2013-05-09 UTC
# 00:00 reidab oh, nevermind, I see what you mean
b0bg0d joined the channel
# 00:41 bret aaronpk: did you ever catch the discussion about using mozilla personas as an indi-auth provider?
# 00:53 bret adding rel= to email links, and comparing that to the persona authentication results
scor, bnvk and spinnerin joined the channel
# 02:06 aaronpk what comes back from the persona authentication? I'd need bi-directional verification of url==email matching
# 02:10 aaronpk in other words, if example.com has a rel-me link to example@gmail.com, I'd need to know that example@gmail.com claims to be example.com
duckbillp, tilgovi and tantek joined the channel
# 02:59 bret aaronpk, im not totally sure, but I think you get an email back.
# 02:59 bret i have not had time to get into the technical details, but conceptually, it makes sense
# 03:01 aaronpk yes, I'm aware of the design, it's nice that it's actually distributed from the beginning, using mozilla's service as a shim
# 03:02 aaronpk i'm pretty sure it would need the bi-directional link though. If it doesn't, then we shouldn't require bi-directional verification for web providers like Twitter.com either.
# 03:03 aaronpk which means the basis of RelMeAuth would need to be updated
# 03:14 bret i have no idea :? I'm still trying to play catch up with getting PuSH and webmentions working with github pages commit hooks
# 03:14 bret outgoing webmentions a little less easy
# 03:14 bret and incoming is going to have to be some external tool im not sure what to do about that yet
# 03:14 aaronpk it's not really set up as a nice service yet, but it has an API and I can give you an account to play with
# 03:17 aaronpk updating the readme since it's a bit out of date...
# 03:27 bret jekyll can create static feed type documents easily, and github can pass messages about git commits easily, so I think outgoing messages should be pretty simple
# 03:28 bret getting incoming messages into the git repo is going to require some work though
# 03:29 bret like if someone sends you a response or a webmention, how to get that data back into a static website based in git
# 03:30 bret ill have to think about that for a while
# 03:30 aaronpk you'll probably need a server somewhere that can edit the git repo
# 03:31 aaronpk a partial solution would be to query the pingback.me API from the browser and display things inline that way. of course the markup wouldn't exist in the page server-side, but the user would see it.
# 03:31 bret with a small script? Yeah that would be pretty cool, and realtime actally
# 03:32 bret that might actually work really well just to get something running
# 03:33 bret i display a live pinboard feed with JS
# 03:36 aaronpk I think I might add in microformats parsing into the pingback service, to pull out basic h-entry data from the linked pages like how news.indiewebcamp.com does
# 03:37 bret I went with microdata when i made the site :/ kicking myself now
bnvk joined the channel
# 04:59 bret you think the weather is going to still be this nice in portland by june?
duckbillp joined the channel
# 05:24 aaronpk I get in the habit of adding the pingback/webmention tags to any site I make now, even if I don't have plans to display comments on the site
# 05:25 bret I want to totally redo my html templates... lots of effort went into the microdata, and its basically used for nothing... plus its got twitter bootstrap garbage all over
# 05:35 bret but its part of the HTML5 spec! I though to myself
tantek joined the channel
# 06:07 aaronpk ... experimenting with embedding content inline in notes
# 06:07 aaronpk ironically started because I wanted to share a text file in a note
# 06:18 aaronpk figured out a clever trick for my linking problem too
# 06:18 aaronpk I don't want the http:// visible, but I do want it in the microformat data so it gets syndicated out properly, so I'm hiding the protocol with CSS now :)
# 06:29 bret aaronpk, Pingback is the total missing link for static sites I think
tilgovi joined the channel
# 06:40 aaronpk but there should be a way to accomplish the same thing with microformats I feel
# 06:48 bret so now write some JS that lists the last N mentions
# 06:49 aaronpk unfortunately you'll only be able to display the URL right now since that's all the API gives back, but I can add some basic mf2+html parsing to provide a title property as well
