2014-04-30 UTC
# 00:01 aaronpk yeah I'm doing a rudimentary check at that point before actually parsing the page.
KartikPrabhu joined the channel
# 00:07 snarfed kylewm and i were talking the other day about how to get bridgy to find your notes feed
# 00:07 benwerd I'm not quite sure what the best thing to do is there, tbh
# 00:07 snarfed maybe if we follow multiple rel-feed links, or whatever they are
# 00:07 benwerd hmm - but I've got like 8 feeds, so that could get heavy on both of our ends
# 00:08 kylewm haha, tantek was surprised that I was using rel=“feed” and not rel=“alternate”
# 00:08 snarfed benwerd: eh i'm fine with it following multiple, we should be able to optimize on our end so it only does it once per silo post
# 00:09 kylewm snarfed: do you mean fanning out to read multiple feeds?
# 00:09 snarfed kylewm: well, multiple rel-feed links i guess, yes
# 00:10 snarfed i'm happy to let you all drive the design decisions (and review), just wanted to confirm that backfeed on benwerd's notes is within reach :P
emmak joined the channel
jonnybarnes joined the channel
# 00:11 aaronpk huh, I didn't know about the "onerror" attribute of <img> tags
# 00:13 kbs I pretty much just went down the cheat list from owasp - keep learning some new dark corners each time I poke my nose there
paulcp and _6a68 joined the channel
# 00:17 aaronpk this stuff is so hard to figure out without actually implementing it
# 00:17 KevinMarks_ hm, the only reason I know facebook's f8 is on is a friend posting photos of a video uplink truck, and a journalist coming into SFO
# 00:18 aaronpk i think I should go build a token endpoint real quick instead of trying to figure it out on the wiki
# 00:19 kylewm snarfed: just realized that I check for rel=feed but don’t check the type is html … that’s not great
# 00:20 snarfed kylewm: feel free to use util.follow_redirects, it checks for that and a few more nice things
# 00:21 snarfed aaronpk: yup. you may bask in my amazing coolness
# 00:22 kbs wouldn't it be neat if they added fragmentions on facebook
# 00:23 bear hmm, how do I know if my check webmention result is good or bad?
# 00:23 kbs bear: unfortunately, have to look back on your page :/
# 00:24 kbs wasn't clever enough to recheck the page, alas :)
# 00:25 kylewm snarfed: (re: follow_redirects) phew thanks, had no idea how to do it otherwise… since rel-feed doesn’t have to have a type
# 00:25 snarfed kylewm: well, it cheats and makes a head request, and then depends on the content-type http header…but good enough
scor joined the channel
# 00:26 kylewm is that cheating? it kinda seems like the only way to do it :)
# 00:27 bear kbs - so i'm thinking that since the /xss left a mention that it failed?
# 00:28 KartikPrabhu gRegor`: might also want to add a "required" attribute so the modern browsers can save a request incase there is no source ;)
# 00:28 kbs bear: so in theory, one of the mentions should succeed, and the other (if not correctly stripped) leaves behind javascript that displays alerts and so forth
# 00:29 bear then i'm "safe" for now because I pull in no text - just hcard items
# 00:29 bear but this is a nice tool to have - great job!
# 00:29 kbs I think there's one more test that I haven't added in (adding js within any hrefs within the hcard itself)
# 00:29 gRegor` It should give you an error message if you submit the form empty, KartikPrabhu
# 00:30 KartikPrabhu gRegor`: yes it does, but only after a request has been sent to your server. If you set "required" on the input, the browser will do it for you :)
# 00:31 snarfed kbs: sounds like you could attack bear by putting xss etc inside h-card data?
# 00:32 aaronpk considers making a test token endpoint on something.oauth.net lol
# 00:32 bear yea, because I just pull in the data without any checks
# 00:32 kbs snarfed: right, it's a missing test-case indeed. *goes back to add* :)
# 00:32 bear @snarfed no worries - this is the kind of thing that should be pointed out - so we can give good examples
# 00:32 gRegor` Ahh. Nice. Didn't know about that.
# 00:32 gRegor` KartikPrabhu++
# 00:33 KevinMarks_ tokenendpoint.oauth.net ? or does the official name have underscores in?
# 00:35 gRegor` Google sent me there already. :)
# 00:36 gRegor` I'm thinking of changing the subheadings. Adding "Respond" before the webmention form and changing "Add Comment" underneath to "Or Add a Comment"
