2015-07-20 UTC
wolftune joined the channel
nedorito and wolftune joined the channel
# 00:16 snarfed micropub server semantics question: i'm guessing we should generally follow the unix philosophy and do exactly what the client asks, right?
# 00:16 snarfed e.g. if they edit an existing post and give us a published timestamp, we should change it on the existing post even though it's technically a "lie," right?
# 00:17 snarfed e.g. they might have set it incorrectly originally
# 00:17 tantek snarfed - it might be a correction of a previous error
# 00:17 tantek e.g. a buggy client that set an incorrect published date, then updated to send the correct value
KartikPrabhu and snarfed joined the channel
# 00:35 GWG snarfed: Then I have to define Postly as a buggy client, but I'll revert
# 00:35 GWG tantek: I've reached out. But haven't heard anything. Will try again
# 00:36 snarfed yup, i'm sure ben_thatmustbeme won't mind updating postly to not send published= in edits
# 00:36 GWG I'm giving him a lot of Issue reports
# 00:38 GWG snarfed: Postly has no options to not send all fields.
# 00:38 GWG snarfed: So his edit is more Replace All, than edit in function
scor joined the channel
nedsauce, KartikPrabhu and scoates joined the channel
# 01:03 bear oh my word - it's a good thing loqi doesn't auto-add flooders
# 01:14 tantek eBay is a marketplace for people and companies to sell goods, typically as an auction with bids and an end time.
snarfed joined the channel
nedorito joined the channel
# 01:17 snarfed !tell pfefferle,acegiak hi! looks like your semantic-linkbacks repos have diverged, and each one has useful new features (filters vs salmentions). mind picking one to be the single canonical repo, and then merge them? thanks in advance!
# 01:17 Loqi Ok, I'll tell them that when I see them next
wolftune and KartikPrabhu joined the channel
# 01:18 tantek nice to know even the big boys have problems making their wildcard certs work
snarfed joined the channel
snarfed joined the channel
# 01:29 tantek hmm - looks like a lot of supposed consuming sites for OpenID are dead now or not consuming
wolftune joined the channel
# 01:42 tantek.com edited /OpenID (+477) "tons of shutdowns of openid consuming support, including SourceForge (?openid URL trick just causes a 404), gitorius etc. verfiied they're all gone, moved to Shutdowns" (
view diff )
nedorito and seekr joined the channel
# 01:55 tantek if anyone knows of any other site besides Open Source Bridge and OpenStreetMap that still consumes OpenID, please add it (and be sure to check the shutdown/problems lists first).
# 01:56 tantek We may have crossed the threshold of where there are more IndieAuth consuming sites now than OpenID consuming sites.
# 01:56 tantek bret - have you tried Stack Overflow recently? It just failed for me with "Message signature was incorrect."
# 01:58 bret Python package index. Mobile right now otherwise I'd make the dfn
nedorito joined the channel
# 01:59 tantek !tell cweiske it appears nearly all OpenID consuming sites have shutdown, shutdown their OpenID login UI, OpenID login endpoints (URL), or are have persistently broken OpenID consumption flows (reproduceable errors). Down to 2 verified functional. If you know of more, please add: https://indiewebcamp.com/OpenID#Consuming_Sites
# 01:59 Loqi Ok, I'll tell them that when I see them next
# 02:00 bret pypi is the Python package index which hosts Python packages that can be installed with pip.
# 02:03 tantek per archive.org, myopenid disappeared in June 2014
# 02:07 Loqi slack/snarfed: confirmed re stack overflow, I've logged in recently with openid too
# 02:10 tantek snarfed, I'm wondering if perhaps Stack Overflow and IndieAuth.com as OpenID provider are not compatible (both aaronpk and I had problems with Stack Overflow login, and we both use IndieAuth.com as our OpenID provider)
KartikPrabhu joined the channel
# 02:16 aaronpk oh darn, i'm not logging openid consuming sites from indieauth.com logins, just logging the fact that someone signed in to an openid site
# 02:17 aaronpk interestingly, there have been almost as many openid logins as wiki logins
seekr joined the channel
# 02:23 tantek.com edited /OpenID (+269) "/* Consuming Sites */ move Stack Overflow up to still consuming since verified by snarfed with WordPress OpenID plugin, note issue with IndieAuth vs S.O." (
view diff )
# 02:24 tantek thanks snarfed, aaronpk - results and issue cited.
# 02:34 tantek Vox Media is a media publishing site that acquired former [[silo ]] company [[SixApart ]], and re-used their vox.com [[silo ]] domain for media publishing, breaking all previous vox.com blogs and permalinks.
nedorito joined the channel
# 02:39 GWG ben_thatmustbeme: Then where is it getting it?
# 02:40 ben_thatmustbeme i send them in mobilepub because the posts can be saved and thus published in the past
# 02:40 ben_thatmustbeme i checked the data sending from the client and receiving on my end, and its not there
# 02:41 GWG But the blank fields are all there.
# 02:42 GWG Yes. I meant my statement that your implementation of edit is a complete replacement of all fields.
# 02:47 tantek ben_thatmustbeme: request already filed for Loqi to ignore "what are" "what is" questions ending in a preposition.