# 06:50 aaronpk will give me a chance to test out the ruby mf2 parser as well
# 06:51 bret I'm going to jump over and work on PuSH since I know that already works
tantek-ipod joined the channel
# 06:56 bret is there a way to get some test mentions?
# 06:58 aaronpk then you can send a mention from that page to your site
# 07:01 bret {"error":"source_not_found","error_description":"The source URI does not exist"}
# 07:02 bret man that www is starting to look really ugly right about now
# 07:02 bret I can do tld... but I wasn't sure what the best thing to do was
# 07:03 bret which seemed weird but im a total novice at DNS
# 07:03 aaronpk yea DNS doesn't let you define CNAME records on TLDs
# 07:03 bret you have to do an A record or something?
# 07:04 bret well since I am working to make URL's a login name in this world, I might as well get it right before i go much furthur
# 07:09 bret i'm just crossing my fingers that the .tc domain prices go back down
# 07:09 bret i wanted a 3 letter domain that was my name you see :)
# 07:11 aaronpk ok now i'm just getting a bit crazy with this embedded media stuff :)
# 07:12 bret can you superimpose your geotagged photos onto that streetview?
# 07:12 aaronpk eventually I want to do inline previews of web pages like twitter/facebook, but not tonight
# 07:16 aaronpk for now: Gists, Vine, Spincam, Youtube, Vimeo, Google Maps, Soundcloud, Slideshare, Images and Video
# 07:19 bret buhhhhhhhh... gh-pages takes 10 minuts to update their redirects
# 07:25 bret shucks. forgot to update my linkback URLs
# 07:29 bret {"error":"no_link_found","error_description":"The source URI does not contain a link to the target URI"}
# 07:29 aaronpk looks like you linked to www.bre.tc on indiewebcamp
# 07:32 aaronpk that doesn't seem right... it should treat no trailing slash as a trailing slash, yea?
# 07:32 bret i mean, yeah it would be nice if that didnt matter
# 07:36 aaronpk alright well that was way more than I expected to do tonight
# 07:38 aaronpk depends on if you think running ruby apps is easy
# 07:38 aaronpk can probably run it on heroku but I haven't done that yet
# 07:39 bret i consider znc the pinnacle of easy server apps
# 07:41 bret basically, something you can crontab open at startup
tantek-ipod and duckbillp joined the channel
# 08:05 aaronpk ... coming up on 21 consecutive hours awake. oops
friedcell joined the channel
barnabywalters joined the channel
adactio and xtof joined the channel
mxuribe1, scor, bnvk, mxuribe, xtof, brianloveswords, eschnou, texburgher, singpolyma, bret and hober joined the channel
# 15:08 tantek re: oEmbed, I think we can achieve nearly all of it with h-entry these days
# 15:08 tantek if we do add u-photo which has been proposed, that takes care of the thumbnail needs
# 15:08 tantek no need for a separate oEmbed endpoint to query, just get the h-entry from the page itself that you want to embed
barnabywalters, tilgovi, tantek-ipod, tantek-ipod_, spinnerin, eschnou, mxuribe, bnvk and mxuribe1 joined the channel
# 18:09 tantek barnabywalters, aaronpk, just as we've been informally using h-as-note and h-as-article, I think it's reasonable to also use h-as-rsvp
# 18:10 barnabywalters might be an idea to make a distinction between “publishing” and “using” — we’re publishing them, are any of us actually using that semantic for anything?
# 18:12 tantek also it has potential utility if someone wants to write an h-entry -> ActivityStreams(Atom or JSON) converter
# 18:12 tantek are either of you using anything like h-as-reply or h-as-comment on your replies/comments posts?
# 18:16 tantek ok I'm not either, though I'm not posting "real" replies/comments (yet)
# 18:16 tantek (just unlinked @-reply notes once in a while)
# 18:16 aaronpk what's h-as-note for? I don't think I have that on my site yet
# 18:19 tantek aaronpk - they're more useful/relevant when you have a composite stream of all your different post types
# 18:20 tantek but since you have separate stream pages for each post type on your site - not sure how much difference it makes for you (if any)
gRegorLove joined the channel
# 18:37 Loqi gRegorLove: tantek left you a message on 5/7 at 1:26pm: why are you display:none'ing your Twitter link on gregorlove.com? I don't think that affects IndieAuth but it might? aaronpk - gregorlove.com appears to be setup with rel-me to @gRegorLove but isn't working. Take a look?
