 14:19 pfefferle I am thinking about a WordPress plugin to support IndieWeb commenting. Adding support for Posts should not that big of a problem, but what about comments. Comments are accessible under the same permalink as the blogpost (in general), so how should the other party now what to interpret?
# 14:25 tantek pfefferle - that shows that comments deserve their own permalinks
# 14:26 pfefferle yes, but that would make it very complex for a "classic" blog
# 14:26 tantek servers don't typically pay any attention to fragments - and can't expect them to receive them
# 14:26 tantek a classic blog should also be able to post indieweb comments as their own posts
# 14:28 tantek pfefferle - if you notice, that's how tweets work
# 14:31 pfefferle yes, and it works very well in twitter, because they don't make any differences between "posts" and "replys"
# 14:34 aaronpk barnabywalters: re: "it doesn’t effectively communicate the order in which the replies happened" <-- that was actually my exact goal with that layout :)
# 14:34 tantek pfefferle - so clearly wordpress's design needs rethinking
# 14:34 Loqi aaronpk: eschnou left you a message 6 hours, 38 minutes ago: Catching up on the surprise of the night, great work on the replies context!
# 14:35 tantek (which we've found to be true about many aspects of wordpress)
# 14:35 tantek (hence nearly everyone in #indiewebcamp is implementing *their own publishing system*)
# 14:38 aaronpk i would love to see more stuff be built for wordpress personally, they've done a great job of making it easy to install on a host, and it's infinitely extensible, so it's entirely possible for it to be a solid indieweb solution
# 15:00 barnabywalters aaronpk: fair enough :) I found it to be confusing compared to, for example, twitter’s layout
# 15:01 barnabywalters I didn’t know whether or not to read it LtoR or TtoB, and I find absolute timestamps really difficult to parse and compare quickly
b0bg0d and tantek joined the channel
# 15:40 tantek aaronpk, barnabywalters - I like the innovation with the two column layout to indicate clustering
# 15:40 tantek however I'd also be ok with simple linear time order
# 15:42 tantek what's interesting in the case of http://aaronparecki.com/replies/2013/04/21/1/ is that all the "originals" that that post is a reply to, are themselves all a reply to a single one of Aaron's posts - so he could also show that post as context above the four that are already there.
# 15:43 tantek pfefferle,aaronpk, that thinking, that "love to see more stuff be built for wordpress" is double-edged.
# 15:43 tantek on the one hand, yes, wordpress is hugely popular and anything to get wordpress folks to be more indieweb would likely get good traction.
# 15:44 tantek on the other hand, wordpress is the top source of indieweb sites being compromised, and unfortunately people switching to hosted silos like tumblr.
# 15:44 tantek in addition, the "love to see more" often gets reinterpreted as "only build on wordpress" which turns wordpress into a monoculture bottleneck.
# 15:44 tantek witness the original DiSo - which had the vision of being implemented as a bunch of WordPress plugins - and stalled out accordingly.
aaronpk|m joined the channel
# 15:46 aaronpk|m barnabywalters: that's the entire point :) the posts above mine don't matter what order they were written in
# 15:47 aaronpk|m It would be clearer if as Tantek suggested, my original post was above all 4
# 15:48 tantek.com edited /projects (+125) "/* WordPress */ separate out and highlight Christophe Ducamp's use of WordPress with plugins for better IndieWeb/POSSE support" (
view diff )
# 15:49 tommorris crikey, ladies and gents, all this hard work on comments... I've got lots to catch up on. ;)
# 15:53 tantek these guys (eschnou, aaronpk, xtof, barnabywalters) are making indiewebcamp history. all I could do so far is record that historical event. ;)
# 15:54 tommorris Yahoo Hack Europe this weekend. not sure whether to build indieweb stuff or make a fun web app to troll^Wconstructively criticise the culture around big companies like Yahoo
morrocco_mole and spinnerin joined the channel
# 16:18 tantek barnabywalters, exactly. I'm about to give it as a bad example in an interview I'm doing about URL usability :)
# 16:18 barnabywalters I have seen some sites actually use the tick in query params — I thought as a joke, but apparently it can help fix some weird IE UTF8 bug (????)