# 00:38 KartikPrabhu gRegor`: I recently changed my webmention to read "send me the link to your response"
# 00:38 gRegor` How did we develop before Firebug and Chrome Web Inspector? Seriously.
# 00:38 gRegor` I noticed that.
# 00:40 gRegor` So bossy. "send me the link! do it now!" jk
# 00:46 kbs *waves* thanks for the feedback on checkmentions
fmarier, b0bg0d___ and hadleybeeman joined the channel
# 00:58 benwerd kylewm: are the header links I just added to my site any good to you?
# 00:59 kylewm benwerd: that’s perfect, except your twitter profile links to benwerd.com instead of werd.io
# 01:00 kylewm (the bridgy thing uses your user profile url to find the h-feed)
# 01:02 gRegor` Loqi should have "annnnd boom goes the dynamite" in its vocabulary
# 01:03 kylewm benwerd: oh darn, it will only work for new posts
# 01:12 kylewm oh… this is bad, bridgy still thinks your domain is benwerd.com
# 01:13 kylewm i suspect it’s cached from when you signed up
# 01:21 kylewm maybe a good feature request for me would be a way to reset your cached syndicated post relationships
# 01:23 kylewm also trying to reply to an old post that didn’t have any previous responses
# 01:29 kylewm sorry kind of spamming your twitter right now
lukebrooker_ and snarfed joined the channel
# 01:37 kylewm benwerd: posse-post-discovery didn’t find it :( I think there is only one piece missing (famous last words)… it looks for the permalink to a post with h-entry -> u-url, and it doesn’t look like yours have u-url
# 01:39 benwerd would you agree that it should go on the anchor tag here?
# 01:39 benwerd <h2 class="p-name"><a href="http://werd.io/2014/looking-forward-to-iiw">Looking forward to #IIW</a></h2>
# 01:40 kylewm I would probably put it on the link at the bottom of the post (the one with the time)
# 01:42 benwerd oh, oops - that was already committed, but I hadn't updated the template on my site
# 01:43 kylewm nice, so I suspect that will work on future posts
npdoty, snarfed and lukebrooker joined the channel
lukebrooker, dybskiy and fmarier joined the channel
dybskiy and fmarier joined the channel
gRegor` joined the channel
tantek joined the channel
# 03:49 KartikPrabhu ok I wrote a response to Joschi Khupal's article. It is online and accepting webmentions but will be public on 2014-05-02. Feel free to openly add any sort of responses through webmentions to it. Link: http://kartikprabhu.com/article/concurrence-cooperation Also try, using fragmentions to reply to specific phrases, I'll play around with implementing that correctly too.
# 03:55 aaronpk KevinMarks: should I set up a copy of indieauth.com at indie.oauth.net? lol
# 03:59 aaronpk and then use that as the address I delegate my OpenID to?
# 04:05 bret i bet the association would bring a lot more attention to it
# 04:05 bret most people I meet know oauth, most dont know much about indieauth
# 04:06 bret aaronpk maybe the github pages thing should be fixed first?
brianloveswords joined the channel
# 04:14 KartikPrabhu thanks! I think it is more aggressive than what others here responded with. Feel free to add disagreements with it using webmention :)
# 04:17 aaronpk KartikPrabhu: I'm not sure I understand the different levels of indentation
# 04:17 bret aaronpk rings kind of close to myopenid
# 04:18 KartikPrabhu aaronpk: eh! it is mostly for some layout funkiness, no specific meaning. :P
# 04:18 aaronpk bret: heh I don't know, the whole openid / oauth thing is kind of a disaster anyway
# 04:18 bret aaronpk imagine there are 3 energy levels...
# 04:21 KartikPrabhu aaronpk: I usually pull out the introduction/conclusion of a section if it has some sort of standout point I'd like to make
# 04:21 bret Personally i think the wake up call to people talking about how we need open versions of silos or federated this or thats on a silio is worth while. What is the simplest thing you can do that others have already shown possible!?
# 04:22 bret just a bummer when people confuse that for something else
KevinMarks_ joined the channel
rektide and gRegor`_ joined the channel
snarfed, gRegor`__ and kbs joined the channel
j12t joined the channel
legastero joined the channel
# 05:27 aaronpk now that I'm actually trying to implement a token endpoint
# 05:29 aaronpk the main problem is the token endpoint gets a request with arbitrary "me" values, so has to first go find the authorization endpoint for the person
eschnou joined the channel
# 05:30 aaronpk then, for the JWT case, it has to go fetch the public key from the authorization endpoint. or for non-crypto version, makes a POST request to the auth endpoint to verify the code
bnvk and wraithgar joined the channel
# 05:34 aaronpk so I don't really see the benefit of the crypto version
friedcell, friedcell1, LauraJ, cweiske, voxpelli, eschnou, tobiastom and krendil joined the channel
friedcell joined the channel
jonnybarnes, carlo_au, erikmaarten, voxpelli, cweiske, v0_, ttepasse, barnabywalters and eschnou joined the channel
carlo_au, scor, eschnou, chloeweil, v0_ and ben_thatmust joined the channel
carlo_au joined the channel
# 13:17 GWG barnabywalters: Did you see what I did last night to try and fix it?
# 13:18 barnabywalters ah yeah I like that bio better — having the elsewhere icons under your name associates them more clearly
# 13:19 GWG Yes, but if you view the page on a smaller screen, they pop back into the old position
# 13:19 barnabywalters it might balance things a little better if there was some whitespace between the header and the body/bio