# 18:37 tantek did you get indieauth sign-in working? or still same problem?
# 18:38 gRegorLove I was just revisiting to get your response. Let me try again. I have it display: none because I don't want to add links to the home page, if possible.
# 18:39 gRegorLove It was finding the link to my Twitter fine, but appeared to not be finding the link back on my twitter profile, but it is there. :)
# 18:39 aaronpk ah the problem is twitter.com is wrapping your link in a t.co
# 18:41 barnabywalters gRegorLove: I noticed you have a github account — github are a really solid auth provider
# 18:43 gRegorLove Looking forward to getting involved more. I use Nucleus CMS and I plan to make a plugin for webmentions.
# 18:51 gRegorLove I don't see a link to my user page on the wiki, but I presume it's User:Gregorlove.com ?
shaners joined the channel
# 19:08 shaners for example, why can i only sign in from the homepage
# 19:08 aaronpk there's a "log in" button in the footer of every page
# 19:10 shaners why does every section BUT the "this article is a stub..." have an [edit] link?
# 19:10 shaners i mean, i found the Edit Page link in the footer, but just seems weird
# 19:15 shaners tantek aaronpk et al: has anyone ever investigated gollum as a potential replacement for mediawiki powered wikis?
# 19:16 tantek shaners - every complaint you've made is about the presentation/layout - not about the engine
# 19:16 tantek so it's a nonsequitor to go from presentational issues -> suggesting backend change
# 19:16 tantek I too would prefer the edit/history links at the top the way mediawiki installs normally do it
# 19:17 shaners totally, non-sequitur. i'm sorry, i was unclear. i just remembered gollum while thinking about wiki stuffs.
# 19:17 shaners not suggesting a switch. just asking about any experience with it.
# 19:18 gRegorLove I put Nucleus under "experimental" because as far as Indie Web it is. As a blogging engine, though, it's pretty mature.
# 19:19 shaners aaronpk: your embedded content post looks good. great job on that.
# 19:19 tantek aaronpk - interesting that we've taken different approaches to embedded content
# 19:19 tantek you're putting embedded content *after* the post footer, in a block by itself
# 19:20 tantek whereas in my notes, embedded content auto-embeds inline where it's mentioned
# 19:20 tantek I do it inline because I think that works/reads/views better when you have *multiple* embeds
# 19:21 aaronpk you also remove the visible URL and replace it with the image, right?
# 19:21 tantek aaronpk - right - a visible .jpg/.png/.gif URL is just visual noise. I do hyperlink the image to the original URL
# 19:21 tantek also - try resizing the window on that note post ;)
# 19:23 aaronpk although how do you handle making the mf2 representation display the URL instead of the <img> tag?
# 19:24 aaronpk the mf2-parsed page doesn't have the image URL, it just omits that part of the note
# 19:24 tantek oh that's an interesting issue. I'll have to think about that.
# 19:25 tantek I wonder if p-* parsing should extract the "alt" attributes from images
# 19:26 aaronpk and since we're all pretty much standardising on plaintext for notes, the plaintext version should be something I can copy to my site as plaintext and keep the full resolution text
# 19:27 tantek so like for example when you're pulling it in for a reply post
# 19:28 tantek if we parsed nested <img> tags' alt (and then src) attributes for p-* elements, that could work
# 19:28 tantek then if you explicitly didn't want any img content showing up you could just set its alt=""
# 19:29 tantek barnabywalters how hard would it be to add this to the php-mf2 parser:
# 19:30 tantek in addition to parsing out the nested plain text, if you encounter any <img> elements, replace them with their alt attribute text, or lack an alt attribute, replace them with their src attribute.
# 19:30 Loqi tantek meant to say: in addition to parsing out the nested plain text, if you encounter any <img> elements, replace them with their alt attribute text, or lacking an alt attribute, replace them with their src attribute.