tilgovi, caseorganic, andreypopp, pfefferle and b0bg0d joined the channel
# 17:06 barnabywalters aaronpk I’m having some problems with php-mf2 and your individual notes pages — converting classic microformats changes the .published on the parent element of dt-published to dt-published, causing the entire element to be processed as the datetime! if the .published is for classic microformats, I’m pretty sure it should be on the time element too
pfefferle_ and tantek joined the channel
tantek and caseorga_ joined the channel
caseorganic, tantek and cweiske joined the channel
# 18:06 cweiske I don't know who is responsible for pingback.me, but the page title is empty
tantek joined the channel
# 18:10 cweiske how does "focused on growing the independent web" relate to the fact that indieauth forces me to use a central login provider instead of my own?
# 18:11 barnabywalters cweiske: by which you mean the indieauth.com domain, or the auth providers (twitter, github, etc)?
# 18:12 cweiske so there is only me and the site I want to login to that have to be running
# 18:12 barnabywalters cweiske: the indieauth.com software is open source, so if you’re concerned about indieauth.com going down then host your own on your own site. As for the auth providers, more can be added, including your own
# 18:12 tantek cweiske - I know we tried to support OpenID (consuming) in IndieAuth but ran into so many weird problems with OpenID providers that it became a support nightmare.
# 18:13 barnabywalters as I mentioned earlier, it was supporting openID, but there were a great many problemsn
# 18:13 tantek barnabywalters - I wonder if we could at least try to support self-hosted OpenID servers?
# 18:13 cweiske I can install my own indieauth server software, but nobody will use it
# 18:13 tantek since delegated OpenID servers we had problems with
# 18:14 cweiske because all the pages that use indieauth will use indieauth.com as form action url
# 18:14 barnabywalters tantek: with aaronpk’s shiny new system which allows us to choose the auth provider, an openID provider might be able to be added to the list
# 18:14 barnabywalters then, if it doesn’t work, the user could choose another provider and/or fix it
# 18:15 barnabywalters cweiske: whether or not to use indieauth.com is the choice of the application developer
# 18:15 tantek cweiske - indeed - your understanding sounds correct.
# 18:15 cweiske barnabywalters, that's my problem. no application devs will bring their own indieauth code but most will rely on indieauth.com
# 18:16 tantek cweiske - hopefully they'll support Web Sign-in directly
# 18:16 barnabywalters tantek: if you build your own, yeah — or run their own instance of the indieauth.com software
# 18:16 tantek cweiske - then the next challenge is how to have your own site be the OAuth provider in the RelMeAuth flow
# 18:17 tantek right now, every RelMeAuth / Web Sign-in implementation has a hardcoded list of OAuth providers that they can validate against
# 18:17 Loqi tantek meant to say: right now, every RelMeAuth / Web Sign-in implementation has a hardcoded list of OAuth providers that they can authenticate with
# 18:17 tantek so OAuth authentication endpoint discovery is the current challenge to this
# 18:25 tantek.com edited /IndieAuth (+768) "add issues section, document state of (lack of) OpenID support, reasons why it would be good (for self-hosted at least), next steps" (
view diff )
# 18:27 cweiske what do you mean with "provided by an indieweb server itself"?
# 18:28 cweiske 1. webapp, 2. indieauth, 3. openid discovery, 4. openid server, 5. indieauth, 6. webapp
# 18:46 cweiske I've also looked at "webmention.org". It's basically the pingback spec without XML-RPC and without any definition of what must be supported
# 18:47 cweiske let pingback be as it is, and nobody likes xml-rpc anymore, but it has one thing: it's well-defined
# 18:47 cweiske the webmention.org "spec" only gives examples how return values may look like
# 18:47 cweiske it doesn't even specify hat json has to be supported
# 18:47 cweiske and what exact schema has to be used for the json return values
# 18:48 barnabywalters cweiske: it’s experimental and in development — all feedback gladly appreciated :)
# 18:48 cweiske a webmention server can implement some obscure SVG responses and still be a server completely complying to the spec
# 18:49 cweiske this is not what I expect from a specification/standard. I expect well defined requirements, schemas and as little as possible optional things
# 18:49 barnabywalters it is explicitly stated that the body of the response makes no difference — it is the same in pingback
# 18:49 cweiske this is a problem because my webmention library does not have a a gurantee that it can understand the webmentino server
