#hmansI get that IndieAuth won't even forward me to Twitter if it doesn't see the rel link on the Twitter profile
#hmansbut if I'm the legitimate user, that is an inconvenience at best
#hmansand if I'm an attacker in control of twitter.com/hmans, I can just set the bio to whatever.
#rasculif you gain control of someone's twitter account, that alone is not enough for indieauth
#mkoRight. You'd still need control over their domain.
#mkoGranted, you could IndieAuth as the person using a different domain, but it'd be a different domain.
#rasculbecause you need the rel=me stuff on the domain
#hmansWell, if someone "finds" my Twitter password, they can log into my indieAuth stuff, no matter how much IndieAuth checks that my Twitter account links back to hmans.io.
#hmansI'm just trying to grok this. I'm probably missing something important.
#rasculif the twitter/domain stuff is already linked, then the twitter account would give you the access
#snarfedhmans: yeah, you're discovering that indieweb is heavily biased toward a single "identity" domain per person. more than one will fail in lots of places, unfortunately
#hmansI'm very eager to do something with IndieAuth... on #pants, I don't ask users for email addresses, all they get is a password, so I'm trying to come up with creative ways to authenticate them without even that (or if they forget it).
#hmansSome have been bugging me about a "password reset" feature, but that's sort of hard to implement without an email address to send stuff to.
#hmansSo the direction I'm going in now is to ask them for links to their "other" profiles on the web and rel=me them for IndieAuth.
#hmansThe idea is that ideally, small communities will form where someone with a better understanding of tech hosts an instances for all their friends.
#mkoI like it. And that person wears the pants in the family. Or. Wait. What?
#hmansuser names are just domain/host names, so every user name and post ID contains all the information needed for communication, no matter where stuff is located.
#fr0zennice tool to generate visual identity: www.logoshi.com
#hmansI'm trying to, on a higher level, make it look and work like a "normal" social network, hiding the decentralized underpinnings for those who just want to "use" it.
#hmansKevinMarks, yeah, I'm using "following" in the code... but it's not a good user-facing term.
#KartikPrabhuhmans: why is "following" not considered a user-friendly term?
#mkoWant to see something awesome? Check out http://aprilzero.com -- beautiful quantified self site. Not sure how much of the data is liberated from silos yet. Just found the site.
#mkoOh my. Java. I haven't touched Java in almost 4 years.
#voxpellisnarfed: nice – mine currently don't pull any hcard data out, just assembles the relations
#voxpelli(had to install java now to get neo4j working locally :P)
#gRegor`Hmm, the checkmention test with in-reply-to and h-cite makes quite a different parsed array than I set my code up for. Trying to figure out how to best deal with that with php-mf2.
#voxpellimko: the challenges in crawling rel-me:s is btw sites like Flickr – which has rel-me:s between each and every page of a user
#KevinMarksto weed out the links on big sites that end up all being the same
#gRegor`I'm on PHP and using php-mf2. It's a fine parser. My question is about using various output
#mkoHuzzah. IndieAuth fully integrated into the prototype of my site including tiered authentication and auto-generating user profiles.
paulcp joined the channel
#mkoOh, I meant to see what output they were using.
#KevinMarksabsorbing it into your relspider would make a lot of sense (and it may need a bit of updating)
#voxpelliKevinMarks: Would be great to reuse that effort! I guess I should check with adactio about that pattern as well as he is doing the same on Huffduffer
#KevinMarksalso it's in JS because BradFitz was psychic and knew that was the future in 2007
#voxpelliKevinMarks: nice! I fully expetced it to be java or something
#voxpellimko: I'm btw only running a small demo instance of the relspider right now – so it will soon be full again :P The neo4j instance can only handle a 1000 nodes/url:s right now
#KevinMarksthe 80:20 tradeoff was that the big sites have well-known URL patterns, and so you cna short-circuit full crawls, whereas the little ones are more likely to have proper xfn and links
#KartikPrabhuI want to approach this from the following POV "Some one has landed on the archive page. I want the archive page designed in such a way that it encourages them to find articles/themes they want to read about"
#Loqigives KartikPrabhu the archive page designed in such a way that it encourages them to find articles
#KartikPrabhuKevinMarks: I hate those things as much as the "5 mins to finish reading" stuff
#KevinMarksright, but I'm imagining a webmention-driven distributed future
#KartikPrabhuKevinMarks: I have been thinking about those too. Incoming webmentions which are not replies or other could be included in "related posts" too
#bnvksnarfed: yah, in SF heading to Dolores park soon, then eating.
#bnvkwe could meet tomorrow before EFF hackathon if ya dig
#snarfedbnvk: i dig! anywhere between the mission and glen park works for me. let me know if you want me to pick a place
#Loqiaral: barnabywalters left you a message 5 days, 5 hours ago: looks like IWCUK is overfull — any idea if the venue is likely to be happy to accommodate one or two extra people?
#gRegor`Re: author spoofing on webmentions, I think I am going to fetch the h-card from the author's claimed URL and see if it matches the claimed author name. If it doesn't, match (or no h-card is found), it will be flagged as potentially a spoof. Does that sound good, or is there a simpler/more efficient way I'm not thinking of? Reference: http://gregorlove.com/2013/05/1154/#w80 which was sent...
