2013-08-15 UTC
# 00:06 shaners hi friends. i've been AFK all day. any big highlights?
bnvk joined the channel
# 00:35 shaners any interest in an #indiewebcamp::hollywood at The Farmhouse? maybe piggybacking on the back of Farmhouse Conf 5?
spinnerin and bnvk joined the channel
spinnerin_, scor and shaners joined the channel
# 03:18 aaronpk I just realized that I should be treating Github as yet another silo, and should be POSSE'ing my comments there too :)
# 03:31 shaners aaronpk: eventually, you'll realize that some emails (say, to public mailing lists) are blog posts. and you'll POSSE from web post to email message
# 03:31 aaronpk i already did post one internal email to esri publicly though :)
# 03:34 shaners I do think that Posterous was on to something though with the idea of using email as the interface to posting
# 03:34 aaronpk i've used email as posting interface for a *long* time
# 03:34 shaners especially now that #hashtags are fully integrated into our lexicon (human and machine parsers)
# 03:35 shaners i def have "email as interface" on my list for eventual HS features
indiewebcamp-vis joined the channel
# 03:43 shaners "Today, we’re announcing a new feature for developers called WordPress.com Connect, a secure and easy way for millions of WordPress.com users to log in to your website or app."
# 03:45 aaronpk it's like OAuth 2 but with a request token from OAuth 1
# 03:46 shaners solo… now you can add wp.com to indieauth, right? :D
# 03:48 aaronpk i'm having trouble figuring out how this is different from their oauth 2 api
OpenWorld joined the channel
xtof joined the channel
# 04:30 shaners i used to post to wp.com via flickr from email a lot back then
tantek joined the channel
# 05:06 tantek email for posting makes a lot of sense since email clients are much more robust about retrying to send on mobile, whereas mobile apps / web apps just fail and give up.
# 05:07 shaners tantek: email has had a longer staying power than HTML.
# 05:07 shaners it means that email doesn't die. for all of its faults and failures, it keeps on.
# 05:07 tantek email is now too polluted to be useful, from spam, to bacn, to all the donotreply@ crap from providers you use, to random people emailing you, etc.
# 05:08 tantek shaners - nah, it's pretty dead already. check it less and less.
# 05:08 aaronpk interestingly, I still find it useful to search my email to find things from years ago
# 05:08 tantek and the kids, they really don't care. maybe they'll use it to reset account info once in a while.
# 05:09 tantek shaners - I do email pretty much only in the context of work these days.
# 05:09 aaronpk for example, looking at a bank statement for a transaction for $79.15 and I have no idea what it is because the merchant info came through as ACME CORP
# 05:09 tantek outside of work / work hours I blow off almost all email, and things are fine.
# 05:09 aaronpk then I search my email for "79.15" and find the receipt for the thing I bought
# 05:09 shaners i still read and write email everyday. work and non work.
# 05:09 tantek shaners - have replaced it pretty much with IM + IRC.
# 05:10 tantek I use email to talk with old people and old organizations only at this point.
# 05:11 shaners i do dream of an app to do "sms" / IM / email all in one place.
# 05:11 tantek anything that lets anyone spam you is a dead end
# 05:12 aaronpk tantek: if you've replaced email with IM+IRC, that's interesting because of the difference in timing between the two media. how can a near real-time medium replace an async medium? or are you saying that async media do not need to exist?
# 05:12 aaronpk shaners: I think you just described gmail. I can do all those from gmail.com now :(
# 05:15 shaners the good thing about all of this, is we can each do whatever we want
# 05:15 tantek but blog posts are both a) optional, and b) discoverable via relevant web search
# 05:16 aaronpk shaners: isn't that true for the whole world in general?
# 05:34 aaronpk my favorite is the follow-up comments from several years later
josephboyle and pfenwick joined the channel
# 05:41 aaronpk tantek: the other useful thing that came from email is MIME encoding. I like the fact that an entire HTML document including css and images can be stored in a single file.
# 05:46 aaronpk ironically (?) the only person to favorite that tweet is @obra
bnvk and pfenwick1 joined the channel
cweiske, andreypopp, donpdonp, pfefferle and eschnou joined the channel
mapkyca, andreypopp, melvster, eschnou, bnvk and pfefferle joined the channel
indiewebcamp-vis and andreypopp joined the channel
barnabywalters joined the channel
alanbur, bnvk and pfefferle joined the channel
emro and xerc joined the channel
# 11:03 Loqi RT @aral: Mark the dates: the 2nd IndieWebCampUK is taking place in Brighton on the 7th &
julien51, andreypopp and marcopolo joined the channel
musigny and barnabywalters joined the channel
josephboyle joined the channel
# 13:44 cweiske re webmention: how should the server behave if the request is missing source or target parameter?
scor and pdurbin joined the channel
andreypopp joined the channel
bnvk joined the channel
musigny joined the channel
# 14:37 barnabywalters has anyone seen any negative tweets or blog posts about the wired article? or is it just comments? I wonder if it’s added accountability or effort which weeds out trolls
# 14:38 tommorris barnabywalters: no, not seen anything negative other than on the article itself
# 14:39 barnabywalters it would be cool if the stuff we’re doing makes people more positive by design
# 14:41 neuro` barnabywalters: nothing in sight. Some interesting discussions in French with Jihaisse, xtof and some other people (on G+, sigh, even though I've PESOSed everything)
# 14:42 neuro` And at least 2-3 IndieWeb barcamps may happen in France: one in Paris, one in the French Alps, and one in the south.
danbri joined the channel
# 14:46 neuro` The only thing that bugs me there is that involved people are also those who were involved in th
# 14:46 neuro` e first barcamps and in the early Microformats days in 2006
# 14:49 aaronpk cweiske: good question. webmention.org says to follow HTTP conventions for malformed requests, etc
# 14:54 cweiske but it also defines error_code and error_message for json
# 14:55 aaronpk should probably clarify that in the webmention spec
josephboyle and xtof joined the channel
# 15:06 aaronpk there's a new 0.2 version of webmention that sandeep has been working on
# 15:10 aaronpk cweiske: oh also we're realizing that there is little reason to return a machine-readable error description, so we're leaning towards returning plaintext body responses
# 15:10 cweiske wtf? how can my webmention client then diff between a "please try again tomorrow" and a permanent error?
# 15:11 cweiske e.g. "webmention already registered" is something that means I don't have to try again
# 15:11 aaronpk there won't be an "already registered" response because we're allowing updates of mentions
# 15:12 cweiske source_not_found may be a temporary error because of the WM server's network connection
# 15:12 aaronpk also interesting is the fact that most implementations will be processing this asynchronously, so the only response they will be able to return is "202 Accepted"
# 15:12 aaronpk verification would happen in the background somewhere
x3rc joined the channel
# 15:14 cweiske I hate it when I have to jump hoops to get an error code
# 15:15 aaronpk so basically, 202 = all good. 400 = you did something wrong. 500 = server did something wrong, try again.
