#indiewebcamp 2015-04-29

2015-04-29 UTC
lukebrooker, KevinMarks, parzzix_ and parzzix joined the channel
#
tantek.com edited /nicknames-cache (+658) "/* Use Cases */ Better Person Name Display, and note Alternate Better Person Icon Display" (view diff)
#
tantek !tell rhiaro: I believe I captured the insights about updating a person's display name, but not necessarily their icon here: https://indiewebcamp.com/nicknames-cache#Better_Person_Name_Display
#
Loqi Ok, I'll tell her that when I see her next
tilgovi, parzzix, snarfed and tilgovi_ joined the channel
#
Loqi slack/snarfed: KevinMarks_: i changed bridgy to use evergreen twitter profile pic urls a while back
#
Loqi slack/snarfed: my bad for being stubborn and not doing that sooner :P
snarfed joined the channel
#
tantek thanks snarfed! that makes sense :)
KartikPrabhu, KevinMarks__, j12t, mlncn and j12t_ joined the channel
#
KartikPrabhu JonathanNeal: any chance of adding new syntax support to fragmention see Example here: http://indiewebcamp.com/fragmention ? Also found that my minification was causing trouble with the fragmention code for some reason
#
KevinMarks_ eh? which example - that works afaik
#
KartikPrabhu KevinMarks_ oh it does!!
#
KartikPrabhu oh shit!! fragmention.js is compatible with that already! sweet!!
#
KartikPrabhu JonathanNeal++
#
Loqi JonathanNeal has 16 karma
#
KartikPrabhu KevinMarks_ the %20 thing is hard to do manually, so having the + for spaces is more user-friendly for just typing I suppose
scor joined the channel
#
KartikPrabhu another fragmention question. Right now marginalia.js reproduces a bunch of code from fragmention.js to find the element corresponding to a fragmentioned response. Would it be a good idea to somehow attach the methods to decode the fragmention and to find the element to the DOM or something?
#
KevinMarks__ Jon was talking about modifying the dom for direct highlights iirc
#
KartikPrabhu KevinMarks_ yes, but I am talking about somehow exposing the functions that fragmention.js uses for other scripts like marginalia.js to re-use
#
KevinMarks_ ah, so more of a refactoring?
#
KartikPrabhu yes in some sense. I feel that the fragmention.js code is more "robust" and I'd like to directly use it. For instance I have somehow hard-coded the ## syntax into marginalia.js which I'd like to get rid off
wolftune, benwerd, lukebrooker, KartikPrabhu, lukebroo_, parzzix, tilgovi, tantek, torrorist, LCyrin, yakker, nloadholtes and snarfed joined the channel
#
tantek https://twitter.com/benwerd/status/593254787111526401
#
@benwerd Our little company is a year old - and we were working so hard today we didn't even notice. A cake will follow. (twitter.com/_/status/593254787111526401)
lukebrooker and loic_m joined the channel
#
kylewm snarfed: are you around? i'm looking at the twitter publish api issue ... i was able to post a tweet using the remote api console -- is that surprising?
#
snarfed kylewm: that runs code locally, so no, probably not
#
snarfed (probably not interesting)
#
snarfed thanks though!
j12t joined the channel
#
kylewm doh, does it really
#
snarfed i did find a few kindred spirits: https://twittercommunity.com/t/post-to-statuses-update-json-started-hitting-error-32-could-not-authenticate-you-with-no-code-changes/36495/6
#
snarfed kidneybingos did some serious investigation
#
kylewm yikes yeah, that is a person at their wits end
j12t_, j12t and j12t__ joined the channel
#
kylewm snarfed: any idea if running urlfetch.fetch in the remote api console actually goes through the real URL Fetch service?
