snarfed, caseorganic and benjaminchait joined the channel
#tantekthanks snarfed - that's a good capture / start
#tantekyou're right, one could receive them in either order
#tantekI wonder if it makes sense for a webmention receiver to keep ahold of the dups as little mini-links in their UI - e.g. (also on Twitter, also on Facebook) but maybe as tiny icons instead of words, e.g. (also on t f)
#tantekanother approach could be to *always* attempt original-post-discovery on webmentions, and then even if you *only* receive a Bridgy webmention of a POSSE tweet copy, go use the original post instead of the link to the Bridgy permalink.
#snarfedre cluster links, sure! similar to visible rel-syndication links
#tantekthat way the POSSE reply -> Bridgy -> backfeed webmention essentially becomes a circuitous but automatic webmention for the original!
#JeenaA couple of people asked if my software was open source too already. (it isn't yet)
#tantek.comedited /webmention (+722) "/* Issues */ always do original post discovery on the source URL as a means of preferring original posts and automatic de-duping among replies and POSSE copies" (view diff)
#tantekaaronpk - captured the "prefer original posts" way of handling potential de-duping
#caseorganicJeena: i'd love to see it. i joined late yesterday and some discussion about what you built
#tantekJeena - how about open sourcing at least bits of it?
#caseorganic.comedited /site-deaths () "(-2675) /* Upcoming */ Removed site deaths from "Upcoming" that had already passed. Added Incredible Labs acquisition by Yahoo!" (view diff)
#tantekoh no - don't remove them! verify them and archive them by year!
#caseorganicis thinking of going to indiewebcamp sf
#caseorganic.comedited /2014/SF (+356) "/* Regrets */ Added Caseorganic.com to regrets list as IndieWebCamp SF conflicts with Esri Dev and Partner Conferences" (view diff)
#jonnybarnesif a micropub request is also syndicated to twitter, should that be part of the response from the micropub endpoint? Like a second twitter-location header?
#barnabywaltersjonnybarnes: my instinct would be no, mark it up with u-syndication in the response body
#jonnybarnesso a micropub endpoint response should be full HTML response?
#barnabywaltersjonnybarnes: what UI do you want to build which demands it not be?
#barnabywaltersone of the reasons micropub is based on form-urlencoded data instead of e.g. JSON or HTML in request bodies is so that it can be tested and implemented using simple static HTML forms
#barnabywaltersso it makes sense for the response bodies to also be HTML
#jonnybarnesI've only just got micropub somewhat implemented in my site, so still ironing out kinks. But lets say you log into my site and post to yours, at the moment my site just takes the Location response from your micropub endpoint and redirects you to it
#jonnybarnesI suppose itd then be exposed in your UI if your site succesfully syndicated to twitter
#barnabywaltersjonnybarnes: yeah, that’s up to the post-creating site to do
#barnabywaltersand if the micropub request was happening from, for example, an iOS app, it would need to parse the response for u-syndication links if it wanted to indicate syndication success
#jonnybarnesso am I doing it wrong? if you were to log into mine, the form points to my own site, and my site then makes the API request to yours
#jonnybarnesaaronpk said you were pointing the form straight at the micropub endpoint
#barnabywaltersI don’t think there’s anything explicitly wrong with that
#glennjonesInteresting, I have not followed micropub, but have gone down the JSON route and built a full CRUD API that takes JSON documents for my site.
#glennjonesI should try and open it up and document it on the wiki to contrast the two
#jonnybarnesso if you logged into my site, your micropub endpoint would create the response with relevant correctly markup'd response, if I logged into your site then my micropub endpoint would make the relevant correctly markup'd response
#Loqijonnybarnes meant to say: the hard part isnt the webmention part itself, its how do I store these mentions and how do I displa them
#barnabywaltersjonnybarnes: storage is an annoying problem to have
#barnabywaltersbut the fact that figuring out how to display them is the only other problem is *good* — it means the protocol is simple enough that you can concentrate on making a good experience
#jonnybarneswhat'd be cool is if you could have a conversation taking pace on your site where the comments are coming from different sources, some facebook and some twitter for example
#Jeenahehe yeah, why not, but I can see all my attempts here, damn, It didn't work well. Now I moved away from apache and passenger to nginx and puma via unix sockets which works much better for my rails app
#jonnybarneswhy does aaronpk have a native webmention endpoint, but then send pingbacks to webmention.io? is webmention.io turning pingbacks into webmentions for him?
#barnabywaltersjonnybarnes: yep, it’s converting ping backs into webmentions so he doesn’t have to parse XMLRPC requests
#pfefferleno... I want to bundle all plugins into a single one, without rewriting or changing the code... and I need a way to prevent errors if someone installs the single plugin and the bundle...
#Jeenahm but as far as I see it is written like this because all those WordPress hooks like add_filter, add_action, etc. need a function name as a string? (omg)
#pfefferleJeena you can also pass array("ClassName", "functionName");
#Jeenato leave out <html>, <head> and <body> is ok
#barnabywalters“If it's reasonable for the Document to have no title, then the title element is probably not required. See the head element's content model for a description of when the element is required.”