# 13:20 ben_thatmust i need to completely redo the front end of my site. I have a lot of stuff to do
# 13:20 GWG barnabywalters: That side bio is hard coded, which is why it is temporary.
# 13:21 GWG barnabywalters: Not sure what you mean.
# 13:22 GWG I'm confused about where on the page we are talking about
# 13:23 GWG Is the top one what you are seeing?
# 13:24 GWG Because my version looks more like the bottom one?
# 13:25 GWG I have Firefox and Chrome installed. And both show a space between the header and the top of the content boxes
# 13:26 GWG ben_thatmust, can you have a look?
# 13:26 GWG Since you are here? Do you see a space?
# 13:27 GWG barnabywalters: Good. I was worried there for a moment.
# 13:27 GWG I spent a lot of time looking at people's bio bars
# 13:28 GWG For example yours and aaronpk's have commonalities in content
# 13:28 GWG Although you align left, he aligns right
# 13:29 GWG A lot of people put the same info as a content area on top
# 13:30 GWG A lot of design for me is looking at other people's sites and saying... "I like this...don't like that."
# 13:33 GWG So, I like having the location. So I'll likely be adding that. But I don't like the idea of being too precise about it.
# 13:34 GWG I also may add an optional location parameter to the posts, which I don't have now. But it is only useful when I'm somewhere worth mentioning.
# 13:37 GWG ben_thatmust: Never mind, it was barnabywalters caching, not the site.
# 13:40 GWG barnabywalters: indiewebify.me is reading the markup for one of the articles instead of the one in the sidebar. Will have to work on that somehow.
tobiastom and v0_ joined the channel
# 13:54 GWG Wordpress assumes multiple authors, and I coded for it as an optional feature, but hid the display. May have to rethink that
eschnou joined the channel
# 13:57 GWG barnabywalters: When there are multiple h-entrys on a page, such as a home page or h-feed, should each have an embedded h-card?
# 13:58 barnabywalters or, as a I do on my homepage, enclose authorless h-entries in an h-feed with an author h-card
# 13:59 GWG barnabywalters: There are 3 options I see. Strip the h-card on the home page only, and keep it on the single page
# 13:59 GWG Strip the h-card entirely as there is a big one in the sidebar
# 13:59 GWG keep the multiple h-cards but mark it up differently
# 14:01 barnabywalters the way I would do it on your site would either to mimic what I do (h-feed with author h-card and child h-entries) or what aaronpk does (flat page with h-entries and h-card on)
# 14:02 GWG Well, that will be something I fiddle with tonight.
snarfed and kbs joined the channel
# 14:42 kbs aaronpk: a thought re yesterday's conversation on bearer tokens. I wonder if OTPs can be (ab)used to simplify the authorization endpoint
# 14:42 kbs and used to obtain a new bearer token.
# 14:45 kbs love to add any testcases you might have used/run-into as well
# 14:46 kbs aaronpk: user (say) signs into indieauth.com, which then provides a TOTP key for each scope of interest. (eg:) totp-key1-write, ... user installs these keys on their mobile device
# 14:46 kbs now, a third-party app/website comes along and wants to publish a comment on behalf of user
# 14:47 kbs now the mobile-app requests the code for a totp-key1-write from the user, submits it to indieauth.com, and gets back a signed token confirming it has the ability to write to the user's site
# 14:48 kbs signed token == bearer token, used/verified with each call to the user-site as usual
# 14:53 snarfed btw kbs, i'm going to (periodically) clean up the test mentions on the polytics post. holler if you need any, i can always restore them
# 14:55 kbs snarfed: :) thanks much - appreciate that you let me test with your page, but don't need them permanently. I'll also give you a heads up before testing again -- should be done for the moment.
gRegor`, tilgovi, KevinMarks and v0_ joined the channel
# 15:22 kylewm I just tried checkmention ... had a false sense of security cause I thought I was using python bleach everywhere
# 15:22 kylewm but apparently I was not using it in the most important place!
# 15:23 kbs oh, cool - glad it's helping! I don't use python, so not familiar with the quirks of commonly used libraries. If you notice corners where things could be tested more, please do let me know
snarfed joined the channel
# 15:26 kbs I need to add a rel=me test link, somehow forgot to add that in the last change. (snarfed also uses nofollow everywhere when embedding links, which also seems like a good idea...)
# 15:29 kylewm cweiske: even for links in comments on your site?
# 15:31 kylewm I'm not sure, in the indieweb, it makes sense to me that they would get a small bump from being on my low-traffic site, but should not get a big bump for me commenting on a higher traffic blog