#
snarfed kylewm: i think all the apis do
#
snarfed https://cloud.google.com/appengine/articles/remote_api#running
eschnou joined the channel
#
kylewm ok hmm, i was able to post through the urlfetch service
KartikPrabhu joined the channel
#
snarfed not sure though
#
kylewm :(
wolftune and loic_m joined the channel
#
tantek cheer-up kylewm!
wolftune, nloadholtes, fkooman, lukebrooker, Jihaisse, tilgovi, KevinMarks, csarven, eschnou, glennjones, petermolnar and pfefferle joined the channel
#
pfefferle good morning
KevinMarks__, KevinMarks, j12t, stream7, pfefferle, marjolein, Sebastien-L, KartikPrabhu, friedcell, adactio, LauraJ, frzn, catsup and gavinc joined the channel
#
KevinMarks_ what are comments
#
Loqi A comment is a kind of post that is in reply to some other post, that makes little or no sense without reading or at least knowing the context of the source post https://indiewebcamp.com/comments
barnabywalters, fkooman, friedcell and pfefferle joined the channel
#
barnabywalters Just released mf2/mf2 v0.2.10: https://waterpigs.co.uk/notes/4auAbf/ — thanks ben_thatmustbeme!
pfefferle and evalica joined the channel
#
@evalica The IndieWeb is a people-focused alternative to the ‘corporate web’ https://indiewebcamp.com/principles via @simonv3 (twitter.com/_/status/593373394768396288)
#
evalica spies :P
#
barnabywalters greetings evalica!
csarven joined the channel
#
evalica ty
#
GWG Good morning.
#
pfefferle GWG good morning
#
jonnybarnes barnabywalters: do you know there’s a typo in the mf2 readme?
#
jonnybarnes highlighted in red by github in this section: https://github.com/indieweb/php-mf2#parsing-microformats-from-a-html-string
#
GWG pfefferle: By the way, I borrowed some of your code again.
#
pfefferle GWG you are allowed to… but only you! ;)
#
pfefferle GWG what are you working on?
#
GWG This still - https://github.com/dshanske/indie-webactions
#
GWG Parsing was being problematic, so I tabled that for a bit to finish the web action functionality.
#
GWG But I added the code you started that adds indie-action tags to the reply links on comments.
#
@korayal @eminaksehirli @EmreSevinc @_anjinsan abi o yüzden vaktiyle ff'te bahsettiÄŸin #indieweb mevzusu pek mühim sanıyorum. (twitter.com/_/status/593382111131873280)
mlncn joined the channel
#
GWG Though, after the MF2 parsing thing, it occurs to me that a library to generate link previews from MF2 might be something useful to work on in future
pfefferle joined the channel
#
tommorris jonnybarnes: fixed - https://github.com/indieweb/php-mf2/commit/b5c85e9c235e03712bb33cc10347742c2a5b0b68
#
jonnybarnes tommorris woop
#
tommorris I didn't know I had commit access to php-mf2. I exercised my prerogative to 'be bold' as per http://enwp.org/WP:BOLD
#
tommorris was preparing to pull request it but then Github just let me commit directly.
#
jonnybarnes presumably anyone in the “indieweb” group can access any repo owned by that group, I don’t think github has more granular controls than that
#
barnabywalters thanks tommorris and jonnybarnes! Tpyo corrections always appreciated :)
#
pfefferle GWG ah ok, have seen that already!
#
tommorris that reminds me: a pre-commit handle for Markdown files that contain embedded JSON/XML etc. to validate them would be very useful.
#
tommorris adds to his backlog.
#
tommorris pre-commit.com already lets you check .json and .xml files in your repo to see that they are syntactically valid. would be lovely to have the blobs inside Markdown files checked too.
#
barnabywalters tommorris: I know of one project where outputs in the readme are generated by running the code samples, ensuring that both will always be correct
#
barnabywalters it’d be nice to make something similar for php-mf2
#
barnabywalters but not exactly a high priority
#
tommorris I know that's something that the Pragmatic Bookshelf folk do - their books are basically a giant script that starts with XML files and emits PDF files, and the code in them is compiled and run as part of the build process for the book
#
GWG pfefferle: I know. Still working on it.
scor and fkooman joined the channel
#
@orco_universe Passer á l'auto-hébergement : http://t.co/vXGD2z09PH, article sur le site du @gwadalug #yunohost #owncloud #ownyourdata (twitter.com/_/status/593408880924880896)
mlncn joined the channel
#
@jonnybarnes Just fixed a problem with my sites not receiving webmentions properly #retryallthethings (https://jmb.so/t/7d) (twitter.com/_/status/593411364351258624)
#
jonnybarnes if bridgy gets a certain number of errors in a row could it be set up to email you or something?