#mayureshIs it theoretically possible to implement POSSE using JavaScript (client side)?
chrissaad joined the channel
#gRegor`Because it's the internet. I don't know everyone's reasons, but trolling, spamming, or maligning someone are all foreseeable reasons. It's not a problem currently, but it could be.
#gRegor`My wm implementation allows me to moderate them before being displayed, so adding this check isn't too difficult.
#gRegor`mayuresh: I'm not sure. I suppose you could have JS open the posting interface for the silos, at least. Not sure about pre-populating the content.
#gRegor`I think Twitter does/did allow you to pre-populate a tweet via URL parameters, and you had to click "tweet" to send it. They might have locked that down for spam abuse though.
#mayureshKartikPrabhu: yeah, that would be really awesome...
#tantek!tell KevinMarks, KartikPrabhu could you guys go through the logs from http://indiewebcamp.com/irc/2014-07-11 and upon reflection of what you said add to the wiki accordingly.
#KartikPrabhutantek: I mean to add some of that to the wiki... thanks for the reminder
#LoqiKartikPrabhu: tantek left you a message 1 minute ago: could you guys go through the logs from http://indiewebcamp.com/irc/2014-07-11 and upon reflection of what you said add to the wiki accordingly.
#tantek!tell mayuresh, gRegor`, kylewm instead of proprietary Twitter intents, please implement /webactions and only use Twitter intents as proprietary fallback.
#LoqigRegor`: tantek left you a message 1 minute ago: instead of proprietary Twitter intents, please implement /webactions and only use Twitter intents as proprietary fallback.
#mayureshtantek: yeah, Sandeep is a cool dude, have known him for a while now.
#Loqimayuresh: tantek left you a message 1 minute ago: instead of proprietary Twitter intents, please implement /webactions and only use Twitter intents as proprietary fallback.
#tantekbecause you can have theoretical discussions about formats and protocols and go in circles all day and never get anywhere - or worse, get distracted by someone's abstract notion of "complete" or "ideal" which has nothing to do with anything a user does or wants.
#GWGtantek, I think webactions have an issue because they require a plugin in the browser to work.
#tantekGWG - could be - you can figure one out using your own site!
#tantekGWG - but for now the point is, INSTEAD of *only* Twitter actions (intents), use webactions <action> tags, and then move those Twitter actions inside
#GWGtantek, maybe the JS route. I really need to pick up some javascript
#tantekGWG - if you're using <action> tags, add yourself to the IndieWeb Examples section on the /webactions page listing those like myself, barnaby, aaronpk etc. that are also publshing <action> tags.
#KartikPrabhubut it might be good to have a JS that fallbacks to Twitter stuff so that anyone can include it on their site for people without browser extensions
#GWGI want a picture of an in pool Indieweb meet up. That sounds atypical
#Loqikylewm: tantek left you a message 5 minutes ago: instead of proprietary Twitter intents, please implement /webactions and only use Twitter intents as proprietary fallback.
#Loqigives GWG a picture of an in pool Indieweb meet up
#LoqiKevinMarks: tantek left you a message 1 hour, 29 minutes ago: could you guys go through the logs from http://indiewebcamp.com/irc/2014-07-11 and upon reflection of what you said add to the wiki accordingly.
#gRegor`Though looking at the Facebook example mko posted, it wouldn't be too hard to extract just the <img> element.
#gRegor`Extracting comments on it would be harder though
#KevinMarkswell, if it is uniform per post type you probably can get something approximating the right thing
#KevinMarksgoogle's exports are better and do use microformats
petermolnar, grantmacken and tantek joined the channel
#Loqitantek: mko left you a message 1 hour, 3 minutes ago: Re Facebook export: It's now a crazy div soup with no discernible microformats, schema, or microdata.
#mkogRegor` KevinMarks GWG: The div soup is severe. I've identified the patterns that are needed to extract, creating proper RegExps for extraction will likely be a more tedious process.
#gRegor`I was thinking of using the DOM, not regex. Madness lies that way. :)
#mkoFor example, to extract e-content out of a wall post, you have to do the following: extract from wall.htm all content within the p immediately following the first </div> until div.meta
#mkoThe DOM might be the better way to go, but with it being a 10MB text file, I feel like the number of nodes might choke the browser, whereas consuming the file and using RegExp to extract the contents won't.
#mko!tell KevinMarks: Both Google and Twitter provide very nicely formatted exports.
#SamBI guess browsers don't provide callback-based parsing APIs ...
#mkoBut this is a serious mess. Most of it is unterminated text content run-on with block level content and a handful of tables thrown in for good measure.
#snarfedmko etc: i'm totally a broken record here, but remind me why you're not using the api?
#mkoThe content of the post is " has activity on FriendFeed." with a link and a description of "Learning Curve of Popular MMORPGS [Pic] (via Digg) - Comment or see your complete FriendFeed »"
chrissaad joined the channel
#gRegor`Says the post isn't available. Probably privacy setting.