KevinMarks joined the channel
# 15:36 kbs oh, neat - didn't realize both of you were self-hosting your git repositories too
# 15:36 cweiske self-hosting your blog isn't as important as self-hosting your code
# 15:37 kbs ha - that's pretty neat :) I like it
# 15:39 cweiske version 0.4 even sends webmentions/pingbacks if it forked a remote site
# 15:40 kbs this is way more advanced than I realized :) neat
# 15:42 cweiske I really need to get the .phar version going, so that people can try it out easily
v0_ joined the channel
# 15:48 GWG I ripped out the part of my page you commented on.
catsup joined the channel
# 15:53 GWG On mobile, the top social bar reappears
npdoty joined the channel
# 15:56 GWG Needs more work though. I have a lot of unused space right now
# 16:01 GWG Yes, but I think I could add more in the sidebar.
# 16:01 GWG I may not like it and take it out
# 16:02 GWG I may try a local time display like some sites
KevinMarks and KevinMarks2 joined the channel
_6a68 joined the channel
j12t joined the channel
brianloveswords, ryana and dybskiy joined the channel
brainTrain joined the channel
# 16:53 bnvk KevinMarks: do you know if the Hypothes.is people plan to allow sites to host their own copy of data?
# 16:54 KevinMarks not sure - they are keen on the standardisation process, and seem to be open sourcing bits
brianloveswords and snarfed joined the channel
# 17:02 aaronpk would rather switch to the tab and scan some text periodically
# 17:06 kbs KartikPrabhu: dunno, do you think it might restart another round of arguments? :)
caseorganic joined the channel
# 17:07 KartikPrabhu it definitely is more aggressive than how others here have responded.
# 17:08 kbs KartikPrabhu: okay :) the 0.02 that struck me, is that Joschi seemed (to me) to be saying that his notion of indieweb includes people who don't actually build things
# 17:08 kbs so there's a chance that both of you might be right in different ways.
benwerd and snarfed joined the channel
# 17:13 snarfed hmm. now users have to opt into giving an app access to a post, etc, *per post*?
# 17:13 snarfed i'm inferring from what he just said. very unclear so far. we'll see
# 17:14 snarfed eh, i actually like stability better. fixing is ok but better to try to prevent the problems in the first place
# 17:14 benwerd curious about what the line about not giving access to stuff your friends share with you means
# 17:15 aaronpk "The Anonymous Login feature allows users to try apps without sharing their identity right away."
# 17:15 snarfed fb's early culture very much glamorized firefighting, which is a very specific culture. (imho not great for eng)
# 17:15 snarfed aha, ok. if he was talking about anonymous login, then no, bridgy should be fine
# 17:15 snarfed lots of goog engineers i know who went to fb hated it and left fast
# 17:16 snarfed since google was much more engineering, less hacking
# 17:17 kbs snarfed: great post, nodding my head a lot while reading it :)
# 17:19 snarfed heh. people said the same thing about open graph a few yrs ago
# 17:19 j12t Did you guys have any doubts about that?
# 17:19 benwerd and wasn't that kind of the implication with Google+ too?
# 17:19 KevinMarks I'm reminded of the "decentralised" slide at F8 that had FB in the centre
paulcp joined the channel
# 17:19 snarfed benward: g+ was maybe more about data/interop standards
# 17:20 benwerd I always assumed they'd attempt to extend the reach of circles to content on the web
# 17:21 KevinMarks I'm sure they wanted to be cross silo, but the other silos said nope
# 17:23 benwerd oh! just realized that j12t will be at IIW as well
# 17:24 benwerd They've made space for an Idno demo - snarfed, cool to show them Bridgy while I'm at it?
# 17:24 snarfed wish i could be there. congrats on getting backfeed working w/o links btw. kylewm did great work there
# 17:25 j12t No. But I don’t think I missed more than 2.
# 17:25 gRegor` KartikPrabhu: The paragraphs seem oddly aligned in that post
# 17:26 gRegor` I think it's a good post, though.
emmak joined the channel
# 17:27 gRegor` Footnote 2 has a run-on sentence, if you wanted grammatical feedback. :)
# 17:27 gRegor` I like the use of "learnt" :)
# 17:28 snarfed aha, KevinMarks et al, here's the f8 part you all will be interested in
# 17:30 snarfed this kind of thing will help tantek's personal comms thing a lot
# 17:31 aaronpk "touch to return to Quip" <- is that part of the applinks sdk?
# 17:31 snarfed aaronpk: not sure. i haven't seen this presentation, just got to use the api a bit
# 17:32 snarfed they have a service where you search for an app and it returns the ios url to use, the android intent to use, etc
# 17:32 Loqi Ok, I'll tell him that when I see him next
# 17:33 aaronpk wait so what is the new thing? the "return to" link at the top?
# 17:34 snarfed no, forget about the return part, the interesting part is finding the app link/intent in the first place
# 17:34 kbs how are applinks different from intents (functionally?)
# 17:34 snarfed the key part is the db that's queryable and kept up to date
# 17:35 snarfed we can do this now manually, but that's way brittle. this helps
# 17:35 gRegor` snarfed: bridgy doesn't have a published time on favorites, presumably because Twitter doesn't supply one, right?
# 17:35 snarfed sure, i'm ignoring ideology. the *functionality* is useful. if we like it we can go build our own clone that supports the same api, etc
# 17:35 gRegor` I got my first brid.gy backfeed. On a post from 2006. :) Only because I linked it in a tweet to someone, heh
# 17:36 gRegor` I think I will match up the twitter profile URL between the like and the comment on the same post, so I can use the comment's published time.
