2013-04-11 UTC
josephboyle, bnvk, duckbillp, spinnerin, b0bg0d, smus and jedahan joined the channel
smus, jedahan, b0bg0d, duckbillp, eschnou, bnvk, Nadreck, seyz, friedcell, andreypopp, barnabywalters, christopheduca-1, tantek, indiewebcamp-vis and scor joined the channel
# 12:20 tommorris app.net is done by parsing the RSS feed and updating the database
# 12:20 tommorris which is currently done by me triggering the parse command in a Ruby shell. eventually, it'll be a cron job
# 12:21 tommorris each item has a link back to the source <atom:link href="whatever" type="text/html" rel="source" />
# 12:22 tommorris so it's just a matter of getting each item, matching them up using that, and then retrieving the URL of the app.net post
# 12:24 barnabywalters ah, of course — so you’re (ab?)using the posted from field to link back. That’s a great approach
# 12:25 barnabywalters much more interesting than linking to the app used, which I think (?) was the purpose?
# 12:26 tommorris I'd know more about how it worked if I became an App.net developer, but $99 buys a lot of cocktails.
# 12:26 barnabywalters tommorris: that you’ve got POSSEing with backlinks set up without forking out for the dev price is impressive
# 12:27 tommorris the free account doesn't let you follow more than 30 people IIRC
# 12:27 tommorris but I don't really care about following people. I have an RSS reader. ;)
# 12:29 tommorris the idea is that you'll hit 30 people and become so addicted you'll pay the $30 or whatever to upgrade
bnvk and b0bg0d joined the channel
andreypopp, b0bg0d and josephboyle joined the channel
# 15:37 tantek barnabywalters, what was the problem with t.co?
b0bg0d joined the channel
# 15:47 tantek.com edited /Twitter (+658) "fix heading hierarchy a bit, not everything is about porting to twitter, POSSE Replies fixes, distinguish POSSE Replies to Tweets and POSSE Replies to Twitter" (
view diff )
seyz joined the channel
# 16:04 tommorris so, app.net's Atom feed uses rel="source" to link from the ADN version to the original version. it'd be nice if their HTML did likewise. ;)
# 16:05 tantek it really should use rel=canonical in the HTML version
# 16:05 tantek although the example you gave only showed your blog post title on ADN
# 16:06 tantek whereas you had a complete post on your own site
# 16:06 tantek I thought from the ADN post that it was only a short note kvetching about mobile internet
# 16:06 tantek it's not obvious that I should click through to an entire blog post
eschnou joined the channel
# 16:06 tantek that's a case where it may make more sense to include the full URL to your blog post inline in the post
# 16:07 tommorris yeah, for posts with titles, I just syndicate the title to ADN. for posts without titles, I sync the first ~400 chars (whatever the ADN limit is)
# 16:08 tommorris I should have a public list on indiewebcamp.com of things I should fix to be more indieweb-compliant
# 16:15 tantek tommorris, what's the ruby code for? I thought your POSSEing to ADN just used IFTTT?
# 16:15 tantek parses which Atom feed and updates which posts?
# 16:16 tommorris ifttt periodically polls my Atom feed and syndicates it to ADN
# 16:17 tommorris then updates the post on my site to link to the ADN version
# 16:18 tommorris which seems very Heath Robinson, but such is the problem of silos
# 16:18 tantek a-ha because the IFTTT runs sometime later, and you have to explicitly do work later to go get the syndication permalink
# 16:19 tommorris in an ideal world, I'd send a copy directly to ADN in a message queue from my site
# 16:19 tommorris but that'd require an ADN API key, and I'd rather spend the $100 on tequila.
# 16:26 tantek.com edited /App.net (+906) "/* POSSE to ADN */ better sumary and flow reinterpreted from TomMorris description in IRC, also pay alternative" (
view diff )
texburgher, spinnerin, duckbillp, seyz, scor, bnvk and b0bg0d joined the channel
# 19:10 tantek aaronpk, barnabywalters, here's a good example of a full fledged article post which is clearly a direct reply to another article post. In this case, copying the first paragraph of the reply article as a comment on the original article (with a "view more" hyperlink to the permalink) would work quite well.
# 19:11 tantek so perhaps that would be a good heuristic for abbreviating an article (without an explicit p-summary) when copying into a "Comments" section - use the first paragraph.
# 19:23 tantek and figure out a way to use the methods described with RelMeAuth to implement TFA (Two Factor Auth) with your own domain
# 19:40 tommorris tantek: I used to do two-factor-auth to log in to my own site. implemented it myself in JavaME, back when that was a thing.
morrocco_mole joined the channel
# 19:41 tommorris I just sha1'd secret string + Unix epoch / 100 (so it'd refresh every 100 seconds)
# 19:41 tommorris then the iPhone happened and I couldn't be bothered to port it over to a smartphone platform.
bnvk joined the channel
# 19:46 tommorris and... someone does a password reset on my OpenID provider. I remember why two factor auth seems like a good idea.
# 19:46 tantek whoa - really? someone is attacking your OpenID provider?
# 19:47 tommorris myopenid seemed to fail too much so I switched over to the wp account I have
# 19:48 tantek is their two-factor auth open sourced somewhere?
# 19:48 tantek would be great to figure out how to adapt that for any indieweb site
# 19:49 tantek sure, it uses GA on the backend, but I mean the UI/glue code
# 19:49 tommorris en.blog.wordpress.com/2013/04/05/two-step-authentication/
# 19:49 tantek oh hey, 2 letter language code subdomains, where have we seen that anti-pattern before ;)
# 19:54 tommorris tantek: implementing 2FA isn't too hard. not sure about Google Auth, but having written my own implementation, it can't be too hard. ;)
# 19:55 tommorris the hardest thing about writing my own 2FA app was that SHA1 isn't part of the JavaME API, so had to write my own implementation of SHA1
# 19:55 tommorris which is just a matter of taking the algorithm and writing it in Java.
# 19:56 tantek tommorris - is perhaps something (that algorithm implementation) that should go into the W3C Crypto API?
# 19:57 tommorris would rather if nothing he'd written crypto-wise would ever run on anybody's devices ever.
# 19:58 tommorris standardising hashing and crypto in JavaScript seems sensible.
# 20:00 tommorris I'd say it seems fairly straightforward, but I've read W3C specs like the old RDF C&AS doc and found it fairly clear while others found it completely imponderable
# 20:01 tommorris especially as I tend to use words like "adicity" without a second thought.
b0bg0d, spinnerin, eschnou, andreypopp and bnvk joined the channel
# 21:03 tantek I wonder what an indieweb version of that would look like
barnabywalters, andreypopp, b0bg0d, friedcell and tantek joined the channel
tantek joined the channel