# 18:50 cweiske how am I supposed to automatically handle errors?
# 18:50 tantek I agree that webmention should be better defined
# 18:50 cweiske the problem is that there is no format that all servers have to implement
# 18:50 tantek too fast cweiske - having trouble keeping up with your issues :P
# 18:51 barnabywalters cweiske: can you add some of these as actionable issues on the webmention github?
# 18:52 cweiske I recently thought a lot about trackback and pingback and remote notification procols, because I need my git-based pastebin software to notify each other when you fork a paste on a remote server
pfefferle joined the channel
# 18:52 cweiske in the end, I decided to go with pingback because reinventing the wheel does seldomly help
# 18:53 barnabywalters cweiske: this is why we’re all supporting pingback before webmention for our projects
# 18:53 cweiske and I as a developer don't care if the servers talk to each other over xml-rpc or via json, the lib handles it for
# 18:53 barnabywalters it’s well supported and defined, but pretty nasty. So, we’ll use it until we’ve got webmention to a more solid state
# 18:53 cweiske barnabywalters, can you list the nasty things in pingback?
# 18:54 cweiske don't tell me you think that "Content-Type: application/x-www-url-form-encoded" is pretty
# 18:55 barnabywalters cweiske: “just the transport protocol” — yes, this is why webmention uses the exact same semantics over a different transport protocol
# 18:56 cweiske so why don't you do "pingback 2.0" or 1.1 and only swap xml-rpc against the new hip stuff?
# 18:56 tantek barnabywalters re: " we’re all supporting pingback before webmention for our projects" - speak for yourself, I plan on supporting webmention *only*
# 18:56 tantek and getting pingback backward compat via pingback.me
# 18:57 tantek barnabywalters - I think aaronpk uses pingback.me instead of supporting pingback directly
# 18:57 tantek barnabywalters - but you said the key word "implemented"
tilgovi joined the channel
# 18:58 barnabywalters tantek: okay, so my phrasing was wrong. My point (that we are all supporting pingback, one way or another) instead of ditching a well defined standard for something new, still stands
# 18:59 tantek barnabywalters - the distinction is important, as it shows where implementation interest is, and where active development / advancement of the protocol is likely to be
# 18:59 tantek pingback support is just for legacy compat as far as I'm concerned
# 19:00 tantek it's insufficient to do good federated comments
# 19:00 barnabywalters tantek: I’ll be more careful with support vs. implementation in the future then :)
# 19:00 cweiske "insufficient to do good federated comments" - why?
# 19:01 cweiske pingback just tells me that my url was mentioned on another url
# 19:01 tantek cweiske - because pingback implementations all SUCK at display
# 19:01 tantek hence - crappy for comments, not good for comments
# 19:01 tantek hence pingback is insufficient to do good federated comments
# 19:02 cweiske the next layer for "good comments" is to do the in-reply-to and microformat stuff
bretolius joined the channel
# 19:02 cweiske to find out if it's a comment to display it properly
# 19:02 cweiske so you still can use pingback as notification mechanism
# 19:03 tantek pingback *by itself* doesn't do good comments
# 19:03 tantek which those of us implementing webmention are doing
# 19:03 cweiske pingback itself does nothing except to notify someone
# 19:04 tantek cweiske, right "does nothing except" = insufficient
# 19:04 tantek cweiske - no, pingback spec specifies bad display explicitly
# 19:04 cweiske then please tell me where webmention does it better
# 19:04 cweiske tantek, i'd like to see that line in the pingback spec
# 19:05 tantek so yeah, we have citations, examples, proof enough that pingback sucks for federated comments
# 19:06 cweiske examples are not part of a spec, only to illustrate it. pingback itself does not define anything about display
# 19:06 cweiske it just tells you what could be done, after pinging
# 19:07 aaronpk sort of, it hints at display such as "The server MAY regenerate the site's pages (if the pages are static)."