# 02:48 GWG Yes, but it means the time reset it on the WordPress side.
wolftune joined the channel
# 02:49 GWG ben_thatmustbeme: That is what I'd expect from Edit.
# 02:49 ben_thatmustbeme although i want to see what happens with edit functionality now that the spec changed so much
# 02:50 Loqi tantek meant to say: chuckles at new vs. old
# 02:50 ben_thatmustbeme has to review new standard, changed without consulting the only person who actually had an edit mp client
# 02:51 tantek that's not process, that's just plain rational courtesy ;)
# 02:51 tantek it's good form to check with your implementers before you change things, and try to get some consensus - that's what we're doing now with all microformats2 parser changes btw
# 02:52 tantek ben_thatmustbeme: had you documented your "edit" support somewhere on the micropub page and aaronpk just missed it?
# 02:52 tantek yeah - an iteration on brainstorming would have been a good intermediate step, now that we have so many live implementations of micropub
# 02:53 tantek ah - sounds like there may have been a lack of communication
# 02:55 GWG I just don't know how to keep backward compatible in this case.
# 02:56 GWG ben_thatmustbeme: I'm just a contributor.
# 02:56 GWG The edit functionality, from what I can tell, has some bugs.
# 02:57 ben_thatmustbeme ahh, good to get this all worked out while there are few implementations of edit then
KartikPrabhu joined the channel
# 02:58 GWG I want to ensure the standard won't change if I fix the implementation.
# 02:59 tantek you cannot ever ensure the standard won't change
# 02:59 tantek all you can do is try to implement the current standard, and report back if it works or not for you
# 03:00 tantek for example, if you find problems while implementing, then the standard will likely change in response
# 03:00 tantek ben_thatmustbeme: possible to capture these as "Issues" on the wiki?
# 03:00 tantek sure that's likely to prompt a preference order of processing which should be there anyway
# 03:01 tantek I'd suggest treating such conflicts in the apparently least lossy way
# 03:02 tantek assuming it's possible to make a form for edits or deletes
# 03:02 tantek perhaps you can upload a screenshot to illustrate?
# 03:03 GWG I think that makes that properties section easier to interpret
# 03:05 ben_thatmustbeme makes for more complexity for the endpoint implementer (having to figure out what is least destructive)
# 03:05 tantek ben_thatmustbeme: my point is that the spec should use that as a design principle and then still be explicit
# 03:05 tantek rather than leave any such cognitive burden up to the endpoint implementer
# 03:06 tantek I was suggesting an explicit spec resolution for your question / issue
# 03:06 ben_thatmustbeme i'm just brainstorming at this moment, getting my thoughts down so i can reference the IRC logs later
# 03:07 ben_thatmustbeme i think you may be right that it really doesn't make sense to have that as a single form for delete/edit
# 03:09 GWG Does it make sense for me to clean up the page a bit?
scoates joined the channel
# 03:09 GWG Or do I not know enough about this to try to organize it a bit?
# 03:15 ben_thatmustbeme realized he can put his template in his actual signature in mediawiki settings, making '~~~~' much more useful
snarfed joined the channel
# 03:24 david.shanske.com created /Micropub_Examples (+8864) "Created page with "Examples of IndieWeb sites that support the micropub API (in order of deployment). === Aaron Parecki === {{aaronpk}}
supports publishing various post types on aaronparecki.com u..."" (
view diff )
# 03:25 tantek huh - why move the examples to a separate page?
# 03:25 tantek it's much more compelling to have them listed right there up front on the Micropub page itself
# 03:25 GWG The page was getting too long, I thought.
# 03:26 tantek the existence of implementations is definitely not something to burry
# 03:26 Loqi tantek meant to say: the existence of implementations is definitely not something to bury
# 03:26 tantek if anything, it's one of the biggest strengths of micropub that should be very obvious up front & center
# 03:26 tantek at a minimum, date-ordered list of implementations, with date (linked to cite) of when deployed and in use
# 03:27 tantek that kind of rapid progress shows momentum for anyone else considering implementing the spec
# 03:27 tantek which is a huge part of what a spec needs to prove frankly, whether it is worth implementing
# 03:27 tantek GWG, the existing examples are in date order - so that's a good start
# 03:27 tantek basically, the existing examples serve that purpose, if you want to change them, please make sure you don't break that aspect of the page
# 03:28 GWG Let me see if I can address both concerns.
# 03:38 tantek GWG, not quite - the loss of headers means loss of numbering
KartikPrabhu joined the channel
# 03:38 GWG Okay. Will revert. It just seems like the page is unwieldy.
# 03:38 tantek I'd say keep the headers (since they were permalinks before), and just give each one a {{main|...}}
link that links to the more extensive version
# 03:38 KartikPrabhu updated homepage to list a Atom feed for Notes generated using hfeed2atom !
# 03:40 KartikPrabhu weird though that Firefox does not show the <content> of an Atom entry only <summary>. Harder to debug.
# 03:41 KartikPrabhu yes. view source does have it. But a more visual representation would be very useful
# 03:41 snarfed hey KartikPrabhu btw, i'm guessing you still haven't seend bridgy backfeed for your notes h-feed yet right?