zero-gravitas joined the channel
#
@eminaksehirli @korayal @EmreSevinc @_anjinsan #indieweb daha çok web'i baÄŸlantılandırmak için diye biliyorum. Yedek alma üzerine de çalışıyorlar mı? (twitter.com/_/status/593413133563232257)
#
@psychoz971 RT @orco_universe: Passer á l'auto-hébergement : http://t.co/vXGD2z09PH, article sur le site du @gwadalug #yunohost #owncloud #ownyourdata (twitter.com/_/status/593413514284371969)
friedcell, chalettu, fourtonfish and snarfed joined the channel
#
fkooman aaronpk, do you prefer bearer authentication for the introspection endpoint or basic authentication?
#
snarfed jonnybarnes: interesting idea! the first catch is that bridgy doesn't have your email address :P
#
snarfed (and ideally doesn't really want it)
#
snarfed jonnybarnes: you might consider https://twitter.com/kylewmahan/status/590906891238178818
#
@kylewmahan enabled email notifications for >= ERROR log messages on my web server last night, learned about a bunch of stupid bugs this morning (twitter.com/_/status/590906891238178818)
#
fkooman aaronpk, i think basic authentication is simpler, so I'd prefer that :)
chalettu joined the channel
#
@benwerd This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyourdomain (twitter.com/_/status/593427639089532928)
#
@jmcclurken RT @benwerd: This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyour… (twitter.com/_/status/593429131502723072)
#
@thierrymarianne RT @benwerd: This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyour… (twitter.com/_/status/593431730306056193)
#
@funnymonkey RT @benwerd: This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyour… (twitter.com/_/status/593431831225053184)
zero-gravitas joined the channel
#
@jbpiacentino RT @benwerd: This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyour… (twitter.com/_/status/593432175900504065)
wolftune joined the channel
#
jonnybarnes !tell sanrfed thanks, will look into that
#
Loqi Ok, I'll tell them that when I see them next
modem joined the channel
#
ben_thatmustbeme i forget that I have desktop notifications of webmentions... suddenly i get a pop up and its like "oh yeah!"
pfefferle joined the channel
#
rhiaro ben_thatmustbeme: nice!
#
Loqi rhiaro: tantek left you a message on 4/28 at 5:12pm: I believe I captured the insights about updating a person's display name, but not necessarily their icon here: https://indiewebcamp.com/nicknames-cache#Better_Person_Name_Display http://indiewebcamp.com/irc/2015-04-28/line/1430266334242
j12t, tantek, snarfed, benwerd, mlncn and LauraJ joined the channel
#
aaronpk fkooman: i'm curious why you think basic auth is simpler
zero-gravitas joined the channel
#
fkooman aaronpk, the error handling by the client of bearer auth is more complicated than that of basic, other than that no reason :)
#
aaronpk how is it more complicated? you get an HTTP 4xx error in both cases
#
aaronpk i guess i don't see an obvious value for username+password for registered apps, whereas a bearer token "api key" style thing seems straightforward
#
tommorris new thing I've started publishing on my site: what I'm listening to - https://tommorris.org/tag/currently%20playing
#
aaronpk ooo
#
tommorris it's just manually posted. I've been listening to a lot of mixtapes/DJ sets at work recently and wanted to keep track
#
aaronpk nice
#
aaronpk i started doing the same (manually) with tv show episodes, it's just in a text file for now
#
tommorris eventually I might add a bookmarklet to post stuff from soundcloud/mixcloud/spotify
#
tantek what is a listen?
#
Loqi A scrobble (AKA a listen) is a passive type of post used to publish a song (music or audio track) that you have listened to https://indiewebcamp.com/listen
#
voxpelli indie-last.fm/audiuscrobbler :)
#
tantek tommorris: add yourself ^^^ /listen#IndieWeb_Examples :)
#
tantek and such "manual" posts definitely count and help inform more design :)
#
tommorris and listening to one of the mixes from Subsoul has prompted me to buy tickets to go to the album launch party next month at Lightbox.
