2013-11-16 UTC
paulcp, caseorganic, andreypopp, paulcp_, bnvk, ryana, thatryana and squeakytoy2 joined the channel
# 01:00 tantek speaking of maps, aaronpk, I believe you and I need to make an indieweb map - perhaps for 2013
# 01:01 tantek let's see what else we can get built by then to raise the bar ;)
paulcp, paulcp_, bnvk, squeakytoy2, scor, josephboyle, squeakytoy, pfenwick and tpinto joined the channel
# 02:09 tantek we should question dopamine-micro-attention-based features like "likes", "reposts", and similar.
snarfed, bnvk, squeakytoy, skinny, tpinto, paulcp, smus, wyomingplease, b0bg0d, shaners and scor joined the channel
# 04:17 bret !tell tantek +1 on the outlook on micro-attention features. They tend to have all sorts of details ripe for argument and dont offer much in terms of significant meaning
# 04:17 Loqi Ok, I'll tell him that when I see him next
bnvk, b0bg0d, squeakytoy and tantek joined the channel
# 04:57 Loqi tantek: bret left you a message 40 minutes ago: +1 on the outlook on micro-attention features. They tend to have all sorts of details ripe for argument and dont offer much in terms of significant meaning
# 04:59 tantek (e.g. where would this make senses on the wiki)
squeakytoy2 and tpinto joined the channel
# 05:11 bret prioritising focus importaint post types like notes and articles video pics comics
# 05:12 bret Not get too distracted on the 'right' way to repost/like things
# 05:15 KevinMarks_ focus on their utility for the person liking/reposting over showing that to the one whose post was liked?
# 05:15 bret It seems to me that those kinds of things tend to skew things in weird ways. People start seeing numbers rather than focusing on how to express themselves or figuring out how to really get to know other people better
# 05:15 KevinMarks_ right - the most toxic thing on twitter was exposing follower counts
# 05:16 bret You might get the rush by getting a like or a fav, but nobody is better for it
# 05:17 bret I'm not sure its a value judgement... Sure its encouraging, but it also sets the tone of a conversation
# 05:17 KevinMarks_ hm, I should have put the indieweb hashtag on that adactio quote, talking of retweets ;)
# 05:17 bret How can we encorage people to really think and communicate?
# 05:18 bret or how can we separate out topic streams from discussion :)
# 05:19 bret half of reddit posts are about karma now
# 05:20 bret Part of me wish a simple like/dislike was private info between the poster and judge...
# 05:21 bret but then the other part of me just wants to tell that person by writing them something short
# 05:22 bret anyway, mini scattered braindump. I dont have a very clear picture of what would be best
# 05:23 KevinMarks_ but as that piece tantek liked to points out, that can be problematic in the long run too
# 05:24 bret I get reposting, as a function of feeds and syndication and conversation or topic context... canned emotions like likes though rub me the wrong way
# 05:25 bret they seem like a nice tool if you wanted to sell 'like' data to an avertiser
smus and bnvk joined the channel
# 05:28 bret Maybe there is some kind of relationship between the magic number of participants in a community before it starts to suck as it gets bigger and micro attention features
smus joined the channel
squeakytoy and pfenwick joined the channel
bnvk and igalic joined the channel
# 05:47 bret just thinking too though that there is value in some kind of 'hey i dont know you and I dont want to distract you, but what you are doing is great' type currency
# 05:49 bret KevinMarks_: exactly, like Dunbar's number but for an online community, not sure if there is another name for that
# 05:54 bret ok i'm out for tonight...i'll try to revisit at some point soon
bnvk, squeakytoy, obensource, tpinto, brianloveswords, tantek, acegiak, skinny, pfenwick, cweiske, friedcell, andreypopp and bnvk_ joined the channel
# 12:48 XgF bret: The way I see it, a "like button" should be a way for me to say "This amused/interested me" without overloading the user's inbox with "++"/"like"/etc notifications
barnabywalters, sfarthing, skinny, pfenwick, bnvk_, bnvk, hidg-, Slyphoria, LauraJ, tpinto, josephboyle, tantek, snarfed, scor and squeakytoy joined the channel