#glennjonesI am sort of getting around some display-requirements by not using there API. My written on my site are sync out from my site or import from a twitter archive file
snarfed joined the channel
#Jeenaah this is interesting, I didn't know that but that is quite cool actually. How often do you import from the archive file?
#aaronpkJeena: it's looking more and more like this is some problem with my server, like an outdated list of root certs that I can't figure out how to update
skinny, scor, pasevin, paulcp, pasevin_, Acidnerd and snarfed joined the channel
#JeenaRyan Barrett send a webmention to https://jeena.net/indieweb but it kind of went wrong a bit and the url goes to his homepage instead to the note and I can't find the note on his website
#Jeenait happened once already a week ago or something
#Jeenasnarfed could you give me the URL to your note so I can change it in my database?
#EHLOVaderbasically it would be neat to offer a signing service built into it, to include the ssl inside the ansible itself
#EHLOVaderso I guess it would also have to be MIT license compatible
benjaminchait joined the channel
#EHLOVaderI am ehlovader everywhere, github, bitbucket, gmail... etc if you wanted to not share it in the public but don't mind me seeing it and using it/basing other code on it
#EHLOVaderI could even PR any improvements back to you if you want, I will probably need more explicit errors and return values for the signing part
#aaronpki'd be happy to. i unfortunately didn't give myself a writeup of how I did it, but I do have a pile of files I could share with you
#tantekalso - I wonder if better than sending a webmention for the fave, would be sending a webmention for the tweet that *linked* to the blog post - as that's more like a "comment"
#aaronpkpretty soon brid.gy will be sending webmentions all over the place
LauraJ joined the channel
#snarfedtantek: yeah, i think we've discussed that before. i was reluctant because i wanted to avoid sending for POSSE posts, and my initial survey a bit ago saw that rel-syndication adoption was low
#tanteksnarfed - but if we put the burden of original post discovery of replies on the receiver, then it should be fine for Bridgy (or anyone) to send webmentions of POSSE replies.
#snarfedright. needs adoption, similar to rel-syndication. i made the initial design choices based on what most people had implemented at the time, and what would give the best signal/noise backfeed
#snarfedi can definitely evolve those as people make their mention handling smarter
#snarfedglad these ideas are on the wiki, that's a first step!
#tanteksnarfed - your implementation could also motivate such updating
#tantekand then have original-post-discovery resolve it
#Jeenaand I will not send webmentions with the one with /posts/93 anymore
#tantekshall we add rel=canonical resolution to original-post-discovery ?
#Jeenatantek btw. why are you not showing webmentions for your notes? Is it just because you didn't get around implementing it or is there something more?
#barnabywaltersif we put rel=canonical discovery in original-post-discovery, it either needs to go after h-entry u-url lookup, or an exception given for twitter.com
#barnabywaltersas tweets have rel=canonical pointing to themselves
bnvk and skinny joined the channel
#barnabywaltersor maybe there’s a silo-independent solution by checking for loops…
#tantekwhere you click on the person first, and then (optionally) choose a protocol that you're connected over
#tantekthe other big omission (from that blog post and Windows Mobile UI) is the TWO PART nature
#tantekWindows UI assumes the person initiating the message/call makes ALL the decisions about which protocol/app to use
#tantekwith the indieweb solution I propose, it's two part, the receiver (person with the indiewebsite), and the initiator (the person going to the indiewebsite)
#Jeenahm not if the person who is called does not use a specific protocol
#tantekJeena - this is all described in the blog post
#Jeenahehe yeah my parents have a paper address book and there under my name they have about 8 phone numbers where all but one is outdated ^^
#tantekbut even then - it's ideal if *before* contacting someone your device gets the latest methods of contact (and in order) from the other person (i.e. their website0
#Loqij12t: tantek left you a message 2 minutes ago: aaronpk and barnabywalters have some solutions to that
#tantekj12t - they just do it on their sites - they mention domains in the posts on their site, and then their POSSE code automatically maps those personal domains to @-references etc. on Twitter etc.
#tantekI think barnaby wrote about his approach but I don't remember where
#aaronpki use the rel-syndication target of the original post to find the person's twitter name
#tantekj12t - there are existing libraries/services for going from a personal domain to various silo profiles
#j12tThat's cool. Would be nice if everybody could use the same code there.
#aaronpki want to do home page -> twitter name, but haven't implemented that yet. I think barnaby may have
#tantekj12t - glennjones (sometimes in the channel) wrote a bunch of awesome XFN based silo profile discovery code
#tantekcan't remember the name of the glennjones's site / service for this
#j12tSo then the "ideal" behavior is that on my own site, I use domain names as identifiers, and when things get syndicated out and back in, it automatically translates back and forth into the appropriate silo names.