# 03:42 KartikPrabhu snarfed: I assumed it would update when I posted a new note and it looks for the feed
# 03:43 snarfed kylewm any idea why the all-h-feeds change might not be working?
# 03:44 KartikPrabhu snarfed: no. it hasn't backfed the old posts yet. Bridgy page still show "No post links found "
# 03:45 KartikPrabhu yay! Feedly show my note feed alright! except it wants to put my logo as "featured image" dumb thing
# 03:47 KartikPrabhu snarfed: about to post note about updated use of hfeed2atom. Let's see if bridgy picks it up
KartikPrabhu joined the channel
lukebrooker joined the channel
# 04:05 snarfed KartikPrabhu: i think i know why. i'll try fixing
# 04:15 tantek GWG, that still breaks the seeing explicit support on what URL since when
# 04:15 tantek Each person/site still needs that explicit summary statement to make it clear *what* has been implemented since *when*
# 04:16 GWG I think I should just revert the section and give up.
# 04:16 tantek GWG - does the preference to keep that information inline not seem valuable to you?
# 04:18 GWG Yes, I just don't know how to balance that.
cweiske joined the channel
# 04:43 Loqi acegiak: snarfed left you a message 3 hours, 26 minutes ago: hi! looks like your semantic-linkbacks repos have diverged, and each one has useful new features (filters vs salmentions). mind picking one to be the single canonical repo, and then merge them? thanks in advance! http://indiewebcamp.com/irc/2015-07-19/line/1437355029097
# 04:45 tantek KevinMarks: odd - just tried my twitter home page and it loaded
KartikPrabhu and wolftune joined the channel
# 04:53 tantek and 7 minutes of distraction later, twitter home page tab closed
indie-visitor joined the channel
# 05:00 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
# 05:02 theRealChupacabr so just curious..and a fairly elementary question I think...but what is the difference between Javascript and AngularJS..and why would I use one over the other? Seems to me like AngularJS is just built on top of JS?
bigbluehat, Garbee, voxpelli, mattl, lukebrooker and tantek joined the channel
# 05:24 tantek seems like theRealChupcabr answered their own question?
# 05:24 kylewm snarfed: KartikPrabhu: i think it's an SSL error so we're not able to do original post discovery on K's site at all
# 05:24 kylewm SSLError: hostname 'kartikprabhu.com' doesn't match either of '*.webfaction.com', 'webfaction.com'
# 05:25 snarfed kylewm: agreed. i just deployed with my dep versions, that may fix it
# 05:26 kylewm snarfed++ for wading through python/ssl version interdependencies
# 05:26 snarfed and i still haven't fixed the requirements.txts, just deployed from my existing virtualenv
# 05:29 kylewm any guesses about why requests-oauthlib + GAE + Flickr might not be working? same code works from my command line but fails in the dev_appserver interactive console
# 05:30 kylewm i'd guess it has something to do with what kevin was saying earlier, the fetch infrastructure
# 05:30 snarfed huh. there was a bug like that a while back that they fixed. let me find it
eschnou joined the channel
# 05:35 cweiske !tell tanktek I don't have a list of sites I logged in via OpenID *because I did not need to write them down* in my password manager
# 05:35 Loqi Ok, I'll tell them that when I see them next
KartikPrabhu joined the channel
rj_ joined the channel
# 05:40 snarfed they're not. it's complicated interop between a number of different libs and app engine
DanC_ joined the channel
kronda joined the channel
# 05:41 theRealChupacabr not sure if anyone answered as I was disconnected. But I asked earlier the diff between AngularJS and JS. If someone already answered, my apologies but could you copy/paste answer again please?
jancborchardt and hugoroyd joined the channel
# 05:41 snarfed KartikPrabhu: oh, you're publishing! sorry, we thought this was just about backfeed. sure, retry now
# 05:42 KartikPrabhu theRealChupacabr: yes AngularJS is as abstraction framework over usual JS just like JQuery
jden joined the channel
# 05:42 theRealChupacabr ok..so the ideal way to learn this is to learn JS first then move to Angular?
# 05:43 theRealChupacabr so whats the diff between JQuery and Angular? Anything worth noting? or just a preference thing
# 05:43 theRealChupacabr just starting to get into web stuff
# 05:44 tantek theRealChupacabr: consider learning HTML first, then CSS, then JS
bret joined the channel
# 05:44 theRealChupacabr I've got a decent grip on HTML and CSS
# 05:45 theRealChupacabr I'm not entirely new to coding....currently a CS student at PSU...but trying to branch out and explore different sides to things
# 05:45 KartikPrabhu theRealChupacabr: sure. but if you just want to learn things without a goal in mind there is too much
# 05:45 theRealChupacabr trying to start projects...building a website is on the "to-do" list
rknLA joined the channel
# 05:46 KartikPrabhu then start by building a website and learn things that you need to make things on the website
# 05:46 theRealChupacabr haha..fair enough.
# 05:46 tantek theRealChupacabr: building a website is a great way to start
# 05:46 tantek and if you're up for building a personal site for yourself, then you've come to the right place!
ramsey joined the channel
# 05:47 kylewm snarfed: I already have the newest python sdk :(
# 05:47 theRealChupacabr yeah...I've been here before A.Case was a guest speaker for a class I had last year...She talked about this
# 05:48 snarfed kylewm: oh agreed, i don't think it's that exact bug, just maybe something similar
# 05:48 theRealChupacabr Its been on my list of things to-do...kinda been busy with school projects
# 05:48 snarfed but if oauth-dropins works in dev_appserver, then…maybe not
bigbluehat joined the channel
# 05:48 snarfed kylewm: yup oauth-dropins flickr works in current dev_appserver
voxpelli and uranther joined the channel
# 05:49 theRealChupacabr Any recommendations on server space? Other than having my own personal server?
# 05:49 KartikPrabhu theRealChupacabr: I learnt everything I know about web stuff by building my own site. Learnt python from scratch on a need to know basis and had a website from scratch in 3 months