# 19:07 cweiske if you say: hey, we need a proper solution to display remote content we got pinged about. here is in-reply-to and microformats to mark it up"
# 19:07 aaronpk I don't see why pingback or webmention should specify how things are displayed
# 19:08 cweiske but saying that the notification protocol sucks because it doesn't handle display, then I disagree
# 19:08 tantek cweiske - the *notification* part of pingback is fine except for the xmlrpc baggage
# 19:08 tantek it's all the hints about display etc. that are crap and have deepended federated notifications for a long time
# 19:08 aaronpk pingback unfortunately has been irrevokabely linked with the way wordpress has chosen to surface the notifications
# 19:08 tantek no, pingback is whatever is in the pingback spec
# 19:08 tantek notification is part of what is in the pingback spec
# 19:09 tantek people have sections on their blogs titled "Pingbacks"
# 19:09 Loqi aaronpk meant to say: pingback unfortunately has been irrevokabely linked in people's minds with the way wordpress has chosen to surface the notifications
# 19:09 cweiske pingback = possible examples of pingback display in the spec?
# 19:09 tantek with horribly double-ellipsed nonsensical extracts
# 19:09 tantek pingback = everything in the pingback spec, including examples of display
# 19:09 cweiske so pingback is doomed beause wordpress doesn't do content extraction properly
# 19:09 tantek note - this is also a good example of why even "just" (bad) examples in a spec is more harmful than none
# 19:09 tantek because for a decade pingback display has been broken
# 19:10 aaronpk pingback is doomed because nobody bothered to try harder than wordpress at using it
# 19:10 tantek because the way pingback expresses display / content extraction is poor
# 19:11 cweiske so fixing the display with microformats is not a solution?
# 19:11 cweiske you need to reinvent the notification system to fix display?
# 19:11 aaronpk so is JSON technically, which is why it's not required by webmention
# 19:11 barnabywalters cweiske: 1) not reinventing, just same semantics over a lightweight transport 2) display is a separate concern
# 19:12 aaronpk webmention is the same mechanics of pingback, with URL encoding to transmit the data instead of ridiculous xmlrpc bulk
# 19:13 tantek but definitely IGNORE all the display examples/hints in the pingback spec itself
# 19:13 tantek the main advantage of webmention over pingback is dropping of XMLRPC baggage
# 19:15 aaronpk btw, i've got another use of webmention in the works that has nothing to do with comments on posts
eschnou joined the channel
# 19:15 aaronpk speaking of the display being a separate part of webmention/pingback
# 19:15 cweiske so pingback is doomed because it had examples and mentioned how you could display them, and because it uses xml-rpc
# 19:19 barnabywalters cweiske: may I ask why you are so bothered by people wanting to build a better notifications solution, whilst still supporting well-adopted technologies for BC?
# 19:19 tantek cweiske except that no one is building any new Metaweblog API implementations nor standards. And AtomPub also failed. Everyone is doing their own *proprietary* posting API. Twitter, App.net etc.
andreypopp joined the channel
# 19:20 tantek cweiske - I *do* want to document any/all issues you have with webmention so we can improve the spec. I'll go back through the IRC log and see if I capture a reasonable representation of what you said/asked.
# 19:21 eschnou tantek, did you share your post on HackerNews ?
# 19:21 eschnou tantek, thanks for the writeup, great that someone takes the time of documenting this !
# 19:21 eschnou if a few of us on this channel upvote, we could quickly get homepage :-)
# 19:21 tantek eschnou - and thank *you* for making history. I'm just documenting it for the record. ;)
# 19:22 aaronpk it takes 3 votes in rapid succession, but you have to upvote from the "newest" page, don't direclty link to it
# 19:22 eschnou well, i'm just writing code, as I understand it, you brought this group together and showed a path that I'm just following :-)