# 17:36 KevinMarks it's not so much getting stuck in a mobile browser as getting stuck in a webview on iOS
# 17:37 Loqi Ok, I'll tell him that when I see him next
indie-visitor joined the channel
# 17:38 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
# 17:38 kbs gRegor`: aha, nice :) thanks for the pointer
# 17:49 Jeena I am thinking of adding pictures or images or something like that to my blog/ and notes/
# 17:50 Jeena but I can't quite decide on how to call it, pics/ images/ imgs/ pictures/ pix/ look all funky
# 17:52 aaronpk Jeena: I also couldn't decide on any of that, or how I even want to host my photos, so I punted for now in favor of an instagram-like stream of photos in my notes
# 17:53 Jeena hm yeah, I'm more in favour of a flickr photostream with only pictures
snarfed joined the channel
# 17:55 Jeena Perhaps I just should call it pictures/ and just change it if I really feel that it was the wrong name
KartikPrabhu, _6a681 and kevinbae joined the channel
dybskiy and caseorganic joined the channel
# 18:30 bret cybe? I mean i cant even think why we would
# 18:31 bret i can only see this used as an example of how NOT to build projects
# 18:31 bret step 1) only talk about goals completely out of reach
caseorganic joined the channel
# 18:38 bret personal projects where the goals are clearly to lofty?
# 18:39 bret its somewhere between people who just talk/make materials to help them talk more
# 18:39 bret and projects that have such lofty goals that they are kind of insane
jonnybarnes and tilgovi joined the channel
paulcp and caseorganic joined the channel
# 18:50 bret feel free to add that to the wiki if you dont see anything else too similar
# 18:53 aaronpk yeah just go ahead and remove it ... not worth people spending time reading it, heh
iangreenleaf and eschnou joined the channel
# 19:31 snarfed kylewm: looking at your PR now. re the test failure, if you run alltests.py with —debug, it should show you more details, including a real stack trace
bnvk joined the channel
indie-visitor joined the channel
# 19:45 tantek If you want to remove Cybe from that page, then please move it to its own /Cybe page and add that to anti-patterns
# 19:45 Loqi tantek: snarfed left you a message 2 hours, 7 minutes ago: in case you didn't see, fb announced that thing i mentioned a while back: http://applinks.org/
# 19:46 tantek KartikPrabhu: your post looks great. honest, from the heart, and positively encouraging.
# 19:47 tantek well, h-as-* does create a microformat object (since it is an h-* classname), it's just not part of any proposed format for doing anything. more of an experiment.
# 19:47 Jeena ah, so I do at least no harm in adding h-as-photo in addition to h-entry
# 19:53 tantek snarfed, quickly skimming applinks.org, looks like some interesting design/engineering behind it. Still a bit confused about the "app to app linking" use case. When would I want that? (as a user)
# 19:53 tantek Actually I should get to bed - don't answer that. :) We can chat more tomorrow. Er, your tomorrow. It's already 2014-05-01 here.
# 19:54 aaronpk is the video going to be posted? the video showed a bunch of good use cases for that
# 19:54 tantek (as Web Workers, let's make sure our own web sites work)
# 19:55 tantek aaronpk - I like your design/methodology better. native app -> micropub -> web and then all web<->web
# 19:56 aaronpk oh this is just so that you can be in the spotify app, click a link, and it launches the songkick app to buy tickets. not really about creating posts or anbything.
# 19:56 aaronpk more like bringing the web's hyperlink experience to native apps
# 19:57 aaronpk the idae being if you publish these applinks tags, you don't need bizdev relationships to be able to link to each others' apps
caseorganic joined the channel
# 19:59 tantek hopefully it will make native app devs think more "web-like" if they end up creating links to various app states/pages
# 20:00 tantek then as we close the gap with webapis perhaps they'll create more web apps
caseorga_ and benwerd joined the channel
# 20:01 Loqi Ok, I'll tell him that when I see him next
wraithgar and dybskiy joined the channel
# 20:08 tantek benwerd - why not link people to a silo to create comments, e.g. Twitter, and just let Bridgy copy them back in?
KevinMarks joined the channel
# 20:09 benwerd tantek - because I don't want to encourage that use if I don't have to. If someone's going to comment in a silo, it might as well be mine ;)
# 20:09 benwerd this way nobody _has_ to create an account on a 3rd party service
hallettj_ joined the channel
# 20:18 KevinMarks And to make sure that sites decide which apps are approved by them
# 20:19 KevinMarks Can I make applinks that say to use Firefox mobile for my site? Not an embedded web view?
# 20:20 aaronpk i still don't quite understand applinks, I'm going to need to try this out with something real before I get it
# 20:21 KevinMarks So, I'm in sf and have a free afternoon. Where should I tempt east coast journalists here to cover f8 to come and sit in the sun with drinks and talk indie web? Near townsend and 7th helps, but lush is key
# 20:22 KevinMarks If twitter were smart they'd be ferrying them to their roof garden from the conf
# 20:23 KevinMarks Applinks are the twitter card for apps, generalised so other apps can do it
# 20:24 aaronpk so I guess app links are a way to say "this web page is also available at this native app URL"
# 20:25 aaronpk and facebook provides an API to answer that question, but you can also just read the applinks tags yourself too
kbs joined the channel
# 20:30 brainTrain I've always been a zeitgeist fan, but that's probably a bit out of the way if you're in SOMA
# 20:31 kbs benwerd: out of curiosity - has anyone [but me now I guess :)] ever expressed a desire to post 'transient' comments? Ie, things that stick around for a few days but bar anyone archiving them, just disappears
# 20:32 aaronpk what is this Prefer-Html-Meta-Tags header that applinks references? the only result for it on google is the applinks docs