#
tommorris.org edited /scrobble (+387) "adding myself, improving lead slightly" (view diff)
#
tommorris confounded markdown/mediawiki syntax
#
tommorris.org edited /scrobble (-1) "/* tommorris.org */ markdown/mediawiki mixup" (view diff)
#
aaronpk hehe yeah
#
fkooman aaronpk, okay, so i'll use bearer as well :)
#
fkooman aaronpk, it is just that for regular oauth 2.0 operations you'd have userid/password anyway, so you could reuse the registration table of clients for introspection as well...but well, it doesn't really matter :)
#
kodfabrik.se edited /link-preview (+148) "Adding Pinterest as another silo example + cleaning up mentions of Open Graph" (view diff)
#
tantek I don't understand how "regular oauth2 operations" implies user/pass in any way whatsoever
#
tantek they are orthogonal AFAIK
#
tantek e.g. you could use a one-time 10 digit code sent to your mobile device to authenticate into your own OAuth2 server on your site.
#
tantek user/pass is NEVER *necessary*
#
aaronpk i think he meant user/pass for the app
#
aaronpk client_id & client_secret in OAuth terms
#
fkooman exactly
j12t joined the channel
#
aaronpk but i'm trying to require that the client_id is the app's public URL
#
aaronpk for IndieAuth
#
aaronpk which means it doesn't require registration
#
fkooman reuse of those credentials is also mentioned in draft-ietf-oauth-introspection as an option
#
voxpelli kind of annoying that oauth 2 unlike oauth 1 is 3-legged only, and can't do 2-legged
#
fkooman i'll make my library support both, so no problem :)
#
aaronpk and the service can get the app's name and logo from the app's h-card on its home page
#
tantek agree with dropping registration requirements
#
fkooman yeah, that sounds kinda cool :)
#
aaronpk voxpelli: what do you mean?
#
tantek that's been an annoying antipattern in the adoption of OAuth2
#
voxpelli aaronpk: probably used the terms wrong, but in OAuth 1 you could use the same mechanism for communicating between client and server no matter if you had a user token or not
#
aaronpk voxpelli: that's the client_credentials grant
#
aaronpk http://stackoverflow.com/a/14251144
#
tantek this sounds specific / confusing enough (especially with the use of all these terms - n-legged, client_* etc.) to be worth adding to /OAuth
#
tantek if only as FAQs
#
tantek e.g. in the "what's a ***" style questions
#
aaronpk i'd rather not use the n-legged terminology anymore since that terminology was dropped in OAuth 2
#
aaronpk probably because it was confusing
#
voxpelli +1
#
voxpelli aaronpk: the client_credentials doesn't match what I believe https://oauth.googlecode.com/svn/spec/ext/consumer_request/1.0/drafts/2/spec.html was describing
#
voxpelli but that wouldn't really be possible to do with OAuth 2 + Bearer tokens, as the client-credentials are now used completely separately from the tokens, rather than used together as in OAuth 1
#
aaronpk voxpelli: i'm not sure what you're asking
#
aaronpk the client credentials grant is a way for an application to get an access token to make requests on behalf of itself
#
@grantpotter RT @benwerd: This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyour… (twitter.com/_/status/593455332837425153)
#
aaronpk if you want a way for applications to obtain an access token for a *user* without asking the user, that is not part of OAuth 2 core, but can be done with an extension grant type
#
tommorris so, PESOS from Mixcloud: http://api.mixcloud.com/tom-morris5/feed/ - my 'feed' on the site. wouldn't be hard to pull that in automatically, so when I 'repost' (i.e. 'retweet') a show on there, it'd post it on my own site.