# 18:24 snarfed hey tantek, falcon on tantek.com accepts webmention replies, right?
# 18:25 tantek snarfed - I have it setup through aaronpk's proxy webmention.io
# 18:25 tantek not at all realtime, at best, moderated once I figure out a way / method to display them
aaronpk|m joined the channel
# 18:26 tantek I feel it's in this kind of thoughtful dialog that there is hope for escaping from the empty dopamine-like-cycle
# 18:27 tantek I have such mixed feelings about likes/reposts
# 18:27 snarfed it's a big wide world w/a huge range of personalities, and human nature doesn't really change
# 18:27 snarfed but we can at least enable different patterns for people who want them
# 18:28 tantek on the one hand, clearly they're silo behavior/speech that it would be good to move to own on our own domains. on the other hand, is decentralizing and distributing the dopamine-like-cycle a good thing? neutral? or possibly *worse* in that we may end up building an even stronger more persistent empty dopamine-like-cycle.
# 18:30 snarfed yup, fair enough. fighting this culture is kinda orthogonal to indieweb
# 18:31 tantek as pioneers in the indieweb, everything we build sets a tone and a direction, individually and in aggregate
# 18:32 snarfed we're also only so many people, with so much bandwidth, so we have to pick our battles
# 18:32 bret Kicking it at unhosted PDX. Let me know if you want me to ask anyone questions on your behalf
# 18:32 tantek yes, the picking our battles constraint helps force us to be deliberate about our priorities
# 18:34 tantek bret - ask the first question - who works on unhosted and is self-dogfooding it own their own domain?
# 18:35 bret He (name? i came late) hosts his site by pointing his web server to the data store
# 18:35 tantek does anyone here, in #indiewebcamp, use unhosted on their own domain?
andreypopp joined the channel
# 18:36 bret unfortunately its all webfinger.... does not allow for just a host name, only email like user@host patterns :(
smus joined the channel
# 18:36 tantek bret - right, which is why it's been largely ignored by the indiewebcamp community.
# 18:37 bret just what mere mortals need, emails that dont actually function as email
# 18:37 tantek so if unhosted is dependent on email for identity, then it is broken
# 18:38 tantek webfinger is unnecessary for the indieweb, so it's not clear why unhosted would or should be dependent on it
# 18:38 bret XgF: maybe the speaker is mistaken, can you use a plain domain as a webfinger id?
# 18:38 XgF Webfinger works just as well for ftp://site.com/blah as it does for https://me.com or acct:me@example.com
# 18:39 bret XgF: per the spec? I wonder if he is refereing to the specific implementation
# 18:39 XgF tantek: Thats the acct: URI scheme, which is optional and orthogonal
# 18:39 tantek and a bunch of unnecessary invisible indirection -
# 18:39 tantek like you just stated - all the well-known crap
# 18:40 tantek yet another FOAF-like mechanism to hide profile information in an invisible file
# 18:40 bret It seems like it would make sense that webfinger is key value on any address
# 18:40 XgF WebFinger is good. Keeps all the discovery crap out of the HTML :-)
# 18:40 XgF Of course if what you're discovering is (in) the HTML its useless
# 18:40 XgF tantek: That still bloats the HTTP response
# 18:41 XgF "Citation needed"? How is it not true that putting data in a file makes it bigger?
# 18:41 tantek so instead, there is unnecessary complication and fragility introduced in the form of webfinger, .well-known, invisible files etc.
# 18:41 tantek sounds like the old school XML/RDF make it all invisible approach
# 18:41 tantek so history is just repeating itself with another generation
# 18:42 tantek who don't realize they are merely repeating the mistakes of the past
# 18:42 tantek XgF - citation needed as in prove it. prove that it makes *any* difference to network traffic to add a link header (e.g. packet sizes etc.)
# 18:42 tantek as opposed to the trivial provable problem of multiple additional HTTP requests etc. due to webfinger
# 18:43 tantek known inefficiency (more HTTP requests) is much worse than hypothetical problem (supposed "bloat" - with that impact? in HTTP headers)
# 18:43 XgF tantek: What if I don't care about the HTML?
# 18:43 tantek bret - what do you like about the unhosted model?
# 18:44 tantek XgF - have you added yourself to irc-people yet?
# 18:44 bret You can host a generic app for people to use, and they can use their own data store
# 18:45 XgF I very much care about the web. I very much lament how much everything is getting ossified behind HTTP
# 18:45 bret A static html + mvc app which can be bult for yoursle but used by all
# 18:45 tantek bret - nice theoretical model - but I don't buy it until I see the creators of unhosted selfdogfooding it on their own domains.