# 20:36 kbs KartikPrabhu: haha - yea, guess that's what it could be called :)
# 20:36 aaronpk KevinMarks: really? that wasn't my impression at all... also how would that ever be enforced?
# 20:37 KevinMarks By it happening when you click a link in Facebook, or any app that calls their library
paulcp joined the channel
# 20:40 kylewm benwerd: it looks like you got a webmention from barnaby’s notes stream? that’s … odd
# 20:42 gRegor` !tell benwerd I'm a fan of flat comments and having a "local" comments option since the majority of people are not webmention capable yet. I actually just implemented interleaving of comments and webmentions: http://gregorlove.com/2014/02/1180/
# 20:42 Loqi Ok, I'll tell them that when I see them next
# 20:42 kylewm now I want to send a webmention from my h-feed to ben and see what happens… O:)
# 20:43 Loqi Ok, I'll tell them that when I see them next
# 20:43 kylewm argh gRegor` I want to delete that “Nerd” comment from your blog post, it bums me out every time!
# 20:44 gRegor` Haha. If it helps, it's from a nerd.
# 20:44 gRegor` So not derogatory.
# 20:44 gRegor` My friend Jon just likes to be a punk
# 20:45 kylewm I am overly sensitive to the “don’t you have anything better to do with your time?” people
# 20:46 kylewm no, not much anymore … it’s cool to be nerdy and obsessive now
# 20:46 gRegor` Related: I showed off my new manual webmention form to a friend last night. He said "Look. Quit doing things and making me feel bad for just sitting here playing Minecraft."
kbs joined the channel
# 20:47 gRegor` So I told him I would probably play after I grabbed some food followed by " . . . and BUILD WEBMENTIONS IN MINECRAFT! jk"
# 20:47 bret i used to play a lot of games, now I feel like i had better things to do with my time
# 20:47 gRegor` I should build an IWC logo, though
# 20:48 bret i didnt used to like the iwc logo that much, but the stickers are really rad i have to say
# 20:53 kbs something I don't yet grasp [on android specifically] is what applinks adds, that isn't already being done by the apps?
# 20:54 kbs it works the other way admittedly - the apps register for urls, and the caller simply asks the system to resolve any given url
# 20:55 kbs now it seems that applinks wants to move that [relatively nice] system-provided-user-mediated resolution mechanism into the calling app, which doesn't immediately seem like a better approach
# 20:56 bret the question is pushup counter going to get applinks?
# 20:58 kbs is bugged by apps on android that pre-decide how I want them to handle a shared intent
# 20:58 bret I'm finding the OS X text edit app more and more appealing for writing articles/papers
# 20:59 kbs pretty soon you'll be joining kylewm and me in the church of emacs
krendil joined the channel
snarfed joined the channel
# 21:01 aaronpk shouldn't that be checking to see if each h-entry is actually in-reply-to the url?
# 21:02 aaronpk sidenote: i'm about to test my note posting interface with emoji, and syndicating the emoji to twitter
# 21:05 bret kbs: i get distracted easily, by plugins and colors... text edit has none of this
# 21:07 kbs bret: :) as a fellow-sufferer of that affliction, I see what you mean
# 21:07 gRegor` Man, I hate new Twitter profiles.
# 21:08 kbs I can't read the time well in those glyphs - so he's invited you at 6pm, and you're going to show up at 6.30 - is that how to read them?
# 21:08 gRegor` "Show me this twitter profile." "Ok, here's some of their tweets." *click link to see all tweets* "Oh, here's all their tweets."
# 21:09 kbs demonstrates the easily distracted symptom :)
# 21:12 kbs oh :) I thought that meant "the usual watering hole"
erikmaarten, _6a68 and snarfed joined the channel
paulcp joined the channel
fmarier and kbs joined the channel
# 21:43 kbs snarfed: does applinks (on android specifically) turn the decision on what app to resolve a given URL on its head? Ie, normally the caller registers for URLs, and the user mediates which app is used. With applinks, the caller/library directly decides
# 21:43 snarfed kbs: i don't actually know it in depth, so i'd have to go read the docs with you for detail q's
# 21:44 kbs er. "normally, the receiver registers.."
# 21:44 snarfed i agree w/you that the original ideas of both android intents and app-independent url schemes are great, and ideally make applinks unnecessary
# 21:45 kbs just curious - what do you find the best benefit from using applinks? (I assumed you were working on it to some extent, apologies if that isn't the case!)