#
voxpelli aaronpk: 2-legged was a way to communicate between a client and the server when there were no users, there's really no such thing with bearer tokens, no question really, just an observation
#
aaronpk voxpelli: that is exactly what client_credentials grant is
#
aaronpk the app gets a bearer token that is only for itself, not for a user
#
tommorris.org created /Mixcloud (+346) "new article" (view diff)
#
voxpelli aaronpk: yes yes, but client_credentials aren't bearer tokens – they are something you use with basic auth (right?) – so while it is possible to still do client-to-server communication without a user, it requires a separate mechanism
#
aaronpk voxpelli: no
#
aaronpk the client_credentials grant type lets apps get an access token using their credentials
#
aaronpk http://aaronparecki.com/articles/2012/07/29/1/oauth2-simplified#application-access
#
aaronpk from that point forward, the app has a bearer token it can use at the API server
#
voxpelli sure, one can get oneself a token that represents the client rather than a user – that was possible with OAuth 1 as well
#
voxpelli that's an extra step though
#
aaronpk yes that's the point
#
aaronpk so that API servers don't need to know about client ID and secret
#
aaronpk that was one of the major design considerations in replacing OAuth 1
#
voxpelli yeah, I know
#
voxpelli can be annoying from a client perspective even if it makes sense from the server perspective :) and therefore basic auth looks pretty attractive for clients
#
voxpelli usual tradeoffs, not saying it was right or wrong
#
aaronpk in practice, some services solve this by giving developers an access token for their app from within the developer console
danlyke joined the channel
#
aaronpk twitter even gives people a token for their own user account so you can make API requests without doing the OAuth flow
KevinMarks__ joined the channel
#
aaronpk i feel like that's a pretty good solution, it essentially makes it the same experience as using just the client credentials directly
#
voxpelli yeah, I like that, makes it easy to get going – good DX
#
aaronpk best of both worlds I think
#
voxpelli +1, and also makes sense as bearer tokens really is an independent spec from OAuth if I don't remember things wrongly
#
aaronpk correct
barnabywalters joined the channel
#
tantek great post about Known and reclaiming yourself yourstuff: http://audreywatters.com/2015/04/29/reclaim-known/
KevinMarks joined the channel
#
adactio Friends—I was thinking of maybe throwing in an Indie Web Camp the weekend after Responsive Day Out. Responsive Day Out is on Friday, June 19th so that would mean Indie Web Camp on June 20th and 21st here at 68 Middle Street in Brighton.
#
adactio If I build it, will you come?
#
adactio I'd like to gauge interest before committing.
#
waterpigs.co.uk edited /Posts_about_the_IndieWeb (+107) "/* 2015 */ added Reclaim Your Domain (With Known)" (view diff)
snarfed joined the channel
#
waterpigs.co.uk edited /Posts_about_the_IndieWeb (+7) "/* 2015 */" (view diff)
#
barnabywalters rhiaro: Saw you had some problems with php-mf2 in the logs — sorry to hear about it, did you get it working in the end? I just revised the installation and usage documentation and would appreciate your feedback as a new user!
#
aaronpk barnabywalters: the main problem was curl wasn't installed :)
#
barnabywalters ah yes that would explain it
#
aaronpk but I do think a simpler example of using the library without composer and in a procedural way would be useful for the docs
#
barnabywalters aaronpk: procedural? as in non-OO? The docs have prioritised simple function calls over OO usage ever since they were added
#
tantek adactio - would love to but I'm in Vancouver that whole next week which makes the travel impossible :(
#
aaronpk barnabywalters: there's no require() example for hwo to load it, i think that would help
#
barnabywalters earlier I revised https://github.com/indieweb/php-mf2#installation and https://github.com/indieweb/php-mf2#usage
#
tantek tommorris, rhiaro - would you go to an IndieWebCamp Brighton 20-21 June?
#
adactio tantek: yeah, I saw your dates on the wiki so I know it's a tricky time.
#
tantek the weekend before that is YxYY which I'm still on the fence for (have not bought a ticket)
wolftune joined the channel
#
tantek adactio: happy to see you're making it to IWC Germany!
#
adactio tantek: Yeah, I think that's going to be *great*—the venue looks really good.
#
tantek adactio: what did you think of the 11-12 July dates for IWC Portland?
#
tommorris tantek: probably.
#
adactio tantek: I don't think I'll be able to make it over this time, alas (I blew all my air miles on a trip to Japan earlier this year)
#
tantek adactio - another possibility for 11-12 July would be Brighton as another simultaneous venue
#
tantek since so far no other simultaneous venues have stepped up yet
#
tantek (last year was PDX, NYC, Berlin)
#
adactio tantek: yeah, I thought about that. We'd miss out on piggy-backing on Responsive Day Out, but then again, maybe there wouldn't me much crossover in interest.