# 15:16 aaronpk best option in 400 case is to show the user the error response body and have them fix it
xtof, earplugs and ozten joined the channel
# 15:54 aaronpk icco: what's the command to create a new database for tumble? I haven't used padrino before, but the guides say something like `padrino ar:create` should work but it isn't for me
tilgovi and erikmaarten_ joined the channel
benwerd joined the channel
gjones joined the channel
spinnerin joined the channel
# 17:00 tantek (also a good example of how my site embeds images better than twitter does :P )
# 17:03 aaronpk we'll get a whole different class of comments here :)
shaners joined the channel
pius joined the channel
bndr and erlehmann joined the channel
# 17:18 erlehmann hello indiewebbers. i have made an art, a blog system in 150 lines of shell script that takes a directory of plain text or html files versioned with git and makes HTML + atom feed out of ti
# 17:19 erlehmann it is really very minimal static blog. what indieweb features can i add easily?
# 17:19 neuro` erlehmann: great. Maybe you can start adding some Microformats?
# 17:20 neuro` aaronpk went faster than me. I sometimes think you're helped by an army of robots
# 17:20 aaronpk erlehmann: are these all posts written by you? if so, you should also add your h-card to the home page
# 17:20 erlehmann can't i just link to a rel alternate of a single atom entry?
# 17:21 erlehmann aaronpk, they are, but it obeys the email address in the commit that added the file
# 17:21 aaronpk microformats-2 is a little bit different from microformats-1, and all of us are only parsing the new v2 version
# 17:21 aaronpk erlehmann: ah that's cool, you can generate a unique h-card for each entry in that case
# 17:21 erlehmann what are consumers for microformats-2 ?
# 17:21 erlehmann aaronpk, sounds like BLOAT. :D
# 17:22 erlehmann is there an easy way to get backlinks?
# 17:22 erlehmann most of my friends use twitter or identica – even programmers, but i found it too limited for my personal use.
# 17:22 aaronpk there is webmention.io which basically receives webmentions and pingbacks on your behalf
# 17:23 aaronpk if you make the entries readable by mf2 consumers you can start having converstations with the rest of the #indieweb folk
# 17:24 erlehmann aaronpk, how can i get a feed of such a thread?
# 17:24 erlehmann also, how have you indieweb people solved dead links?
# 17:24 Loqi [@AAinslie] Meet the Hackers (@WardCunningham, @KevinMarks &
# 17:25 erlehmann yeah, i saw that on hacker news. so i thought i should upgrade hipster news to indieweb standards.
# 17:25 erlehmann so can someone tell me why all of you are using pingbacks?
# 17:26 aaronpk well pingback is supported for backwards compatibility, but everything going forward is using webmention. it's like a simpler version of pingbacks
# 17:26 erlehmann why not just parse the referer header and check if someone has a link with in-reply-to ?
jontyw joined the channel
# 17:26 aaronpk i might go add that to my site. sort of an "automatic" discovery of inbound links as people click around the web
# 17:26 tantek erlehmann - because link headers are invisible?
# 17:26 erlehmann it seems to be the obvious, asynchronous way for me. no one spams referer headers AND includes proper semantic markup.
# 17:27 erlehmann tantek, invisible?
# 17:27 aaronpk you'll quickly discover a heavy focus on visible HTML here
# 17:27 tantek in contrast, replies tend have actual visible hyperlinks to what they're replying to
# 17:27 tantek and those are more trustworthy than anything in the header
# 17:28 erlehmann a[rel=in-reply-to]::before{content:'@' attr(in-reply-to) ':';}
# 17:28 aaronpk I like the idea of my server noticing the referer header and checking for inbound links, but that's more of a bonus, I wouldn't consider that part of a core indieweb protocol
# 17:28 tantek aaronpk - yeah, Hixie had that idea years ago when I was working at Technorati
# 17:28 erlehmann suddenly, visibility
# 17:29 aaronpk tantek: did anybody ever experiment with implemeting it?
# 17:29 tantek aaronpk - every site I've seen that did ends up loaded with referrer spam
# 17:29 aaronpk yea I'd be curious to collect it for a while to see what I get
# 17:30 erlehmann tantek, gibe examples. referer checking + checking for semantic markup interests me.
# 17:30 barnabywalters obviously you’d still fetch the page and verify that it does actually link, as well as parsing mf
micflan joined the channel
# 17:30 erlehmann so there can be no ”ž@search“ because twitter was too dump for proper namespacing?
# 17:31 erlehmann my spam filter – for years – is ”žwhat is the first name of franz beckenbauer“
# 17:31 tantek erlehmann - no one is smart enough for "proper namespacing" because there is no such thing.
# 17:31 erlehmann a friend uses z”žwrite apple in the following field“
# 17:31 erlehmann tantek, if you are having categories, you can make a category error
# 17:31 aaronpk erlehmann: I use client-side javascript to change a value in a form. works 100% surprisingly
# 17:32 jontyw Hi! Only just discovered that you're hosting an indiewebcamp after dConstruct, so I'll be attempting to come along.
# 17:32 erlehmann aaronpk, i despise you (this may sound like joking, but i do). why do you hate accessibility?
# 17:32 jontyw The maker faire is on the Saturday, but I think I can hand off any responsibilities to other people and come along.
# 17:32 tantek to have better guest list instructions like the other camps
# 17:32 erlehmann i actually once wrote a script to extract email addresses using a headless webkit, because the script would not run in my actual browser.
# 17:33 tantek I think that's partially what Aral is complaining about - no actual explanation for the sign-up
# 17:33 erlehmann like, the client-side script unobfuscating it.
# 17:33 aaronpk barnabywalters: erlehmann: accessible to browsers that don't interpret JS? or what?
# 17:33 jontyw Anyway, I'll lurk in here from now on.
# 17:33 tantek erlehmann - try adding h-entry + p-author h-card markup to your blog posts and post a sample here
# 17:33 erlehmann aaronpk, report to the principle-of-least-power re-education chambers now.
# 17:34 erlehmann aaronpk, i for one have an attention deficit disorder, so i browse without images and javascript so nothing flickers.
# 17:34 erlehmann i actually cannot use services like google+ or twitter that do not allow full recognition before processing because i am permanently distracted.
# 17:35 erlehmann like, every 5 seconds something pops up and i cannot disable it.
# 17:35 aaronpk luckily none of the indieweb commenting stuff requires forms or server-side validation, so all of this is a moot point :)
# 17:35 aaronpk as in, I don't have to host a form and process it for you to comment on my stuff
# 17:36 erlehmann aaronpk, *if* you create something, better make it accessible. otherwise you add to a power imbalance. have you checked your computational privilege today?