mattl joined the channel
# 05:50 theRealChupacabr I know what hosting is :) I meant...recommendations as far as companies to go with
# 05:50 cweiske tantek, I could simply ask my openid provider for a list of sites
# 05:50 tantek theRealChupacabr: recommendations are on that page :)
# 05:50 KartikPrabhu theRealChupacabr: see the Loqi reply after my question. with the hyperlink
# 05:51 theRealChupacabr ah..gotcha..my bad..didn't see that.
# 05:52 tantek cweiske most such lists are out of date - that's the point
# 05:52 cweiske the ones (apart from the stackoverflow family and personal sites) I logged on in the last year
CaptainCalliope joined the channel
# 05:54 tantek really? where was the login? vox.com has been dead (of its blogging and openid support) since 2010-09-30
benward joined the channel
# 05:59 tantek hmm - I get "Sorry, there was a problem authenticating you with OpenId. Error: invalid_credentials"
rjr4 joined the channel
# 06:01 cweiske probably. openid is likely the most untested part of indieauth.com
# 06:13 tantek cweiske - I'll keep documenting the issues, we can investigate as needed
theRealChupacabr joined the channel
# 06:16 cweiske tantek, so you're counting gitorious.org as openid shutdown when the site itself is being shutdown?
# 06:16 cweiske openid login on our gitorious instance @work is still working
# 06:17 tantek cweiske - that's gitorius as consuming software
KartikPrabhu joined the channel
# 06:22 tantek.com edited /OpenID (+47) "/* 2015 */ scope vox shutdown to SixApart's Vox.com in particular, since modern vox.com has its own openid consuming support" (
view diff )
evalica joined the channel
nxd4n_, nxd4n, ttepasse, sanduhrs, Jihaisse, evalica, KartikPrabhu and loic_m joined the channel
# 07:07 cweiske tantek, do you have a feed with only the longer articles? without the tweets and likes and so?
# 07:12 tantek cweiske: in two forms, not quite equivalent, it's definitely on my Working On list
# 07:14 cweiske you have an updates.atom feed, but not for articles
theRealChupacabr joined the channel
# 07:17 tantek cweiske - I did a one-off just for articles, but it's not discoverable because the URL is a bit of a hack
eschnou joined the channel
# 07:18 tantek (though I think it also works for s/article/note and s/article/like as a side-effect of how it's coded)
# 07:18 tantek cweiske: the other thing is that I often write long notes that are very article-like
# 07:19 tantek so the explicit post type distinction is not necessarily the most useful
# 07:19 tantek like I said, not the best solution, temporary, but you're welcome to use it if you're ok with those conditions :)
lewisnyman, KartikPrabhu, gRegorLove and lewisnyman_ joined the channel
ttepasse, voxpelli, evalica, Phae, ramsey, KevinMarks, JonathanNeal, benward, Garbee, rknLA, loic_m, bigbluehat, hugoroyd, jancborchardt, jden, tantek, mattl, csarven, petermolnar, bret, uranther and kronda joined the channel
petermolnar, KevinMarks, tantek, ttepasse_, interactivist, Sebastien-L, evalica, j12t and LauraJ joined the channel
interactivist, stream7 and j12t joined the channel
friedcell and j12t joined the channel
evalica joined the channel
LanceyWork, scor, j12t and frzn joined the channel
# 11:56 voxpelli my micropub endpoint now supports multipart file uploads, so soon I will be able to use OwnYourGram – yay!
# 11:57 Loqi Ok, I'll tell them that when I see them next
# 12:10 GWG I seem to be having trouble with creating unit tests for micropub, something snarfed had suggested. I may need to step back.
squeakytoy joined the channel
# 12:13 GWG voxpelli: I'm not sure how to approach this. Any advice is always welcome
# 12:14 GWG I was thinking of working on something else for a bit and returning.
# 12:14 voxpelli GWG: have you got any tests set up right now? or is the challenge you're having to get the first tests off the ground?
# 12:17 voxpelli GWG: I could try and get the basics of unit tests set up for you if you want?
j12t joined the channel
# 12:18 GWG voxpelli: I wouldn't want to put you out on this.
# 12:19 voxpelli GWG: I will need to do it someday anyhow, will probably be creating some more WordPress things soon considering that I will be freelancing in the fall rather than working for a startup like the last year :)
# 12:20 GWG voxpelli: If you want to take a stab at it, you are always welcome. I've never done any unit testing and am still trying to figure out how to design one.
# 12:21 voxpelli GWG: my node.js micropub endpoint already has some unit tests so can probably convert some of those
# 12:24 GWG In the meantime, I still have to fix some bugs in my last PR.
Pierre-O and scor joined the channel
# 12:35 Loqi slack/snarfed: GWG: did the pippin tutorial not work?
# 12:35 GWG snarfed: Not exactly. It's based on a blank WordPress database.
# 12:36 Loqi slack/snarfed: sure, that's fine. I'm ok with that. iirc it clears it every time
adactio joined the channel
# 12:45 GWG It's just a matter of wrapping my head around all this. Having trouble with that.
csarven, j12t and theRealChupacabr joined the channel
# 13:13 GWG voxpelli, snarfed: My usual approach when something is stumping me is to take a break and work on something else for a bit, and see if I come back with a fresh pair of eyes, if it seems to help. IWC Edinburgh is this weekend. Might put it off till after that.