# 19:22 eschnou so thanks as well :-)
# 19:23 aaronpk don't share a link to the HN post, or your upvote doesn't count
# 19:23 eschnou aaronpk, yes, that is correct
# 19:24 eschnou also my friends can't upvote for me anymore :-)
# 19:24 tantek thanks eschnou - great to have you here. and keep on cranking on your coding!
# 19:24 eschnou but within this group it may work since we didn't upvote each others much in the past
# 19:24 eschnou I'll post and let you know
# 19:24 eschnou I'll upvote then :-)
# 19:26 aaronpk eschnou: if you enjoy reading hackernews, check out the ##hackernews channel on freenode!
# 19:26 aaronpk cool, now hopefully that it's on the front page it'll get a couple more upvotes from the HN community and it'll stay there
# 19:27 eschnou now we'll get Loqi twitter spam :-)
# 19:33 eschnou aaronpk, thanks for pointing me to the HN channel, excellent idea!
# 19:34 aaronpk you just missed tantek's article being posted there!
# 19:48 cweiske Q: "But what if multiple people are behind the same domain" - A: "People figure it out, this is not a domain-specific issue."
# 19:49 cweiske my mom, dad, wife, kid and other relatives all have mail addresses on my domain. they do not want to put "my homepage is cweiske.de" in their *cough* github profile
# 19:49 cweiske they don't have any of those providers that are currently supported in indieauth
# 19:49 cweiske what they have is an email address on my domain, and automatically an openid
# 19:50 eschnou cweiske, indieauth used to support openid, I think it being removed is temporary, more a code issue than a philosophical one.
# 19:50 cweiske and you're telling me that adding a link attribute (rel="me") is harder than adding a <head> link? come one
# 19:50 Loqi cweiske meant to say: and you're telling me that adding a link attribute (rel="me") is harder than adding a <head> link? come on
# 19:51 eschnou cweiske, no, I think it was an issue on the indieauth side of processing the openid flows etc.
bretolius and andreypopp joined the channel
# 19:57 cweiske all in all don't understand why you think that new specs (that have much less implementations than the older specs) will take off where the old ones failed (indieauth vs. openid, webmention vs. pingback)
# 19:58 aaronpk don't know if you noticed, but each of those is an attempt to improve what has come before, by using as much existing tech as possible
# 19:58 eschnou cweiske, I wasn't there when you started this discussion, but what I can say is that i love this movement exactly because they don't try to invent new things, or at least try :-)
# 19:59 cweiske no, I didn't really notice. It mainly looks like bikeshedding and NIH
# 19:59 eschnou my two previous attempts (onesocialweb, ostatus) where both complex new things that in the end no one cared about and were way too complex.
# 20:00 eschnou here they just use the gool old web, good old markup, good old pingbacks, seems good old enough to me :-)
caseorga_ joined the channel
# 20:01 eschnou cweiske, indeed, I haven't looked at webmention yet, the federation we got today is based on pingbacks
# 20:02 eschnou it seems webmention is simpler to implement, so I'll definitively have a look and support both if it can help other implementers.
# 20:02 eschnou no, great to have someone challenging some ideas :-)
# 20:03 eschnou what I like in this community is also the positive mindset of just do it and iterate, vs debating on specs forever and never implementing them.
# 20:08 eschnou aaronpk, thx :-)
# 20:09 aaronpk cweiske: out of curiosity, how did you find out about indiewebcamp?
# 20:10 cweiske my federated feed reader (tt-rss) told me about it today
# 20:11 cweiske and since i'm all for self-hosting, I looked around
# 20:25 eschnou aaronpk, I like your idea of whitelisting approved domains
# 20:25 aaronpk the whitelisting could even be done automatically, based on domains that you've included in your own posts
# 20:26 eschnou aaronpk, absolutely, so many cool ideas to implement !
# 20:27 eschnou aaronpk, yeah ! great to have a few different implementations to experiment with
# 20:28 eschnou fyi: 10 simulataneous users right now on my site, hasn't dropped since we are on HN. I wonder how many visits tantek will have.
b0bg0d joined the channel
# 20:28 aaronpk wow that's impressive for a link off the main post
tantek joined the channel
# 20:34 aaronpk tantek: the HN post is holding strong on the front page! you should check your google analytics
nsnow joined the channel
# 20:42 tommorris aaronpk: we can make the comment thread angrier. just need to start mentioning women in tech/diversity, Rails scaling issues etc.
pfefferle joined the channel
jfranusic joined the channel
pfefferle and barnabywalters joined the channel
# 21:18 cweiske most interesting thing is probably the diary (tagebuch)
# 21:29 cweiske but I still use some, e.g. a tool to administer playlists on my UPnP server
# 21:46 aaronpk cool! @wmf replied: "The point of this Federated Indieweb thing is to move comments (and all other content) out of the Web 2.0 silos like Facebook and Disqus because those companies aren't trustworthy stewards of our data."