#
adactio tantek: i.e. the people coming to Responsive Day Out might not be interested in Indie Web Camp and visa-versa.
#
tantek hmm - IndieWebCamp has tended to attract a more user/design-centric crowd than other "distributed / decentralized [social] web" type meetups
#
tantek however I think you're right
j12t joined the channel
#
tantek adactio - also - there's no Aral July conf this year right? So maybe IWC Brighton 11-12 July could help fill that gap.
#
adactio tantek: true. I'll have a ponder on this. Looks like the venue is available on those July dates in any case.
#
barnabywalters aaronpk: I added a code block in the non-Composer section demonstrating basic non-composer usage: https://github.com/indieweb/php-mf2/blob/master/README.md#installation
#
barnabywalters hopefully that, along with the other improvements, will make things easier for new users
KevinMarks joined the channel
#
tantek barnabywalters++ thanks barnabywalters!
#
Loqi barnabywalters has 80 karma
#
tantek hopes to be a new user of php-mf2 one day
#
voxpelli adactio: having the annual Brighton meetup at June 20th and 21st? Not sure I can make it, will move to a new apartment around then and a bit like that – planned to go to the Brighton one this year otherwise though
KevinMarks__ joined the channel
#
adactio voxpelli: So July 11-12 would work better for you?
#
barnabywalters is IWC UK still planned for September this year? I’m unlikely to be able to come to one in the UK in summer
#
voxpelli adactio: yes, that would be better for me
#
adactio voxpelli: Cool. Good to know.
#
tantek barnabywalters: I'd be up for the september one too
KevinMarks___ joined the channel
#
tantek the July 11-12 is the main annual IWC - with multiple locations (PDX, … )
#
barnabywalters ooh, I *might* be able to go to a Berlin July IWC then — rhiaro is listed as the only person interested currently on the wiki: http://indiewebcamp.com/2015#Confirmed_Cities
#
tantek.com edited /2015 (+168) "/* Candidate Cities */ add Brighton with adactio, voxpelli; add barnaby to Berlin - all per IRC, sort "o" organizer votes to top of each city" (view diff)
#
tantek barnabywalters: presumably you meant this link: https://indiewebcamp.com/2015#Candidate_Cities
#
barnabywalters ah, yep, that one
#
adactio voxpelli: I'm not sure if there will be an Indie Web Camp in Brighton this September. This year that's the weekend that Clearleft will be celebrating its 10th birthday so there may not be time for an Indie Web Camp.
KevinMarks joined the channel
#
tantek adactio - makes sense. even last year you pointed out how hard it was to do both dConstuct and IWC back to back
#
tantek another weekend during September might work though
#
tantek (and goodness, 10 years of Clearleft - wow 2005 really was a watershed year)
#
barnabywalters wow, congraulations adactio for 10 years of Clearleft!
#
adactio barnabywalters: thanks!
#
adactio tantek: yeah, there's a lot of 2005 nostalgia going round this year. ;-)
#
voxpelli adactio: congrats! good to know about September
LauraJ, frzn, pfefferle, LaurieJ, friedcell, zero-gravitas, KevinMarks, lukebrooker, fkooman, scor, fkooman_ and eschnou joined the channel
#
@kevinmarks RT @benwerd: This piece about Known makes me so happy. Owning your space is about being free. http://audreywatters.com/2015/04/29/reclaim-known/ #indieweb #reclaimyour… (twitter.com/_/status/593494987796512769)
scor joined the channel
#
kylewm watching Eric Meyer's talk at Fluent Conf from last week https://www.youtube.com/watch?v=r38al1w-h4k
#
kylewm lotta js;dr parallels
#
tantek yup
#
tantek we chatted a bit about that during/after HWC last week
#
kylewm lol, one of his slides is just a screenshot of instagram with NoScript
#
tantek nice
#
tantek which only proves you can sell a company for $1b without supporting JS-off
#
kylewm did they even have much of a website at all when they sold?
#
kylewm instagram was iOS-app-only for a long time right?
#
aaronpk yeah i think the web thing was after FB, but not sure
#
tantek good questions
#
tantek creating web access to the content *after* acquisition is certainly the opposite of most site-death acquisition scenarios!