# 21:46 snarfed definitely haven't worked on it. at a high level, it seems useful for 1) translating single-app web links (e.g. a quip doc) to an intent/url that opens it in the native app, for all ios apps and android ones that don't correctly/fully register intent filters
# 21:46 snarfed 2) reliably generating a specific app's url/intent across breaking changes by the app itself
# 21:47 snarfed (since the applinks db should be kept up to date)
# 21:47 kbs would that take into account breaking changes by app-version?
# 21:47 kbs naively assumes that intent-filters are best kept with the app, rather than in a global version-free db :)
# 21:48 snarfed in general, i agree with you that it's the ideal solution. the problem is that reality hasn't really followed
# 21:49 KevinMarks Snarfed the problem is that you no longer get to choose what app to use
# 21:50 snarfed KevinMarks: eh, i don't know that this changes that much
# 21:51 snarfed it seems like applinks is useful for when you already know you want a specific app
# 21:51 kbs the navigation flow docs for android certainly puts the choice right in the caller, as far as I can tell
# 21:51 kbs rather than in (the android-centric view) of having the user mediate it.
caseorganic joined the channel
# 21:51 snarfed it doesn't disable generic cases. generic android intents are still powerful and useful independent of applinks
# 21:51 aaronpk but won't applinks stop the android intents from working?
benwerd joined the channel
# 21:52 aaronpk cause applinks will translate the web URL to an app URL, then ask the OS to open the app URL instead of the web URL
# 21:52 Loqi benwerd: gRegor` left you a message 1 hour, 9 minutes ago: I'm a fan of flat comments and having a "local" comments option since the majority of people are not webmention capable yet. I actually just implemented interleaving of comments and webmentions: http://gregorlove.com/2014/02/1180/
# 21:52 KevinMarks For when the site wants only one app. Yes. It works for quip where your site is a stub. It's anti web
# 21:52 snarfed right. most web links are already app-specific, web intents notwithstanding
# 21:53 KevinMarks Yes, though if the app isn't installed it may fall back on the web url
# 21:53 kbs also prefers me (the user) to decide which app -- browser or otherwise -- to handle any given url
# 21:54 KevinMarks So, can I declare to Facebook to open my site in chrome on iOS with this, instead of a web view?
# 21:54 aaronpk let's use twitter as an example. you're in the Quip app, and click a link to a tweet. Quip uses the facebook SDK which translates the link to twitter.com/xxx to a twitter app link like twitter://xxx, then hands it off to the OS which launches the twitter app.
# 21:54 snarfed KevinMarks: i don't know enough about the specifics to answer detailed questions, sadly
# 21:54 aaronpk so instead of apps providing the OS with web links, they now provide the OS an app link, breaking the ability for other apps to handle the twitter.com/xxx link
# 21:55 snarfed aaronpk: right, i see. for content where you reasonably have multiple clients, this isn't the right thing to use
# 21:55 snarfed …or maybe it is, if applinks can return multiple apps for a given url. not sure.
# 21:56 aaronpk but... if twitter one day decides to include the applinks tags, then all apps using the facebook SDK will suddently change
# 21:56 KevinMarks They do let you express multiple apps. But the website declares them. It gets a veto
# 21:56 snarfed sure. all true., i definitely grant those poinst. there are tons of things that don't have multiple clients though. this is a huge improvement on the status quo for those cases.
# 21:57 snarfed if it turns out to tie the multiple-client drawback with the single-client benefit, that's a tradeoff to consider
# 21:58 snarfed i don't know much about the current state on ios. afaik it doesn't let multiple apps register for the same url schemes, right?
# 21:58 snarfed ie you can only do generic intents on ios right now if you specifically look for each candidate app separately…?
# 21:59 aaronpk no, an app can register for a custom URL scheme, but behavior is unspecified if multiple apps register the same
# 22:03 kbs is still quite perplexed on the android end :)
j12t joined the channel
# 22:03 jonnybarnes benwerd: kylewm's posts have also ran amuck in your comment section
# 22:03 kbs eg: if the quip android app registers for a suitable filter - doesn't it automatically work like all similar apps? What benefit do I get by using the applinks navigation mechanism from (say) my messaging app that gets a link to a quip doc?
# 22:04 kbs also (to KevinMarks point) it lets the user decide which app to use, which seems nice to me
npdoty joined the channel
# 22:08 snarfed kbs: yes, it's much more useful on ios than android. again though, as a user, you can't meaningfully choose which app to use in every case
# 22:09 snarfed quip docs are the example here. it's a document silo
# 22:11 snarfed also, minor android point, correctly registering for http(s) urls to your app is subtle and surprisingly few apps get it entirely right
# 22:12 kbs snarfed: true - but fixing any of this is an update away, while applinks appear to shift the burden to all the calling apps :)
paulcp_ joined the channel
# 22:15 kbs I see your point about access-protected URLs on the web, with no browser-client UI to open them. [I'm not quite as opposed to non-browser clients as KevinMarks :)] but I do also feel it takes the choice away from the user. Maybe this is fundamentally an ios-vs-android thing as well
# 22:16 snarfed agreed. it definitely does help interop by hurting user choice a bit, and helps ios more than android
# 22:19 snarfed it does help user choice in a different way, though
# 22:20 snarfed a *very* indirect way: most apps don't register to handle their http links at all, on either platform. app links will theoretically let me make "glue" apps like open-in-app handle way more apps
# 22:20 snarfed which will let users choose to open way more links in native apps, if they want, instead of in their browser