csarven joined the channel
mlncn, interactivist, wolftune, j12t and schmarty joined the channel
tantek, scoates, evalica and snarfed joined the channel
j12t, tilgovi and fourtonfish joined the channel
chalettu, KevinMarks, KevinMarks__ and Deledrius_ joined the channel
hober2, kronda_, KevinMarks, Kongaloosh_, ben_thatmust, edsu_, shaungui1e, GWG-, jacus, MylesBraithwaite, Unifex, arlen, minsky, hidgw, renoirb, DanC_, ehlovader, modem, KevinMarks__, j12t and evalica joined the channel
nedorito joined the channel
KevinMarks, shiflett, schmarty, friedcell, Sebastien-L and j12t joined the channel
# 15:49 voxpelli aaronpk: thanks! now I need to fix my endpoint so that it understands the requests sent, so trying to set up OwnYourGram locally now
# 15:49 aaronpk cool! you could also use a localtunnel thingy to get ownyourgram to post to your local site
# 15:52 voxpelli aaronpk: the ownyourgram repo seems to be missing an sql definition btw
# 15:52 voxpelli tried the one from Quill, but it's missing some fields
# 15:54 aaronpk sometimes I am lazy and don't add the SQL table because I don't actually expect anyone to use these :)
# 15:57 voxpelli aaronpk: also: a weird thing – out of laziness I tried my local OwnYourGram towards my live endpoint, the one I also tried with the live OwnYourGram. The local one worked, but the live one didnät
# 15:57 aaronpk Jeena just added support for sqlite to Quill, so I may try porting that to Ownyourgram too
# 15:57 aaronpk i assume you're using the "test" button from the site?
# 15:58 voxpelli aaronpk: I'm logging the requests to my endpoint and they are completely empty after I've tried parsing stuff from them – so something in the multipart parsing seems to be failing
# 16:00 voxpelli the file seems to be picked up, but none of the other data, but when I did the request from my local machine everything worked well
# 16:01 Jeena aaronpk the ORM zou're using, can it also create tables and perhaps even do migrations? If so this would be neat to write the schema with it and have it available and updatable everywhere
# 16:02 aaronpk but I could write PHP code and migrations manually
# 16:04 aaronpk if the database were more complicated i would do it, but it's just one table and worst case if the table goes away people just log in again, so it's not like it's canonical data
# 16:06 Jeena is there a keyboard shortcut to change between keyboard layouts in os x?
# 16:07 voxpelli aaronpk: any idea on what could be wrong with the multipart request on the live ownyourgram from the perspective of my endpoint? :P
# 16:08 aaronpk voxpelli: you're sure you're running the latest version locally?
# 16:08 voxpelli I installed it locally using "composer install", so the lock file should have made it be the same version of everything
# 16:08 aaronpk because I just switched it to encode the multipart requests itself
# 16:10 voxpelli aaronpk: are you sure your live server updated the dependency properly? :P
j12t joined the channel
# 16:15 aaronpk i broke it for adactio but fixed it the other day
# 16:17 aaronpk wow lots of people using ownyourgram with a Known site
# 16:18 aaronpk voxpelli: what server-side code are you using to handle the request? just built-in PHP parsing?
# 16:19 voxpelli aaronpk: so in other words: I use my own and am trying to get it to work :)
# 16:21 rhiaro attending.io is a lightweight silo for events registration
# 16:21 aaronpk voxpelli: ooh okay, no idea what those do. it's sending content-transfer-encoding:binary
# 16:21 KevinMarks_ granary is A library and REST API that frees you from social network API chaff and and exposes the sweet social data foodstuff inside as HTML and JSON with microformats2, ActivityStreams, Atom, XML, and more. https://granary-demo.appspot.com
# 16:22 tantek Loqi, you should add the period at the end of a definition if someone doesn't provide it, since you seem to need it yourself.
# 16:23 voxpelli and well – it works from one of the ownyourgram instances :)
nedsauce joined the channel
# 16:24 aaronpk i can run node stuff on my laptop so I can try to take a look too
arlen and tilgovi joined the channel
tantek joined the channel
# 16:42 voxpelli one thing that makes tunneling a bit hard is the coupling between ones identity and ones endpoint – I have to make my micropub endpoint part of an IndieAuth identity or pick another IndieAuth-compatible provider :P
# 16:43 aaronpk do you use any of the non-social logins for indieauth.com?
# 16:43 voxpelli or rather: the coupling between the discovery of ones micropub endpoint and authentication endpoints – that both has to be discovered from the same URL
jciv joined the channel
# 16:44 aaronpk easiest way I've found to test this stuff is just make a subfolder and put a GPG key or email address there, then you can point to your dev micropub endpoint from the subfolder
# 16:44 voxpelli aaronpk: nope, but right now I would just like to say to OwnYourGram: Hey, you know me – I'm "kodfabrik.se", but today I want to be crazy and post to a blog at "a-tunneled-example.com" rather than my ordinary site
wolftune joined the channel
# 16:44 aaronpk but you don't want to update your micropub endpoint on your site?
# 16:46 tantek aaronpk: IIRC tommorris uses Persona for IndieAuth
# 16:47 voxpelli I think it makes sense to ask two questions – "Who are you?" and "Where do you want to post?" – sometimes one can be implied from the other, but at other times not