# 17:36 jontyw tantek: I'm on the overground zooming across london, so my connection is a bit flaky to attempt that at the moment. I'll get on the list tomorrow.
# 17:36 erlehmann there are sites that break on my smart phone because of heavy javascript use and race conditions and the devs – instead of fixing them – just say i should buy a new phone. however, they do not give me the monies.
# 17:37 aaronpk erlehmann: you should try adding some of the basic microformats2 markup to your feed. if you do that and post a link here we'll take a look at it.
# 17:37 erlehmann aaronpk, why add it to the feed?
# 17:38 erlehmann the feed already contains all the information.
# 17:38 erlehmann why not add it to the actual HTML site?
# 17:39 erlehmann oh, why? there are actual consumers, lots of them.
# 17:39 erlehmann like, my email client. or podcast clients (podcasting actually gave me a reason to hate on RSS for a lack of proper content modeling)
# 17:39 erlehmann and every blog provides them
# 17:39 erlehmann so why not re-use what is already there?
# 17:40 erlehmann i actually implemented a system that found metadata about podcast episodes using the feeds
# 17:41 erlehmann all while podcasters where debating which microformat to use (they are still debating, AFAIK)
# 17:43 erlehmann please stop using round brackets, parsing them is a mess
# 17:43 erlehmann (like, around urls)
# 17:44 erlehmann barnabywalters, why does it violate DRY? to me, it is just another instantiation of the resource. like having a picture in SVG, PNG or JPEG
# 17:44 erlehmann i generate atom feeds from html snippets and plain text files in git repository
# 17:45 erlehmann never writing them myself :)
# 17:45 aaronpk the problem is ATOM is another view to maintain, one which people often forget to keep up to date when they make changes to their HTML
# 17:45 erlehmann i do the same with my podcast buildsystem, have one audio file and make mp3, vorbis, opus out of it.
# 17:45 erlehmann does that violate DRY?
# 17:45 barnabywalters erlehmann: exactly — you shouldn’t have to write them, and other people shouldn’t have to read them
# 17:46 erlehmann barnabywalters, but how would feedreaders work in such a world then?
# 17:46 erlehmann i would actually prefer only publishing feeds
# 17:46 erlehmann but there is one stupid browser that cannot process feeds
# 17:46 erlehmann every other browser can.
# 17:46 erlehmann like, even IE
# 17:46 erlehmann and shows the feed as human readable. chrome does not even show the XML structure
# 17:47 aaronpk the nice thing about html is all the browsers know how to render it :)
# 17:47 erlehmann even elinks can show a feed with HTMl content.
# 17:47 erlehmann aaronpk, even chrome, the IE6 of the new web!
# 17:47 barnabywalters atom is only ever really going to be a container for HTML, so why even wrap it?
# 17:48 erlehmann oh, that is actually wrong.
# 17:48 erlehmann have you ever heard podcasts?
# 17:49 erlehmann barnabywalters, rather than what?
# 17:49 erlehmann i do use html for intra-document semantics and atom for container semantics.
andreypopp joined the channel
# 17:49 erlehmann as there is no format allowing me to have several html documents in a container other than rss and atom.
# 17:49 erlehmann (no plain text format)
# 17:50 erlehmann is there a commonly consumed MIME multipart stuff?
# 17:50 erlehmann like, similar to how many browsers can do html or atom?
# 17:51 aaronpk funny enough if you give someone a MIME-encoded file ending in .eml they can most likely open it on their computer
# 17:53 erlehmann i think i'll stick with atom feeds until there is something that is more often consumed. (which does not mean i will not add microformats, only that i am not buying the argument from novelty)
# 17:54 barnabywalters sandeepshetty built a h-entry => atom converter, and both him and eschnou are working on feed readers which consume mf
# 17:55 barnabywalters but by all means still publish atom feeds! a big part of the indieweb way is to gain independence whilst staying connected
# 17:55 erlehmann i often have low bandwith (7kb/s) so i am hesitant to include too much information in one way
# 17:55 erlehmann so what do you think about my reply links?
# 17:56 erlehmann it is funny, some people are actually angry when finding they can reply to me.
# 17:56 erlehmann and not just load off their drivel at my page.
# 17:56 erlehmann i then tell them to post on their own page that they hate me and link to me :D
# 17:56 erlehmann no one ever does
# 17:57 erlehmann so i only get helpful emails with corrections for posts
# 17:57 erlehmann yeah, it is entirely social
# 17:57 erlehmann i even thought to make it an email form, but that would be a deceiving interface.
jgw joined the channel
# 17:58 aaronpk does <form action="mailto:you@example.com"> work?
# 17:59 erlehmann i think it does. you have to give a proper encoding
# 17:59 erlehmann wait for it, i have a test case
npdoty joined the channel
# 18:08 erlehmann aaronpk, data:text/html;charset=utf-8;base64,PCFET0NUWVBFIEhUTUw%2BPHRpdGxlPmVtYWlsIGZvcm1zPC90aXRsZT48c3R5bGU%2BKnt3aWR0aDozM2VtfTwvc3R5bGU%2BPGgxPmVtYWlsIGZvcm0gKHRleHQvcGxhaW4pPC9oMT48Zm9ybSBhY3Rpb249Im1haWx0bzp0ZXN0QGV4YW1wbGUub3JnIiBlbmN0eXBlPSJ0ZXh0L3BsYWluIj48aW5wdXQgbmFtZT1zdWJqZWN0IHZhbHVlPXN1YmplY3Q%2BPHRleHRhcmVhIG5hbWU9Ym9keT5ib2R5PC90ZXh0YXJlYT48aW5wdXQgdHlwZT1zdWJtaXQ%2BPC9mb3JtPg%3D%3D
# 18:10 erlehmann does it work?
# 18:10 tantek wonders if data: URLs in IRC is an attack vector - e.g. what security context do they run in … hmm...
# 18:12 erlehmann attack vector?
# 18:12 aaronpk anyway yeah give the h-entry thing a try on your HTML
# 18:13 erlehmann aaronpk, this is a html document base64 encoded in a data uri. your browser shows the HTML in it.
# 18:13 erlehmann wait, i can give you a proper hosted web page
# 18:26 erlehmann something strange is happening
andreypopp joined the channel
# 18:30 erlehmann in the moz rendering engine
brixen, pavelz_, ceh and redge joined the channel
# 18:56 erlehmann how does it werk?
# 18:58 erlehmann pay attention to the third form. it renders wrong in gecko
# 19:00 erlehmann lawl, works in w3m
# 19:00 erlehmann but not in elinks (wants to invoke mutt)
# 19:04 bret Tantek: re: thing about zeldman.. I was just reading his twitter response to wired with an old context. It was a silly q.