# 18:45 bret tantek: i think the speaker does selfdogfood
# 18:46 bret tantek: its more of the back end of how you actully work with data
# 18:46 bret rather than how one would host a site
# 18:47 bret but you are using locally running webapps with offline support which is a neat idea
# 18:47 XgF tantek: Anyway, let us not forget that the "Your ID is your URL" attempt failed because, well, for the 99% who don't own their own domains, its' cumbersome
# 18:47 tantek when you find out the name/URL of the developer who actually selfdogfoods unhosted
# 18:47 tantek and some way to verify it - or is it all self-claimed since it is purely backend?
# 18:48 tantek XgF - sorry that's a misconception, and one failure (OpenID) does not a proof make.
# 18:48 XgF tantek - (re "hiding data") - sorry that's a misconception, and one failure (XML/RDF) does not a proof make ;)
# 18:49 tantek every use of meta tags, XML/RDF sidefiles, etc. has failed
# 18:49 tantek with perhaps *one* exception (in the long term), Google support for meta description
# 18:50 tantek there's plenty of overwhelming data on invisible data failures
# 18:50 tantek XgF - yes, the silos themselves spam the OGP tags of FB
# 18:50 bret tantek: the speaker was Michiel de Jong i belive
# 18:50 tantek so you see summaries like "Twitter is the best site for posting short updates"
# 18:51 XgF EVERYTHING gets spammed. Deal with :-)
# 18:51 tantek visible data tends not to get spammed as much
# 18:52 tantek that's the evidence, and subsequently the design principle
# 18:52 XgF You can only distinguish "Visible data" if you parse the CSS, ignore stuff display: none
# 18:52 XgF ignore stuff color == background-color...
sfarthing joined the channel
# 18:53 XgF Really? Have you never seen the spammy shit which is all over HTML?
# 18:53 tantek XgF - e.g. and beyond spam, invisible data just breaks, rots, etc. through neglect
# 18:54 XgF tantek: Demonstrating that said data was actually unimportant
# 18:54 XgF If the data is important, people notice it. If people notice it, people fix it
# 18:54 tantek e.g. when we were parsing both blog RSS and the HTML of blogs, we found that 30-40% of feeds were broken, out of date, spammed, or otherwise completely wrong
# 18:55 tantek XgF - it's not just importance. greater visibility causes more feedback loops
snarfed joined the channel
# 18:55 bret tantek: rob lord was talking about using remotestorage/unhosted ideas for indieweb stuffs
# 18:56 tantek since it's not actually a primary content hosting project
andreypopp and bnvk joined the channel
sfarthing joined the channel
dietrich joined the channel
dietrich joined the channel
# 19:18 tantek like is he a founder of unhosted, a core developer, an occasional contributor?
sfarthing joined the channel
# 19:22 hadleybeeman Oh wow... That's going to keep my IRC alerts busy for the duration of this conversation… :)
# 19:24 hadleybeeman But given that you've got my attention, Tantek, I'll respond to your question. I wouldn't call it a silo because it doesn't act as a repository. Its data is transient.
# 19:24 hadleybeeman (But it's not an interoperable service.)
# 19:26 tantek hadleybeeman - that's my conclusion too. silo implies storage, containment. snapchat doesn't store anything.
# 19:26 tantek nor is there any social interaction other than replying
# 19:28 hadleybeeman "Transient bcc mms." I like that.
# 19:30 hadleybeeman I wondered, when I first saw that pie chart, what the data sources were. And what the parallel numbers are for iMessage and whatsapp.
bnvk, smus and KevinMarks joined the channel
# 19:52 KevinMarks just powerwashed my ChromeBook, as developer mode was making audio playback stutter
KevinMarks2 joined the channel
LauraJ joined the channel
# 20:06 bret Michiel works for 5apps the company that is doing html5 app hosting for firefox phone etc
dietrich and wyomingplease joined the channel
# 20:34 tantek you sure you discovered my webmention endpoint and sent one?
# 20:35 snarfed huh, odd. not sure. i use pfefferle's wordpress replies plugin, and it's sent replies to a number of other sites, but iirc not yet falcon
LauraJ joined the channel
# 20:36 snarfed i'll take a todo to fix that and send this one manually for now
# 20:36 tantek might need to check to see if the plugin's discovery code handles rel="webmention" links
tpinto joined the channel
# 20:40 XgF Meanwhile IETF still use the above (for HTTP, for example)
# 20:40 XgF Note that the last registration went into the IETF registry on the 1st of this month
# 20:41 tantek XgF - well if a rel value actually becomes relevant to implementations and real world content, we can add it to the official rel registry.