# 16:47 tantek voxpelli: perhaps that routing to another blog should be done by your site?
# 16:47 aaronpk the problem with that is it adds an extra UI step every time you sign in
# 16:47 tantek e.g. it sounds like a case of a private personal site post that you publicly POSSE to another site
# 16:48 voxpelli tantek: I currently have two standard sites + one development sites I want to have set up and I can only point my site at one of them
# 16:48 tantek aaronpk - the silo analogy here is the Twitter client's support for posting to multiple Twitter accounts
# 16:48 aaronpk yeah I would recommend setting up a separate indieauth ID, either a subfolder or subdomain
# 16:48 voxpelli aaronpk: it only needs to add an extra UI step if it only found one of the two pieces it needed
# 16:49 aaronpk voxpelli: but every client would have to implement that, and it would be a huge burden on clients
# 16:49 aaronpk I wonder if the auth server could let the user choose it
# 16:49 tantek aaronpk: makes sense to make it a client burden
# 16:49 tantek supporting multiple accounts (like the Twitter app does) is more of a UI problem, and thus the focus should be on the client being smart about it
# 16:49 voxpelli not that much of a burden, just an extra request to ask for a place to discover the micropub endpoint if one isn't found at ones identity
# 16:50 aaronpk like if the auth server let the user choose which micropub endpoint
snarfed joined the channel
# 16:50 tantek voxpelli: that could be determined via the rel=me identity network
# 16:50 aaronpk it would mean the response of verifying the auth code would also contain the location of the micorpub endpoint
# 16:50 voxpelli tantek: not necessarily, micropub could be used for multi-author blogs that don't necessarily tie in to a rel-me network I think?
# 16:50 tantek so your site would just rel=me link up with all the other places you like to post
# 16:51 tantek and then they could have their own endpoints at their own URLs
# 16:51 tantek micropub for a multi-author blog has yet to be figured out AFAIK :)
# 16:51 voxpelli I guess rel-me could be a way to suggest sites that a user could may want to edit, but I think the user should be able to provide one itself as well
# 16:52 kylewm tantek: I have a fork of the known micropub plugin that does multiauthor
# 16:53 aaronpk not sure it helps with multi-author blog, but does help with one ID posting to multiple endpoints, similar to signing in to tweetdeck with one account and having the ability to post to multiple
# 16:54 tantek aaronpk - is there a name for that feature? multiaccount? or something?
# 16:56 ben_thatmustbeme specifically i was trying to figure how it would work if you wanted to post "ONLY" to the chained destination
# 16:56 aaronpk "add an account" -> oauth screen -> click cancel -> "Error! Please return to your TweetDeck client and try again."
# 17:00 voxpelli if one could just verify that a token gives access to a mircopub endpoint then it would be fairly easy to add new sites, one wouldn't have to do the oauth dance more than the first time
# 17:02 aaronpk imagine when you request an access token, the token endpoint also returned the micropub endpoint to use
# 17:05 voxpelli aaronpk: why would the token endpoint have to care about what endpoint one would use?
snarfed joined the channel
# 17:08 voxpelli aaronpk: how would it find out what were available? currently the indieauth endpoint has no knowledge about any of that I think?
# 17:09 voxpelli also – it would be nice to be able to identify and post to a place without advertising that fact publicly
# 17:09 aaronpk also has built-in fallback for the existing method which would just use the first one
# 17:09 aaronpk if you build your own auth server you could do it all internally without exposing anything
# 17:10 aaronpk imagine that auth screen is not indieauth.com, but a page on your own site. you'd control what goes into the list
# 17:11 tantek that's a good way to encourage experimentation in this space
# 17:13 voxpelli I'm thinking that it would be enough to auth once with the client and then use that one auth to communicate with lots of micropub endpoints?
eschnou, gRegorLove and mlncn joined the channel
friedcell, KevinMarks and benwerd joined the channel
# 17:37 aaronpk i'm thinking the only change needed to support this is having the token response include a micropub endpoint
keroberos joined the channel
# 17:44 aaronpk a token verification endpoint is fine too, probably worth adding that to the spec
# 17:45 aaronpk i'm actually in the middle of a discussion about this with the oauth mailing list right now
# 17:46 voxpelli there can probably be many ways one can populate the site column in – more auomated ones, more manual ones or a combination of the two
# 17:48 aaronpk that doesn't even require a change to the micropub/indieauth spec though
# 17:48 aaronpk that's just a client doing its own cookie login and associating multiple IDs with it
# 17:49 voxpelli aaronpk: it does require clients to support multiple sites for one identity, but nothing in the spec stops that from happening currently, it's just that a /verify resource could make for a better experience
# 17:50 aaronpk it just requires clients have their own concept of identity, and tying multiple indieauth identities to it
# 17:50 aaronpk you can go do this without any change to indieauth right now
# 17:51 voxpelli sure, but the specs doesn't require you to couple the indieauth identity with the micropub endpoints currently either I think? so why do that just so you can create another grouped identity rather than reusing an existing identity?
# 17:52 aaronpk i guess it's sort of implied that they're tied together right now
# 17:53 voxpelli in my Editorial.app workflow, the only client I've built so far, I separated the authentication from the choosing of publishing target – worked well there (even if the flows are yet a bit rough due to a tough dev environment)
tantek joined the channel
# 17:54 voxpelli I guess I will try to find time to experiment with some other clients to try out the flows better
keroberos joined the channel
# 18:03 voxpelli aaronpk: perhaps you want to add some thoughts there regarding the oauth mailing list discussions
j12t joined the channel
# 18:04 aaronpk voxpelli: out of curiosity, why did you make a separate endpoint for that, instead of just responding to an empty GET request at the micropub endpoint?