#
loqi.me created /Foundation (+84) "prompted by tantek and dfn added by KevinMarks_" (view diff)
#
KevinMarks_ their website was "go download the app" at the time - they'd just added the android app
#
KevinMarks_ there was an api, but no web ui at that point
#
KevinMarks_ there still wasn't much of a web one when I hacked up pestagram.com
fourtonfish, friedcell, KartikPrabhu, zero-gravitas and LauraJ joined the channel
#
KartikPrabhu for avatar and mixed-content issues: https://carlo.zottmann.org/2013/04/14/google-image-resizer/ ?
#
KartikPrabhu found via adactio's links of course
j12t joined the channel
#
KartikPrabhu also of possible interest here: http://morning.computer/2015/04/why-facebook-hates-me/
scor joined the channel
#
KevinMarks_ if you want to try out the to-do app that is indieweb friendly, go to http://app.willsomeone.com/
#
KevinMarks_ and send me feed back
#
aaronpk feedback #1: let me sign in with my domain instead of google
#
KevinMarks_ yeah yeah
#
KevinMarks_ an indieweb replacement for appengine's user api would be a fine thing
#
KevinMarks_ maybe I'll hack on thta next
sammachin joined the channel
#
KevinMarks_ aaronpk: if you sign in, you can make that a public to-do
#
KevinMarks_ :D
#
snarfed KevinMarks_: cool! we could maybe add it to https://oauth-dropins.appspot.com/ . it has all the scaffolding already
#
sparverius win 35
#
KevinMarks_ ooh, handy snarfed
#
KevinMarks_ does that have a way to get names and profile pictures (which the appengine user mode fails at completely)
#
tantek KevinMarks - that's called - get p-name and u-photo from the representative h-card of the author's URL
#
tantek s/author/user
#
Loqi tantek meant to say: KevinMarks - that's called - get p-name and u-photo from the representative h-card of the user's URL
#
snarfed tantek++
#
Loqi tantek has 183 karma
#
snarfed KevinMarks_: it does for some services
#
aaronpk that's what the indieauth OmniAuth gem does!
#
snarfed it lets clients provide arbitrary scopes
#
KevinMarks_ I know how to do it properly, it's google's built-in user api that is wierdly crap
#
KevinMarks_ it gives you their email, but not their name
fkooman joined the channel
#
snarfed ah, you mean the app engine users api
#
snarfed this doesn't use that
#
snarfed but i think the G+ drop-in fetches name
#
snarfed yup it does, try clicking the G+ button on https://oauth-dropins.appspot.com/
#
KevinMarks_ what I was thinking was to have a replacement for the app engine users api that supports indieweb login and adds helpful things like name and phot
#
snarfed KevinMarks_: right
#
snarfed by replacement, do you just mean a python lib that works on app engine? or specifically a lib with a compatible API so you can port existing code with minimal/no change?
#
KevinMarks_ yes, just swap "from google.appengine.api import users" for "from indieusers import users"
#
snarfed how would porting work? any app with existing users couldn't port because they'd block those existing users, right?
#
KevinMarks_ well, in theory existing apps can have URLs as well as email addresses 'cos that api used to support openid
#
KevinMarks_ so you can distinguish based on email versus url
#
KevinMarks_ https://cloud.google.com/appengine/docs/python/users/
#
snarfed :P ok so it'd work if your existing users are all email-based
#
snarfed ok
#
snarfed if that's your specific use case, then yeah, ignore oauth-dropins
#
snarfed i'd still also love to have indieauth in oauth-dropins though!
#
KevinMarks_ hm , now they've dropped openid, do federated_identity() go away? https://cloud.google.com/appengine/docs/python/users/userclass
tilgovi, benwerd, zero-gravitas, lukebrooker, j12t, KartikPrabhu and loic_m joined the channel
#
aaronpk another way to handle the problem of URLs not mapping cleanily to filesystem names for /IndieArchive could be to store the files on disk with a name that is a hash of the URL
#
aaronpk the filesystem would be a mess of filenames, but it would work
scor joined the channel
#
aaronpk might be a good idea to keep an accompanying .meta file for each, which could store the original URL as well as things like the HTTP headers from the request
LCyrin joined the channel
#
aaronpk somebody talk me out of this
#
KevinMarks_ if you're going that far just use a hash based filesystem
#
KevinMarks_ Brad ahs one you can use in camlistore
#
KevinMarks_ is the problem when you crawl things with ? & in the paths?