# 19:04 bret Aaron.. the worst thing about GH is that the issues are a complete silo, although the API is pretty liberal at letting that data out… for now. Good call on the POSSE to GH issues, actually makes me want to think about POSSE myself, or move my issues inside my site structure somehow.
# 19:05 bret neuro`: Way cool on all those meet ups in france :)
# 19:05 bret erlehmann, I dig your site! looks really nice.
# 19:05 erlehmann bret, which one? hipster news?
# 19:05 erlehmann if you like it, look at the source code (linked at the bottom)
# 19:06 erlehmann it really is just 150 lines of shell script and a git repository (and redo, which is a simpler, more powerful alternative to make, but that can also be installed as a shell script)
# 19:07 erlehmann i almost always do my CSS in a simple, not obstrusive way.
# 19:07 erlehmann bret, i actually found out that bashblog (which people suggested) has like 700 lines. hehe ;)
# 19:08 erlehmann yeah, but assange beat me. before i knew it. his site is the original hipster news.
# 19:08 erlehmann > Readers have asked what software is used to run IQ.ORG. A mere page of handwritten ruby constructs the site out the most robust future proof storage form imaginable. A flat directory of text or html files.
# 19:09 erlehmann i did not know that
# 19:09 erlehmann but it sounds cool
# 19:09 erlehmann i like static blog compilers that do not do bullshit (bashblog does)
# 19:09 bret I started using jekyll a while ago, its okay
# 19:10 bret Fairly easy for someone like me who is new to web stuffs
# 19:10 erlehmann i found that git repository + shell is very easy
# 19:10 erlehmann after all, git is a distributed network file system
# 19:10 erlehmann and it even has namespacing
# 19:11 erlehmann bret, your site has a ”ž0“ prefixed to each post
# 19:11 erlehmann and one of the links is a pyramid Δ
# 19:11 erlehmann lol utf-8 ↬
# 19:11 erlehmann i'll send you a screenshot
# 19:12 bret the delta was leftover from an old octopress install
# 19:12 bret I changed the permalink symbol to that, still need to change that
# 19:12 erlehmann bret, you should seriously consider to change the black on black title font
# 19:13 erlehmann you probably have perfect eyesight, but the majority (i, for example) has not.
# 19:13 erlehmann i often mistake bad-contrast text for ornament
# 19:13 bret Im teribad at css, and I started using bootstraps theme
# 19:14 erlehmann 1. stop using bootstrap.
# 19:15 erlehmann bootstrap is almost as bad as ”žjust use jquery“
# 19:15 bret hey, I wrote my own webmention display in pure JS
# 19:15 erlehmann that is so stupid.
# 19:15 erlehmann why do you hate accessibility?
# 19:16 erlehmann shaners, ok!
# 19:16 erlehmann i'll be helpful.
# 19:16 bret I don't, I'm wicked new at this stuff and I'm just trying to learn
# 19:16 erlehmann bret, sorry.
# 19:16 bret its okay, I can handle brash words ;p
# 19:16 shaners bret: do you have a handy link to the @zeldman response to the wired article?
# 19:17 erlehmann bret, it is not okay, ”žbrash words“ are not helpful.
# 19:17 erlehmann bret, easy rule for high contrast. *always* set color AND background-color properties in the same scope in CSS.
# 19:19 erlehmann bret, if you read the text, i'll be happy to answer any questions regarding that.
# 19:20 erlehmann side note: i am always amazed how people who use open, crawlable resources manage to build something that is not easily re-usable.
# 19:21 shaners erlehmann: because it's often faster / easier to solve for *Just Me* for now and extract later.
andreypopp joined the channel
# 19:21 erlehmann shaners, faster, yes – easier, most often not.
# 19:21 erlehmann bret, webmention display in js is not easily crawlable.
# 19:22 shaners e.g., iamshane.com is a mess of duct tape, baling wire, cotton candy, marshmallows, tears, hardcoded urls, hopes and dreams. all very specific to me. but it allows me to quick and dirty test assumptions and ship something.
# 19:22 bret I want to get it into my data structure somehow, but that was the quickest path for me to get something done
# 19:22 erlehmann bret, that is what you are telling yourself. spoiler: there is a non-zero chance you'll change it, but you probably will not.
# 19:23 erlehmann (that is what always happens)
# 19:23 shaners erlehmann: the current state of all this #indieweb hacking is that everyone is working on parts that interest them and using the tools that they like / know
# 19:23 erlehmann i know a developer who quit her new job at the first day.
# 19:23 bret The problem is that I don't know of a good way to automatically write to my data structure and commit that to git
# 19:23 shaners in bret's case, he doesn't have a ton of programming experience. so he's doing what he can *now* to ship something.
# 19:24 shaners i have high confidence that he's learning more every day and getting better. and with it, his code / site will too.
# 19:24 erlehmann shaners, i am not belittling him.
# 19:24 bret Its primarily an educational experience, I come from a physics grad student background :p
# 19:24 shaners erlehmann: certainly not. def not saying you are.
# 19:24 erlehmann this “ship now” mentality is what resulted in cryptocat
# 19:24 erlehmann shaners, implying implications!
# 19:25 bret I could start using the Github API, but that isn't nessisary the rout i want to go down,
# 19:25 shaners the reality is that bret is still one of just a handful of people that are doing webmentions. and that's totally awesome.
# 19:25 shaners we're exercising all of these ideas in lots of languages/frameworks/environments/hosts. and that's a Good Thing™.
# 19:25 erlehmann shaners, “shipping” is not a value in itself. if you create software that for example discriminates racially, it might better to not ship *right now*
# 19:26 erlehmann if you are using turing complete languages where context-free or regular languages suffice, you are actively discriminating against computationally underprivileged
# 19:26 bret I like those guidelines, don't get me wrong.
# 19:27 erlehmann oh, no. but i am one of the underprivileged at times. i have an old laptop. and there are web sites that make it slow to a crawl.
# 19:27 erlehmann bret, that is actually a very good idea
# 19:27 erlehmann if you use old hardware as a developer, you will probably not make un-usable software
# 19:28 erlehmann if i were ever to command a squad of programmers, every single one of them would get an unterpowered netbook
# 19:28 erlehmann that is at least 3 years old
# 19:28 erlehmann bret, do you have a static host?
# 19:29 erlehmann bret, if you find a fault within these guidelines, tell me.
# 19:29 bret Before we go any further, let me reiterate that my tinny personal project represents my own interest in what makes up a much larger indiewebcamp project
# 19:29 erlehmann oh, of course.
# 19:29 erlehmann but whenever you are interfacing with others, you must be careful.
# 19:30 erlehmann example: a girlfriend of mine uses twitter as a blog software. that software seems to be defective, feeds stopped working.
# 19:30 erlehmann she now has effectively locked me out of her social life unless i go learn about twitter API and stuff.