# 18:05 bret empty request is what the wiki says to do right now?
# 18:05 voxpelli aaronpk: since I felt it was about two different operations + took inspiration from Twitter
snarfed joined the channel
# 18:06 bret aaronpk: i though i read that yesterday
snarfed1 joined the channel
# 18:08 aaronpk the advantage of using hte same endpoint is you don't need to figure out where the other one is
# 18:09 bret the advantage to a separate endpoint is less endpoint logic (in the case of twitter, theres only one twitter), but I think in this case maybe the same endpoint argument has a big advantage
keroberos joined the channel
# 18:13 voxpelli Mine is a subresource, but both works for me – a bit less explicit to have it on the same – error codes can then mean either lack of implementation or lack of access
# 18:13 aaronpk hm yeah, would need to define an explicit error response if the token is invalid
# 18:14 voxpelli I guess many just check for 2xx and considers everything else a failure :P
snarfed joined the channel
# 18:15 aaronpk everything else is a failure, just might be a failure for different reasons
friedcell joined the channel
snarfed joined the channel
# 18:26 aaronpk I really like this idea of selecting the micropub endpoint during login, but don't have a reason to use that myself yet
snarfed1 joined the channel
# 18:27 GWG aaronpk: Did you look at the /Micropub page? I tried to refine the page without changing the meaning
# 18:28 aaronpk actually I wonder if that would help clients even for the single endpoint use case
# 18:28 aaronpk i've heard some people complain about parsing the html twice looking for the token endpoint and then the micropub endpoint
keroberos joined the channel
# 18:29 tantek needs to update/generalize his link/rel header/html parsing PHP lib
# 18:29 aaronpk there are two complaints. one is parsing the HTML, the other is doing it twice
# 18:29 KevinMarks_ also, if you find multiple endpoints you can give multiple syndicate options?
# 18:29 aaronpk KevinMarks_: the idea is to remove the burden of that from cilents
# 18:29 tantek anyone complaining about parsing is likely not a selfdogfooding publisher themselves :P
# 18:29 aaronpk tantek: i just use the microformats parser since it does rel value parsing
# 18:30 tantek true - I was offering up the simple(r) minimal PHP lib if they just wanted to quickly get those rels
# 18:31 aaronpk is thinking about asking this as a question to the oauth mailing list
# 18:31 KevinMarks_ “Now, I happen to work a media company, and I happen to run a website that can be bloated and slow.”
# 18:31 aaronpk since they sidestep the whole endpoint discovery in the spec
# 18:32 aaronpk "In addition, this specification leaves a few required components
# 18:32 tantek because that's where silos insert their proprietariness and ToS for getting an app key etc.
# 18:32 aaronpk partially or fully undefined (e.g., client registration,
# 18:32 aaronpk authorization server capabilities, endpoint discovery)"
# 18:32 aaronpk "This framework was designed with the clear expectation that future
# 18:32 aaronpk work will define prescriptive profiles and extensions necessary to
# 18:32 tantek exactly:leaves undefined - for the silos to insert propreitary/bizdevy crap
# 18:33 Loqi tantek meant to say: exactly:leaves undefined - for the silos to insert propreitary/bizdevy carp
j12t joined the channel
# 18:33 tantek going to start saying "carp" to indicate where fishy things may be going on
# 18:33 tantek and more language-sensitive friendly than "crap"
# 18:34 tantek and not sure a smiling emoji would make it any better
keroberos, lewisnyman and tantek joined the channel
fkooman, KartikPrabhu and snarfed joined the channel
# 19:14 bret has anyone uploaded native video to twitter via the API?
# 19:14 bret and show how to do a multipart *by hand*
eschnou and tilgovi joined the channel
schmarty joined the channel
# 19:41 kylewm twitter's way of doing multipart with oauth1 is awkward iirc, thankfully the python oauthlib supported it directly
# 19:44 aaronpk it occurred to me that choosing the micropub endpoint at the auth server only solves half the problem, because in that case all of the micropub endpoint options would need to share a token endpoint still
cweiske, Fran0, tantek, sparverius, danlyke and friedcell joined the channel
chreekat, tvn and KartikPrabhu joined the channel
tantek, snarfed1 and snarfed joined the channel
# 20:24 tantek KartikPrabhu: YES! I've been doing that since 2008 :)
# 20:24 tantek I think before that I was using my own HyperCard stack
# 20:26 aaronpk google calendar handles it reasonably well, but it's the only thing i've found that does
KevinMarks and KevinMarks_ joined the channel
# 20:26 aaronpk you have to be proactive about adding things in the correct timezones and such tho
# 20:26 tantek except Google Calendar pretends to be Google Spider Searchbot when it is requesting .ics files :P
# 20:27 tantek just goes to show Google is likely just using ONE GIANT CACHE ;)
keroberos and eschnou joined the channel
# 20:43 kylewm I don't know what Hypercard is except that someone says "whoa, Hypercard" every time it is mentioned
keroberos joined the channel
# 20:55 KartikPrabhu One cache to index them all, One bot to find them // One spider to search them all, and in the silo bind them
# 20:58 tantek One bot to retrieve them all, one spider to find them, one cache to index them all, and in the silo bind them.
# 20:59 tantek gets pedantic about what the bot vs the spider do ;)
snarfed joined the channel
# 21:01 tantek also it just sounds creepier for the spider to find you
keroberos joined the channel
# 21:04 aaronpk maybe the mobile web sucks when you build terrible websites
# 21:04 aaronpk maybe also native apps suck if you build terrible apps
# 21:04 KartikPrabhu aaronpk: that article has gotten that back-reaction from web designers on Twitter
# 21:04 tantek content is made harder when you make it dependent on software, or only viewable with custom downloaded software. per js;dr
lewisnyman joined the channel
# 21:08 aaronpk wow. 25s for me and i'm on a computer on pretty good wifi.
# 21:12 KartikPrabhu my homepage is 0.48s 10 times faster. my article with custom post style and a video embed is still 1.12s... Verge is not doing this right at all
keroberos and daf joined the channel
KartikPrabhu, keroberos and Pierre-O joined the channel
friedcell joined the channel
# 21:39 tantek I feel like that came up in discussion before
# 21:40 snarfed sure! the interesting part seemed like the project/community, not the license
# 21:41 tantek snarfed, ironically, that URL doesn't mention "license" at all.
# 21:41 snarfed i expect they also think the project/community is more important than the license :P
# 21:42 tantek snarfed, I'm not sure I care much for a community only around a code-of-conduct (in theory) rather than being developed in practice for a specific community. Feels a bit abstracty/astronomy to me.