# 20:41 tantek if you have a specific one you'd like to propose, feel free to bring it up in #microformats.
# 20:42 XgF "Official rel registry"? As I see it, we have two registries pupporting to be official
# 20:43 tantek well for HTML there is one, as linked to from the HTML spec and the HTML5 spec.
# 20:43 tantek and since we're specifically talking about <link rel=webmention>, it's the only one that matters, officially, since that's HTML.
# 20:44 tantek anyway - that's not really indieweb specific so if you'd like to discuss it further, we can take it to #microformats
andreypopp joined the channel
# 20:45 XgF Actually, while on the subject I was thinking it would be nice to see Webmention submitted to the IETF as a potential RFC
# 20:46 tantek nah, it's not mature enough for RFC shenanigans
# 20:46 tantek essentially specs have to be fairly finished before dealing with IETF mailing list crap and being overwhelmed by talkers
# 20:46 XgF Is there much more it needs besides "Hey, this page is in reply to this page"?
# 20:47 tantek and I'd say until we have working automatic spam blocking figured out, webmention is not ready for an RFC
# 20:47 XgF The same way it works for Pingback/etc? (i.e. bayaesian filters e.g. akismet)
# 20:47 tantek it doesn't work for Pingback. Wordpress is about to drop support for it.
# 20:47 tantek I feel like I'm explaining a lot of history here.
# 20:48 XgF Ooh, is there a link to them removing it?
# 20:50 XgF You could always go for a hashcash type approach (e.g. you must submit hash=sha-256(reply_url || replying_to_url || nonce) where hash has X zero bits at the end
smus joined the channel
# 20:50 tantek XgF - you're welcome to try building that and seeing how it works
# 20:52 XgF Hashcash has the nice property of asymetric cost - the pinger needs to do 2^(n-1) attempts for n zero bits at the end
# 20:53 tantek XgF - perhaps you could write up a blog post showing how it would work step by step.
# 20:54 XgF tantek: Simple option: require that the hash of the post body evaluate to an SHA-256 where X bits at the end are zero. The sender needs to brute force this (On average 2^(n-1) attempts)
# 20:54 XgF the reciever can just hash and check the end
# 20:56 tantek xgf - sorry don't have time to evaluate your ideas on this right now - hence suggesting you write up a flow in a blog post. Others have written up their ideas on the wiki.
# 20:57 tantek and if it's simple - go ahead and build it. ;)
smus joined the channel
# 20:58 XgF tantek: Of course with the problem that my blog would then be returning 403 erros for any webmentions, of course :-)
# 20:59 tantek when aaronpk first implemented receiving webmentions, there wasn't anyone to send them to him. (I think he was the first to implement)
# 21:00 XgF Need to setup a PHP development environment locally so I can hack up idno's webmention endpoint/sender...
# 21:02 tantek snarfed - did you send the webmention manually from your post? I see it now in my queue.
# 21:07 Loqi Ok, I'll tell them that when I see them next
# 21:16 tantek !tell snarfed and it's back - guess you were rebooting it? ;)
# 21:16 Loqi Ok, I'll tell them that when I see them next
sfarthing joined the channel
sfarthing, friedcell, bnvk and snarfed joined the channel
# 22:05 Loqi snarfed: tantek left you a message 48 minutes ago: and it's back - guess you were rebooting it? ;)
# 22:06 snarfed !tell tantek yeah, i sent a manual webmention, and yeah, i nudged the site. shouldn't have been all the way down, but meh. thanks for checking!
# 22:06 Loqi Ok, I'll tell him that when I see him next
# 22:12 KevinMarks2 Proof of work seems an odd thing fit the Indie Web "make sure you can only be spammed by Google, Amazon the nsa and major botnets"
barnabywalters joined the channel
# 22:18 XgF KevinMarks2: Eh. Making the person sending you a message/spamming you do some work to do so seems sensible
tpinto, bret, josephboyle, dietrich, obensource, snarfed and bnvk joined the channel