# 19:31 erlehmann (the web interface is rate limited)
# 19:32 erlehmann she actually started using twitter because she felt locked out.
# 19:32 erlehmann and now, she is a victim-perpetrator, locking me out (she has a wordpress blog, but does not update it anymore with event stuff)
# 19:32 bret Right? She could feel like she could never match your technical chops and guidelines! ;)
# 19:33 erlehmann she could. that is the point.
# 19:33 erlehmann she can actually program enough.
# 19:33 erlehmann most people i know could make something like hipster news. but they chose not to, maybe of some authoritarian desire to submit.
# 19:34 bret erlehmann: I use githubs free hosting, I'm new to the sysadmin world, with my first job in it this summer. I have a VM running some IRC software, but still learning about nginx before I use it as my own host.
# 19:34 erlehmann i think the difference is between best case analysis and worst case analysis
# 19:34 erlehmann like, some people only think about what the best case is
# 19:34 erlehmann and some people think about what the worst case is
# 19:34 erlehmann this creates two measures of quality, everywhere
# 19:34 tantek erlehmann - speaking of shipping, have you had a chance to markup one of your posts with h-entry? e.g. the one you linked to above?
# 19:34 erlehmann like, if you only care about the best case, a lot of proprietary software is really better than the free equivalent.
# 19:35 tantek erlehmann - until it stops working on the newest hardware/os. but enough with the complaining - let's talk about your site.
# 19:35 erlehmann if you care about the worst case, you want to fail easy.
# 19:35 bret Ok, i have heard enough about the delta permalink… killing that now
# 19:35 tantek erlehmann - we care about making things work incrementally.
# 19:35 tantek thus, you've got posts on your domain - great!
# 19:36 erlehmann tantek, i have not yet. can you show me a list of consumers for h-entry?
eaburns joined the channel
# 19:36 tantek erlehmann - every indieweb site that supports indieweb comments
# 19:37 erlehmann like, a h-entry metadata extractor?
# 19:37 tantek erlehmann - it's data, not metadata (since it is visible, and the user cares about it)
# 19:37 tantek but if you're looking for a test parser - pin13.net is good
# 19:38 erlehmann why does it escape forward slashes (“/”)
# 19:39 erlehmann this seems to be an ugly format composed entirely of dictionaries and lists containing strings. any reason it is used?
# 19:39 bret is confused about the context of erlehmann's lawing
MattJ joined the channel
# 19:40 bret uF2 also maps nicely to the structure, but you could parse it to something else, right?
# 19:41 tantek erlehmann - the specific items array etc. were re-used from the canonical JSON output of parsing microdata - so that both microformats and md parsed to a similar JSON output.
# 19:41 tantek wonders if he should add these questions to an FAQ
# 19:42 erlehmann tantek, yeah, but i do not understand why. it seems both more verbose and less human-readable, then say, XML.
ozten joined the channel
# 19:42 aaronpk you could parse uf2 and output as xml if you really wanted, but that would mean dealing with XML
# 19:42 erlehmann seems kinda LISPy
# 19:42 erlehmann lots of brackets
# 19:42 erlehmann but maybe that is the destination of things
# 19:42 aaronpk everybody is passing around HTML to each other, it doesn't matter what each stores it as internally
# 19:42 erlehmann having lisp everywhere
# 19:43 erlehmann aaronpk, i know. but i wonder why people prefer this JSON
# 19:43 tantek how's the markup going on your post erlehmann?
# 19:43 aaronpk exactly, that's the reason we're all ultimately dealing with HTML
# 19:43 erlehmann i know it from javascript internally. maybe it is the modern ”ži'll just serialize this struct“
# 19:45 bret I mean, you could get in an argument with douglas crockford :p
# 19:45 erlehmann tantek, nice that you ask, but it motivates me more if you stop asking that every ten minutes and ask, like every hour or so.
# 19:45 erlehmann tantek, but i'll try.
# 19:45 tantek erlhemann - just trying to keep you on track - you seem to be easily distracted by tangential issues ;)
# 19:45 erlehmann took my ritalin
# 19:45 erlehmann tantek, attention deficit disorder.
# 19:46 tantek erlehmann - glad to hear you're taking care of yourself.
# 19:48 bret well, because I used the bootstrap icons XD
# 19:48 erlehmann the 0 should be an icon?
# 19:48 bret that should have a little chat bubble next to it
# 19:48 erlehmann signifying that there are 0 people in a chat?
# 19:48 bret it indicates the number of web mentions the article has
# 19:49 erlehmann you should maybe write that on the page, then! ;)
# 19:49 erlehmann also, contrast is really bad, light grey on white. i can see it, but it takes work.
# 19:49 erlehmann easily fixable!
# 19:50 erlehmann bret, do you have a site with your projects?
# 19:50 erlehmann i am interested!
# 19:50 erlehmann oh, found it
# 19:50 erlehmann > I have not had a chance to get these up yet.
# 19:50 bret and I was asked not to put my homework online
# 19:50 bret since the prof reuses the grad assignments
# 19:51 erlehmann plan: put it in cold storage, release when you are not in danger of incursing his or her wrath.
# 19:52 erlehmann you'll be the snowden of your university course!
# 19:52 bret yeah, i have it in a git repo across 3 computers
# 19:52 bret I'm just not as fast at latex as I thought I was, but the ones I typed up sure looked nice
# 19:53 erlehmann bret, i see. problem: it is still not descriptive. have you thought about having an alt text to the pic?
# 19:53 bret its some bootstrap markup… I should you are right
# 19:53 bret I want to switch over to a glyph font instead though
# 19:54 tantek bret - there's probably an emoji you could stick in as the alt text ;)
# 19:54 erlehmann it will take longer, but you'll always learn more if you forgo frameworks if you build something the first time
# 19:54 erlehmann a friend of mine learned x86 assembler as her first programming language.
# 19:54 erlehmann i thought she was crazy, but now she is doing C and python and can actually understand a stack trace.
# 19:55 bret erlehmann: totally, but when I started, I had different motivations. Funny who you can start something with the goal of avoiding learning about something, then realizing that you actually are way interested in it
# 19:55 erlehmann bret, wrong, there is the free symbola true type font. still, emoji is the wrong idea. use plain text as alt, like ”žcomments:“
# 19:55 erlehmann or just write ”žcomments: 0“
# 19:55 erlehmann and don't use fluff.
# 19:55 erlehmann yeah, CSS is funny
# 19:55 erlehmann do you know fefes blog? it is a german blog that is a CSS playground
# 19:56 erlehmann it takes CSS that you give it
# 19:56 Loqi I added a countdown scheduled for 8/15 1:56pm (#5235)
# 19:56 bret erlehmann: ill take some of this advice in mind going forward
# 19:57 erlehmann tantek, ritalin takes 30 to 45 min to work. until then, i'll be distracted and distracting (at the same time, superposition)
# 19:57 tantek erlehmann - another good channel for chatting about web standards, semantics, good web practices in general is #microformats - might have better luck there (e.g. with your article on good web practices)
# 19:57 erlehmann here, get your anaglyph glasses out!
# 19:58 erlehmann it is like a minimalist CSS zen garden :D
# 19:59 tantek hi erlehmann - I read through your web practices recommendations. let's discuss in #microformats
# 19:59 bret puts on my Dr. Lawrence Jacoby glasses on
# 20:02 pius sorry, I should de-link the "In Private Beta" badge
# 20:03 shaners erlehmann: wrt mf2 parsers outputting json: one of the big benefits is that lotsa programming languages can output json. so we use it as a way to confirm that all of our mf2 parsers are doing the same thing.