#
KevinMarks_ hi Lynn
#
aaronpk the ? and & and othe rweird chars can cause problems
#
aaronpk but more criticall the problem is when part of a path is both a file and a folder
#
aaronpk http://aaronparecki.com/tags http://aaronparecki.com/tags/indieweb
#
KevinMarks_ your index.html workaround sounded a good one
#
aaronpk but then I can't store http://aaronparecki.com/tags and http://aaronparecki.com/tags/
#
KevinMarks_ when are those usefully different urls?
#
KevinMarks_ http://microformats.org/wiki/rel-tag#slashes+in
#
aaronpk i don't think making an assumption like that is a good idea for a long-term archive
#
KevinMarks_ this is probably worth talking to internet archive people about
#
aaronpk oh yeah, here's an example: http://indiewebcamp.com/2015 vs http://indiewebcamp.com/2015/
#
aaronpk web.archive.org has both pages separately
#
KevinMarks_ what is 2015
#
Loqi https://indiewebcamp.com/2015/
#
aaronpk now i'm thinking a path like archive/example.com/DDD/SSS/HASH_OF_PATH would work well on disk
#
aaronpk right now I do archive/example.com/path/to/file.html (which has already caused problems)
yakker joined the channel
#
tantek it loses the inspectability advantage
#
tantek aaronpk here is a dumb answer: use "tag" for the name of the file, and use "tag/" as the name of the folder (no seriously)
#
tantek since you can detect the "/" on the end of the name, and then conclude it's a folder not a file ;)
#
aaronpk i'm missing something... folder names on disk can't have '/' in them
yobj joined the channel
#
aaronpk oh another problem with using the path directly, is some filesystems are case sensitive and others are not
wolftune joined the channel
#
aaronpk so on OSX, indiewebcamp.com/Events and indiewebcamp.com/events clobber each other even though they are unique URLs
#
gavinc aaronpk: I assume you've looked at WARC and found it wanting already? It takes care of a large number of these sorts of issues
KartikPrabhu joined the channel
#
aaronpk it's been a while.. let me refresh my memory
#
gavinc wget already spits them out, keeps headers, etc
#
aaronpk i think i was never able to find good docs or code for actually reading these files later
#
aaronpk various github repos were last updated >2 years ago
#
aaronpk what would be great is if any of these projects gave an example of 1) creating a warc file for a specific URL, and then 2) outputting the HTML and headers from that warc file for a given URL
#
aaronpk there's plenty of examples of tools that create warc files, but i can't find anything that says what to do with them later
#
gavinc oh, I've just used the python library
#
gavinc never had any issues with it
#
aaronpk yeah, reading the github and readthedocs page doesn't really give me any sense of how to do that
#
aaronpk literally no examples of finding the contents of a URL here http://warc.readthedocs.org/en/latest/
#
gavinc WARC-Target-URI ;)
#
gavinc which has a short name of target_uri ;)
#
aaronpk so to find a specific URL i have to "for record in f" and find the record['WARC-Target-URI'] that matches the URL i'm looking for?
#
gavinc yep, or given the file format create an index into them after scanning them all once
#
aaronpk yeah none of that is obvious
#
aaronpk it just seems like warc is for a different use case
#
gavinc Mmm, but it already HAS the complete archiving story solved
#
aaronpk and as far as i can tell, there is only a python library
#
gavinc wget produces them, as does the IA
#
aaronpk for reading
#
gavinc I mean, it's a GZIP'd text file
#
gavinc and the responce part is literally an HTTP responce
#
aaronpk i can't seem to add a file to an existing warc archive with wget
snarfed joined the channel
#
tantek aaronpk - but in OSX you can put a "/" in the filename in the Finder
#
tantek since you mention OSX
j12t joined the channel
#
tantek case-insensitivity is a different problem
KartikPrabhu joined the channel