# 21:42 tantek but the absence of a license (when so many - nearly all - others have a license, especially the ones they cite, is odd)
# 21:43 Loqi tantek meant to say: but the absence of a license (when so many - nearly all - others have a license, especially the ones they cite), is odd
# 21:43 tantek aaronpk, lol. it's open because I say it's "Open"!
# 21:43 tantek one would think that an *open source* code of conduct would remember to put the license on there up front. just sayin.
# 21:44 tantek no "copyright" nor rel=license either (yes I checked the source)
# 21:44 tantek it's likely just incomplete, a work in progress, as the end of the document has an obviously * bulleted list that's collapsed into a single paragraph (like someone forgot to do the markdown -> HTML conversion on that part)
friedcell, snarfed and keroberos joined the channel
# 22:06 kylewm it's great that github is adopting a CoC for their open source projects
# 22:10 tantek is that related to the one snarfed mentioned?
lukebrooker and Pierre-O joined the channel
# 22:26 kylewm yeah, they are working on it and have adopted it for their projects
# 22:27 kylewm "We are proud to be working with the TODO Group on the Open Code of Conduct'
# 22:27 kylewm oh, yes it's on the github blog, looks visually similar to any github file though
# 22:28 tantek might have to make a minor correction about "the"
# 22:28 tantek since any of the openly licensed codes of conduct are *an* "open code of conduct"
gRegorLove joined the channel
# 22:36 tantek I just got an email that broke a URL at 59 characters
# 22:37 tantek so those of you who depend on slugs for uniquely identifying a post, and 404 if the slug is wrong, 59 characters is the most you may get (including the http... part)
wagle joined the channel
# 22:51 aaronpk Google+ Photos will be shut down August 1. It will be replaced by Google Photos.
# 22:51 aaronpk also hilarious autogenerated slug which removed the "+" making the slug very misleading
yakker and snarfed joined the channel
# 23:20 tantek as in, what a simple checkin micropub client could work / look like
# 23:21 tantek the checkin UI is as easy as dropping a pin on a map.
# 23:21 tantek that checks you in, and then you can add any details you want asynchronously, choice of venue (auto-suggested), snap pin to venue, who you're with, etc.
# 23:22 tantek point being, you don't have to wait for a list of venues to load, but when they have asynchronously loaded (i.e. after your checkin), you can then tap one to pick it for your checkin and snap the pin to the venue
# 23:22 tantek the more important point is the simplest / fastest possible checkin UI so you can checkin and move on
# 23:22 tantek and when you have another moment you can *optionally* confirm suggested venue etc.
# 23:23 tantek might even just start with showing a map with the blue dot, and when they load, highlighted polygons for the nearby venues so you could tap them to check into them, or long press to get their info in a hover bubble
# 23:23 tantek on any polygon with multiple venues, it shows a " > " menu icon on it
# 23:23 tantek and tapping it does the usual screenswipe left to reveal a checkie style menu of what's at that polygon
# 23:24 tantek for extra credit, tapping would do an auto-smart 3D zoom (including perspective shift as necessary) and show sub polygons lit up to tap on
# 23:30 aaronpk they show the lat/lng now! I used to have to install a google maps extension that added a right-click menu to show the lat/lng of a location
# 23:30 tantek yeah no idea why they were so stingy with that in the UI before (even with URL hacking etc.)
# 23:32 aaronpk ooh zoom levels are floating point now because vector maps
# 23:33 aaronpk ironically I was about to build that "drop a pin on the map and find the lat/lng" feature into Atlas
# 23:41 tantek just wait til they deploy time-indexed tiles with a slider
# 23:45 KevinMarks google uses their crawl cache because if you had most of the active web backed up in RAM, you would too
# 23:46 tantek speaking of which KevinMarks, I was just tweaking my mod of that
# 23:46 tantek One spider to crawl them all, one bot to retrieve them, one cache to index them all, and in the silos bind them.
# 23:48 KevinMarks One Ring to rule them all, One Ring to find them, One Ring to bring them all, and in the darkness bind them
# 23:48 tantek see here I was worrying first more about semantics, then second more about temporal order of operations
# 23:50 KevinMarks One spider crawls them all, one bot to find them, one cache to index all, and in the silos bind them.
# 23:50 tantek I think the keeping of the "to"s is mor important
# 23:51 tantek is it a subtweet if I don't deliberately reference Google?
# 23:52 aaronpk haha yes i believe that is actually the definition of subtweet
# 23:55 GWG kylewm, I need to know what he is running.
# 23:56 GWG At least once a month, someone tries to install the WordPress repository and the Github version of my plugins and I get a support complaint
# 23:57 GWG I have been meaning to put in an error handler for it.
# 23:58 GWG The Semantic Linkbacks issue may be something misconfigured. I need to know more details
KartikPrabhu joined the channel