# 20:03 melvster i assume it's OK to store our microblog under another domain than your homepage and provide a link?
# 20:04 tantek melvster - multiple domains? sounds overly complicated ;)
# 20:04 bret It will only delay the h-entry implementation
# 20:04 tantek only you can decide how much complication you're OK with in your life :)
# 20:05 melvster tantek: you can follow a link from my homepage to find my mircoblog, much like you do rel="me" to find my twitter
# 20:05 tantek except Twitter is a silo. hopefully your microblog is not.
# 20:06 tantek and the biggest use of rel=me links here is now IndieAuth - does your microblog support OAuth2?
# 20:06 tantek melvster - open source isn't what makes something a silo or not.
# 20:06 melvster tantek: I may one day support OAuth2 ... but not a high priority right now
# 20:06 tantek both Facebook and Twitter share plenty of FLOSS
# 20:06 tantek and Wordpress.com is almost entirely FLOSS as well
# 20:08 erlehmann pius, white on white text
# 20:08 melvster actually timbl wrote most of the code (its MIT license) ... I've just been gluing things together and testing ... now I have mircoblogging, following and favourites
# 20:09 tantek melvster - do you have an example post permalink with h-entry we can take a look at?
# 20:10 melvster tantek: not this second as I only just got it working, but I plan to blog about it in the next 1-2 days so ... I can give you a link after that?
# 20:10 bret melvster: link to your site/microblog?
# 20:10 tantek melvster - it takes the average person mere minutes to add h-entry
# 20:11 tantek it'll take you orders of magnitude less time to add h-entry than to blog about it
# 20:13 melvster right now it just has content / date / creator type things
# 20:14 tantek markup just what you show on your HTML page - that's the microformats way
skh joined the channel
# 20:16 erlehmann melvster, why?
# 20:16 erlehmann if it is client side, that makes it hard to parse.
# 20:18 tantek melvster - yeah, erlehmann is right, if you can't curl it, it's not on the web.
# 20:19 erlehmann melvster, i actually have to run headless webkit because of stuff like that :/
# 20:19 melvster ok thanks will think it over ... but back to the original question ... it's ok to have a microblog *linked* to my homepage, right?
# 20:20 tantek if it's not serving its content in the HTML, is it really a microblog?
# 20:20 tantek blogs have permalinks to their posts, and their posts have their content in the HTML
# 20:20 erlehmann if a tree falls in the woods, but you have to have javascript enabled, does it make a sound?
# 20:21 melvster tantek: totally agree with you, it has permalinks, but the html view uses some JS to render the screen at the moment, as it forms an interactive client
# 20:23 melvster thing is you can unpack the data, server sider with node.js and display the HTML. OR you can unpack it client side with a script, and render the page.
# 20:24 tantek perhaps we *should* make don't depend on JS for your content (i.e. make it search engine indexable) as an explicit IndieMark Level 0 achievement.
# 20:25 pius melvster: I'm using javascript as well for lots of rendering. for me, it was a hard compromise between the unclean (losing CURLability) and the awesome of meeting my project goal of allowing people to deploy without a backend. I'm mitigating with WAI-ARIA and a caching service that crawls with headless webkit and provides cached version
# 20:25 melvster well my data is crawlable to the extent that it's public ... of course privately shared data wont get into google
# 20:25 pius tantek: aaronpk: I'll be unpopular and say that the javascript piece is an orthogonal issue.
# 20:26 tantek pius what is orthogonal to what in particular?
# 20:26 melvster so you can have google style circles or facebook styles groups
# 20:26 pius not depending on JS and stated goals of indieweb
# 20:26 melvster but the public posts will be crawlable through content negotiation
# 20:27 pius like, I understand crawlability as a criterion
# 20:27 pius if you use js for rendering, you have to work more to achieve it
# 20:27 pius but shouldn't fundamentally disqualify
# 20:28 melvster pius: a half decent search engine should have little trouble crawling the data
# 20:28 aaronpk pius: one of the building blocks is being able to have my server fetch content from your site via an HTTP request
# 20:29 aaronpk as long as an h-entry in HTML comes through when I make an HTTP request it'll be fine. you can load whatever with JS to build on that, but the markup needs to be in the response
# 20:29 pius aaronpk: yup, that makes sense. one thing I wonder, though, is if there should be a <link> standard that points to pre rendered versions of content.
# 20:30 pius so, if someone is using JS really extensively, they can still point to the rendered version, something like rel="rendered" or something
# 20:30 tantek pius - that's why it's wordly exactly as it is - in terms of capabilities
# 20:30 erlehmann pius, if you use js for rendering, your content is not accessible.
fmarier joined the channel
# 20:30 erlehmann what aaronpk says.
# 20:31 erlehmann pius, why not just give proper html and built your javascript skyscrapers from that foundation?
musigny joined the channel
# 20:32 erlehmann pius, there is a halting problem of networks. if you give 200 ok and the content is not there because of ajax trickery, you are essentially deceiving the user agent.
# 20:32 erlehmann so why not return the content and then progressively enhance it?
# 20:32 erlehmann or render the javascript part server-side.
# 20:33 pius erlehmann: one of the goals for the project I'm building is to make it easy for anyone who can ftp or S3 files to trivially deploy it
# 20:33 pius erlehmann: talk about "computational privilege," not everyone wants to run a backend
# 20:34 pius erlehmann: trust me, it'd be much easier for me to write this project in Ruby, Erlang, or whatever. :)
# 20:34 erlehmann pius, so what part of that needs javascript rendering?
# 20:35 erlehmann web pages are more often read than written. thus, if you do the work on the client, you are increasing the workload by several orders of magnitude *and* cannot be sure people will be able to read it.
# 20:36 pius tantek: I see that, but to me it looked like you were just proposing that heavy use of javascript is incompatible with the mission of the indieweb. that's what I was pushing back against.