# 19:30 tantek we don't need to do that for video/audio btw because they can already contain fallback text content
# 19:31 tantek if that seems reasonable, I'll add it to the microformats2 parsing spec
# 19:31 barnabywalters tantek: shouldn’t be too tricky — actually there are a couple of security points about e- vs p- content I need to bring up
shaners joined the channel
# 19:34 tantek let me know if that makes sense, and if so, hopefully we can get it added to php-mf2.
# 19:35 aaronpk the embed code for gist is <script src="https://gist.github.com/aaronpk/44fffbf6c61e4348e08f.js"></script> so I think you could put the URL inside the script tag for fallback and mf2 parsing
# 19:35 tantek scripts seem more dangerous to parse anything out of
# 19:36 tantek and in general I'd prefer to not depend on *anything* inside of a script for moving content around
# 19:38 barnabywalters the potential security issue I mentioned was something I noticed when implementing note reply contexts
# 19:38 aaronpk yea, that's why I would only ever copy over plaintext from other peoples' posts
# 19:38 tantek aaronpk - do you have documentation of the URL patterns you're searching for and how you're converting them to embed codes?
# 19:39 aaronpk tantek: I should probably add that to the post I just wrote
# 19:39 barnabywalters and I’m html-encoding the result of that run in order to preserve any <, > & characters in the plaintext
# 19:40 aaronpk I am, for soundcloud since that's the only way they support finding the embed code
tantek joined the channel
# 19:52 tantek oembed seems like extra work to support (as a publisher) when you can simply publish rich h-entry markup instead
# 19:53 tantek aaronpk, perhaps we need oembed.me that converts sites with oembed to a simple h-entry page for consuming
# 19:55 tantek aaronpk, re: your auto-embedding post, I'm looking forward to see what you do with this: "My next step is to parse arbitrary URLs looking for an h-entry, and display a mini preview of the content linked to."
# 19:55 tantek ideally implementing that should replace any need to support consuming oembed at all
# 20:02 tantek and if we build conventions for parsing / previewing / embedding like that
# 20:02 tantek we can point to those to get others to publish h-entry as well
# 20:03 tantek the embedding/previewing use-case, plus the commenting use-case, make a good case overall for doing it
# 20:03 tantek shouldn't be too difficult to overtake oembed in terms of deployment/support
xtof joined the channel
# 20:19 tantek aaronpk, you were asking what types of links that I auto-embed
# 20:20 tantek Falcon uses the #cassis.js function auto_link() to do all its auto embedding
# 20:21 tantek custom code for vimeo.com and youtube.com embeds
# 20:21 tantek oh and of course .jpeg, .jpg, .png, .gif image embeds
# 20:22 tantek @-name references get linked and marked up with a "h-x-username" class name on the hyperlink - thus giving the client the option to do more things with it (as I think Taproot does)
fmarier joined the channel
# 20:24 tantek.com edited /p3k (+0) "/* Features */ p3k embeds are in the footer, e.g. after the datetime published stamp, rather than inline with the content of the note" (
view diff )
xtof joined the channel
tilgovi and spinnerin joined the channel
# 21:53 aaronpk tantek: I'm going to be in SF again this weekend, would be fun to meet up again on Sunday if you're around!
# 21:53 tantek oh awesome! racking up those frequent flier miles :)
jedahan and xtof joined the channel
# 23:26 bret Also, how do I know if my PuSH atom feed works
# 23:27 tantek bret - you can try to subscribe to your homepage from a Status.net account
# 23:27 aaronpk bret: heh, PuSH is kind of hard to test. I've had the best luck by using a google talk bot that IMs you when it receives a PuSH mention
# 23:27 tantek if your PuSH feed works, you'll see updates appear in real time on your status.net account
# 23:27 bret i also just found an error on my atom feed
# 23:28 bret my www redirect is still broken from my switch to a TLD last night
# 23:28 tommorris bret: yeah, I have experience with FOAF. but can't help this evening as it's rapidly approaching 1am London time and I need sleep. ;)
# 23:28 aaronpk sounds like I should make atom.me to convert h-entry feed to an atom feed :)
# 23:29 tommorris bret: email tom@tommorris.org and I can help you with FOAF stuff asynchronously or poke me about it tomorrow sometime.
# 23:29 bret actually wait nm no error.. grrrrr why no worky
# 23:29 tommorris bret: or you can ask in #swig - that's the hang out of the upper-case Sem Web types
# 23:30 tommorris also, one observation: I always wanted to know what it was like to develop an AOL app. Over the last week, I managed to replicate what it must have been like by building a facebook app.
singpolyma and bnvk joined the channel