# 21:53 tantek aaronpk, barnabywalters - what formats do you "Accept:" in you webmention endpoint?
# 21:54 tantek rather than making it open ended / HTML / JSON
# 21:55 barnabywalters cweiske: problem with JSON is that it may be just like XMLRPC in that it’s all trendy now, but in a few years not so popular
# 21:56 aaronpk could just go with form encoded responses I suppose, although that's kind of weird
# 21:56 tantek webmention is a protocol - it should probably stick to the headers
# 21:56 tantek aaronpk - how about you? what kind of "Accept:" does pingback.me support for webmentions?
# 21:57 aaronpk so far p3k doesn't handle its own, but the plan is to make it support webmention natively and use pingback.me to translate pingbacks to webmentions
# 21:57 barnabywalters the conneg idea currently in the webmention spec is cool, but probably overengineering
# 21:57 aaronpk i believe pingback.me only responds with JSON right now
# 21:58 tantek so we should either (a) pick JSON and stick with it, or (b) specify precisely how its done with HTTP headers and nothing else.
# 21:58 aaronpk it returns HTTP 202 for success, and 400 for all failures
LukasM joined the channel
# 21:59 aaronpk how would you indicate the different types of error cases in an http header?
# 22:00 barnabywalters aaronpk: probably similar to the current status header, as the fail conditions are all number + message
# 22:01 aaronpk interestingly, if there's an error, there isn't much the sender can do
# 22:02 barnabywalters aaronpk: yes, which is one of the reasons I don’t see the handling of the response body as particularly useful
# 22:04 tantek right, if the fail conditions don't result any different handling, then why deal with a response body?
# 22:05 tantek barnabywalters - nicely done with including context on your reply posts!
# 22:07 tantek can your software do the right thing for you?
# 22:08 tantek and then automatically change your reply post accordingly
andreypopp joined the channel
# 22:09 tantek and if that h-entry has a rel=syndication link back to the tweet
# 22:09 tantek then congratulations, you've found the original
# 22:09 tantek use the original post in your reply post on your own site
# 22:09 tantek but then when you POSSE your reply to Twitter, make sure the POSSE reply's in-reply-to-status-id links up to the tweet
# 22:10 tantek that way your indieweb reply post will thread with the original indieweb post
# 22:10 tantek and your POSSE'd reply will thread on Twitter with the POSSE'd copy of the original
# 22:10 tantek providing a threaded experience for users of either of your indieweb sites/posts
# 22:12 tantek btw, any time you notice that you (or anyone) is doing something "out of force of habit" - make a note of it! that's likely something important to design for.
b0bg0d joined the channel
# 22:13 tantek I'm going to document it as part of original-post-discovery
# 22:14 tantek because the idea is that when you find (a copy of) a post on the web, you want to be able to automatically find the original
# 22:14 barnabywalters apart from possibly checking for the … or () POSSE permalink patterns, and your own permashortid pattern, the pseudocode looks complete
# 22:19 barnabywalters btw tantek, any plans to include a p-author h-card on your articles and notes?
# 22:20 barnabywalters I did a local test of reply context with one of your articles and it picked up the title fine, but no photo or author name
# 22:20 tantek barnabywalters - very much so - in the middle of redesigning them accordingly
# 22:20 tantek my permalink pages need to follow my own advice and use the <body> or <html> element for h-entry
# 22:20 tantek actually this whole discussion about how to POSSE replies to Twitter forced me to go down the redesigning how my posts look rabbithole
caseorganic joined the channel
caseorga_ and b0bg0d joined the channel
# 23:00 tantek rel canonical is the right meaning for discovering an original post from a syndicated copy right?
# 23:34 tantek going with rel-canonical until/unless someone comes up with a better idea (how POSSE'd copies should link back to their original posts)
# 23:51 Loqi Ok, I'll tell him that when I see him next