# 20:36 erlehmann pius, heavy use of javascript for core content is incompatible with the web at large. that is because you are creating a single-page domain-specific web browser then and *must* lie regarding status codes.
# 20:37 erlehmann i run a link checker. if you use javascript for rendering, you cannot return a 404. therefore, all links seem valid.
kudu joined the channel
# 20:38 tantek pius - use as much JS as you want for enhancement. at least that you think your readers can tolerate on their devices.
# 20:38 erlehmann tantek has altered the pool. pray he does not alter it any further!
# 20:39 pius erlehmann: part of creating things is evaluating tradeoffs and choosing accordingly. incidentally, it's not the case that you can't have proper status codes with the approach I'm talking about.
# 20:40 erlehmann pius, fact: if you have the server side processing to decide that, you do not need client side javascript rendering.
# 20:41 erlehmann i think the question is one of object- and meta-language.
# 20:41 erlehmann but now off to include microformats in hipster nws
# 20:41 erlehmann hipster news
# 20:41 tantek erlehmann - iteration, evolution. ship early ship often.
# 20:42 erlehmann tantek, i did ship early. first version of hipster news was only 100 lines and took 2 evenings. but it had only a single site.
# 20:42 erlehmann and the feed timezone was off by two hoers
# 20:42 pius tantek: fair enough. that said, CURLability is not the same as searchability. understandable if both are values, but they're distinct and ought not be conflated. Meanwhile, I'll start using a link rel="rendered" and hope that my cowpath gets paved. :)
# 20:43 erlehmann pius, you are splitting hairs here. ”žrel=rendered“ is not a standard rel value.
# 20:43 erlehmann pius, why not make the thing behind ”žrel=rendered“ the main page?
# 20:45 erlehmann pius, elaborate?
# 20:46 shaners ok, everyone. here's your periodic reminder: part of the #indieweb ethos is a diversity of tactics is a Good Thing™. We all work on what interests us now. That's the #selfdogfood. We don't speculate. Instead, we ship, observer, iterate. Someone will *always* have a different strategy than you. That's ok. Especially, in the beginning. We interop at the simplest level: html and http. Past that, it's up to the builder. Let's not get
# 20:47 shaners For example, I don't like PHP. So, I'm not using it. But I don't care that aaronpk is. We only touch at the highest of levels.
# 20:47 erlehmann shaners, yo, ima let you finish, but javascript page rendering hates html and http interop. :3
# 20:48 shaners erlehmann: i agree. but pius will get there. for now, this is what he's doing. and that's ok.
# 20:48 neuro` shaners: funny that's part of the conversation we had sooner (in French) with xtof and a few people interested in Indie Web.
# 20:48 shaners it's easier to make something better that already exists, than to make something exist.
# 20:48 erlehmann shaners, i think that some things can not be built iteratively unless you think you can throw away your code.
# 20:48 shaners i.e., it's easier to shift from 1st to 2nd, than park to 1st.
# 20:49 erlehmann however, other people *will* use your code and not update.
# 20:49 neuro` About building, dogfooding, iterating, and stop losing time with sterile discussion on how we should translate POSSE and why working on WP is better etc...
# 20:49 erlehmann tantek, why should i include the permalink url if i have rel=bookmark?
# 20:49 tantek as long as you have the discussion by posting on your own site, I see no problem :)
# 20:49 pius shaners: there's no "getting there" … I'm well aware of the cons of javascript rendering, but also aware of the benefits. I didn't troll folks for implementing core functionality in browser extensions, so I think I ought to get some slack for using JS, if that's the price of not having to run a backend. :)
# 20:50 neuro` tantek: the discussion happened on a silo :-/ (I PESOSED at home though)
# 20:50 shaners pius: totes legit. i'm not giving any guff. you're shipping. that's rad!
# 20:50 tantek neuro`: "discussion happened on a silo" that was the mistake ;)
# 20:51 pius shaners: word and 1000% with you on diversity of implementation and just getting things shipped
# 20:51 erlehmann what is a POSSE and a PESOSED?
# 20:51 erlehmann pius, can you tell me what you are actually building?
# 20:51 neuro` erlehmann: POSSE : publish own site, syndicate elsewhere. PESOS: the opposite
# 20:51 erlehmann sharens, give URL?
# 20:51 tantek erlehmann, in general, you can answer "what is an XYZ" by trying indiewebcamp.com/XYZ
# 20:52 erlehmann oh, so PESOS is the dumb way and POSSE the smart way?
# 20:52 erlehmann thanks neuro with the accent
# 20:52 shaners pius: by "get there" i only meant the interop part. if JS rendering precludes indieweb-commenting, and you want to do indieweb-commenting, then you'll make whatever needed changes. but only when it's actually need. (just as a hypothetical example).
# 20:52 erlehmann i think sascha lobo (a well-known german blogger) tried PESOS
# 20:53 pius erlehmann: So, I want people who don't have the skills that we have but share the values that we do to be able to publish on the open web with an extremely gentle learning curve to actually hacking their own platform
# 20:53 neuro` erlehmann: I POSSE to twitter but had to PESOS all my history first :)
# 20:53 erlehmann pius, and why does this necessitate javascript client side rendering?
RedType joined the channel
# 20:54 pius shaners: I hear that, and I appreciate the context. I definitely want to implement indieweb-commenting and understand I'll be making it harder to implement with my approach. :)
# 20:54 shaners it's fine. you don't have to do it OR read his sites.
# 20:54 erlehmann > Note the reclaim wall, which shows you any content you left somewhere on the web.
# 20:55 neuro` Gentlemen, I'm leaving the grid, only way for me to code without being tempted to read what you have to say. Have a good night.
# 20:55 erlehmann pius, i do not see content there. only marketing drivel. where is the source?
# 20:55 erlehmann am i not seeing it?
# 20:55 erlehmann i am confused
# 20:56 Loqi Countdown set by tantek on 8/15/13 at 12:56pm
# 20:56 pius aaronpk: Can I just say that Loqi is awesome?
# 20:56 erlehmann Loqi, you are a dog. it does not even fit.
# 20:56 erlehmann tantek, question was ”žwhy use a h-entry permalink when i have rel=bookmark“
# 20:57 shaners you shouldn't rel-bookmark each article in a feed, only on the permalink page
# 20:57 tantek erlehmann - rel="bookmark" was the old way until we realized it was harder to maintain (across permalinks vs. aggregation pages like home and archive pages)
# 20:58 erlehmann tantek, but isn't rel=bookmark scoped to the nearest sectioning root? (let me look it up)
# 20:58 tantek so now rel="bookmark" is for backward compat only (you may skip it)
# 20:58 pius erlehmann: not sure why you're trolling me, but if you want to see some source code, visit me at http://github.com/pius. I release stuff when it's ready to accept contributions and be useful to someone else.