# 22:22 GWG snarfed: I have a Bridgy question for you
# 22:22 GWG Right now, I use Bridgy for backfeed, but not for publishing.
# 22:22 GWG If I disabled the plugin I use and switched, what are the pros and cons, do you think?
# 22:23 kbs thanks snarfed! think I'm slightly less perplexed after hearing how quip finds it handy, and possibly other situations. I was initially only able to appreciate what KevinMarks and aaronpk were saying, good to get another perspective on it :)
# 22:24 snarfed GWG: mostly that you can publish things that aren't just posts
# 22:24 snarfed e.g. comments/replies, likes/favorites, event rsvps
# 22:24 snarfed kbs: np! yeah, unlike them, i'm not really a big web advocate per se
# 22:24 GWG What might I want to do on my side if I did?
# 22:24 snarfed also re quip, fb mostly found it handy, not us :P
# 22:26 GWG I was more thinking...is it worth adding post meta boxes, etc on the Wordpress side rather than hand coding
lukebrooker joined the channel
# 22:27 GWG Basically, metadata in the post editor
# 22:27 GWG Rather than hand adding the classes to links
# 22:28 snarfed you might look at idno, i think it has good uis for specific actions like replying, rsvping, etc
# 22:31 GWG snarfed: Long term vs short term planing
# 22:31 gRegor` snarfed: Is it by design that bridgy comments shows the author URL as the site linked in the profile, but bridgy favorites shows the twitter profile URL?
# 22:32 snarfed gRegor`: i think it's a sad implementation detail. i have to get favorites by scraping html, so i get much less info
# 22:32 gRegor` No API for favorites?
# 22:35 gRegor` So it's scraping the HTML of each of my tweets when it gets a reply? Or periodically?
# 22:37 GWG snarfed: I'm very cautious on getting into Custom Post Types. But it would take it away from the user
# 22:37 snarfed gRegor`: twitter returns favorite count, and bridgy remembers it and re-scrapes when it changes
paulcp joined the channel
# 22:39 gRegor` What if the comments included a second u-url, for the twitter profile?
# 22:40 gRegor` I guess that could mess up some people's implementations if they're only checking for one u-url in the h-card
# 22:40 snarfed even though all mf2 props can technically be multiple
# 22:40 snarfed in any case, i'd accept a PR to do that if you want
KevinMarks_ joined the channel
# 22:42 GWG snarfed: I don't either. But would you if there was a plugin for it?
# 22:42 GWG snarfed: That is what I'm thinking about.
# 22:43 GWG snarfed: It is on my 'roadmap', the list of ideas for the fture.
# 22:48 kbs do you also predict a day when facebook will charge for the "preferred-launched-app" to handle some particularly desirable endpoint? :)
wraithgar and pauloppenheim joined the channel
# 22:50 gRegor` I'm a python novice, so that probably won't happen, snarfed. :) No biggie, though. the u-uid might help, actually. Is that from the twitter API or something bridgy generates?
# 22:50 kbs :) it's a clever idea. Be curious how it evolves - clearly there's a huge incentive for everyone to add these tags so they can show up directly from links from the facebook app.
# 22:51 snarfed gRegor`: bridgy generates it. in retrospect, should be the actual url intsead of a tag uri, but meh
# 22:51 KevinMarks_ remember when half the web went down because the facebook login iframe redirected the container to an error page?
# 22:52 kbs surfs with noscript and ghostery, and probably half the web is already invisible!
# 22:53 gRegor` Can you point me in the repo where the uid is generated?
# 22:54 gRegor` I see they're both "tag:twitter.com,2013:LucyStag" for the comment and like. Wondering if I can reliably extract the twitter name. wasn't sure about that 2013 in there though
# 22:54 gRegor` Or I suppose if they always match, I could group them together that way (on a single post)
# 22:58 kylewm gRegor`: what are you trying to do, out of curiosity?
# 23:03 gRegor` I would rather have them combined like "John Doe liked this and said: blah blah blah"
# 23:03 gRegor` Still separate webmentions. I just want a way to group them together on my end and display them.
# 23:04 gRegor` Or more likely some small text underneath the comment "liked this post"
# 23:04 snarfed btw kylewm, we definitely don't need to block on that __doc__ mf2py PR
# 23:05 gRegor` Ooh, I didn't realize Github issue comments updated in realtime.
# 23:05 gRegor` I'm still pretty new to Github
# 23:09 kylewm thanks for the reply Kartik,totally agree with your line of reasoning
mko joined the channel
# 23:09 KartikPrabhu neat! I was looking to do this myself, but got caught up in other things. If you get to it first, send a PR! :)
mko joined the channel
j12t, j12t_, KevinMarks_ and caseorganic joined the channel
caseorganic joined the channel
# 23:56 acegiak aaronpk: did we decide my markup on acegiak.net was still subpar?