# 20:58 tantek erlehmann - no, rel values are scoped to the document in HTML5
# 20:58 shaners erlehmann: "nearest sectioning root" isn't well understood/implemented.
# 20:58 erlehmann thx pius. i am not trolling.
benwerd joined the channel
# 20:59 erlehmann i like the model of plain text / markup in git repositories
# 20:59 erlehmann i use it myself
# 21:00 erlehmann > The bookmark keyword gives a permalink for the nearest ancestor article element of the linking element in question, or of the section the linking element is most closely associated with, if there are no ancestor article elements.
# 21:00 erlehmann good enough for me
# 21:00 pius erlehmann: Sure you are. I don't think calling someone's work "drivel" is constructive. Also … I now have no idea what you're talking about.
# 21:00 erlehmann pius, oh, i was referring to ”žcoming soon“ and ”žprivate beta“ ;)
# 21:01 erlehmann i actually have no idea what this means.
# 21:01 erlehmann since i am not in the private beta
# 21:01 erlehmann sorry if i seem obnoxious, i do not try to antagonize.
# 21:01 erlehmann aaronpk, these are fish. they should keep fishing!
moystard joined the channel
# 21:02 pius wants the last 15 minutes of his life back :/
# 21:02 tantek erlehmann, rather than call something "drivel", consider suggesting improvements instead. :)
# 21:03 pius Also, erlehmann, I know you must ship a lot of stuff … I'd love to follow your work. Where can I do that?
# 21:05 erlehmann though i tend to not use github for new projects, their interface constantly becomes worse
# 21:05 erlehmann like, everything works without javascript (i disable it because it interferes with my shortcuts used for browsing) *except* forking.
# 21:05 erlehmann because someone used a <span> and javascript instead of <input type=submit>
# 21:05 erlehmann and they won't change it back
# 21:08 erlehmann pius, the questions that <http://ether.pw> does not answer: what does ether do exactly, how does it compare to other solutions (screenshots) and why should i care?
# 21:08 erlehmann i am not doubting you are developing something interesting
# 21:08 erlehmann i just have no idea what it is
# 21:08 erlehmann after reading the web site
moystard and barnabywalters joined the channel
# 21:09 pius erlehmann: why should you care? you shouldn't.
# 21:10 pius erlehmann: I know you're very distractible.
# 21:10 erlehmann pius, but then what is the web page to do?
# 21:10 erlehmann pius, let me show you what i mean. i have a (very minimal, sorry) web site for an o'reilly book i wrote with a friend http://internetmeme.de/
# 21:11 erlehmann people told me they have no idea what “internet memes” are. so now the first 3 sentences tell that it is about pop culture in the web and gives examples and when it came out.
# 21:11 erlehmann and people stopped asking me what internet memes are! :)
# 21:11 erlehmann (i'm the guy on the left with the hipster glasses)
# 21:11 erlehmann i am working on h-entry as i write
# 21:13 erlehmann pius, when will you release ether.pw?
# 21:13 erlehmann (i'll stop asking after that question.)
# 21:13 erlehmann shaners, done. :3
# 21:13 shaners ok. no one distract erlehmann until his h-entry markup is live on his site.
# 21:14 tantek erlehmann, ask not when others will release their projects, ask when you'll update yours. :)
# 21:14 erlehmann tantek “rel is always document scoped in HTML5” is wrong according to WHATWG spec.
# 21:14 tantek erlehmann - topic more appropriate for #microformats
# 21:14 erlehmann tantek, i just wanted closure to the conversation.
# 21:15 tantek erlehmann - right, it's about a microformats FAQ - hence please in -> #microformats :)
spindritf, musigny, josephg, state, rawtaz and benwerd_ joined the channel
# 21:34 josephg ... I'm glad there are other people out there working on this stuff
# 21:35 josephg is trying to slowly reignite the old opensourced google wave project
# 21:36 tantek shaners - now what was that you were saying about diversity of projects/approaches? ;)
# 21:36 josephg waveprotocol.org is the old site that barely ever gets updated
# 21:37 josephg ... but the interesting stuff happens on the mailing list
# 21:37 josephg its an official apache project, which means its hidden in their labyrintine SVN repositories
# 21:39 josephg = the mirror. It takes a bit of trouble to wake the giant
# 21:39 shaners personally, i couldn't do what do what you're doing. i think doing small for yourself is where it's at.
# 21:43 donpdonp josephg: interesting that wave is built on xmpp, which goog has given up on
# 21:43 josephg ... and most implementations are barely compatible
spindritf joined the channel
# 21:43 MattJ josephg, [citation needed]
# 21:44 josephg well, wave was a particularly insidious case anyway - it was built as an XMPP server extension
# 21:44 donpdonp josephg: im trying to remember the name of the 'philosophy' of federation that wave uses. do you know it?
# 21:44 state is here for the same reason as josephg
# 21:44 donpdonp document _something_, there's a wikipedia page on it but i cant find it
# 21:45 josephg (for which most servers have subtly different interpretations of the spec), and then it sent messages by encoding protobufs in CDATA in XML in XMPP TLS connections
josephboyle joined the channel
# 21:45 erlehmann wave, was that base64 encoded protobufs via xmpp?
# 21:45 MattJ Yes. More layers needed!
# 21:46 josephg anyway, regardless of the strengths of XMPP as a chat system, ... well, there be dragons in wave's federation protocol
# 21:46 tantek welcome state, what brought you here in particular?
# 21:46 josephg - which we're in the process of investigating fixes for!
# 21:47 state tantek: thanks! josephg: thanks!
# 21:47 state been looking at camlistore and thinking about security
# 21:47 state how to re imagine personal hosting for aws
# 21:48 erlehmann josephg, what problem does wave solve? etherpad with cat pictures?
# 21:48 erlehmann it looked to me like a modern mailing list crossed with an imageboard over xmpp.
josephboyle1 joined the channel
# 21:49 josephg state: camlistore is very cool - its been aroudn for so long now, I'm surprised it doesn't get used more
# 21:50 tantek josephg, are you using camlistore on josephg.com?
# 21:50 josephg erlehmann: wave should solve the 'I have some data that I want to collaboratively edit with shaners@example.com'
# 21:50 state josephg: yeah, i had been imagining building something somewhat similar
# 21:50 tantek josephg - I like ward's Federated Wiki approach to solving collaborative editing.
morrocco_mole joined the channel
# 21:51 shaners (erlehmann: stay focused. you're already in one conversation right now in the other channel.)
# 21:51 erlehmann josephg, just use git + sparkleshare or something.
# 21:51 erlehmann tantek, care to give me a link?
# 21:51 tantek josephg - you said "I'm surprised it doesn't get used more" - yet are you using it yourself on your own site? if not, then don't be surprised :)
# 21:52 josephg no - I'm using leveldb, because I wanted to use leveldb for something
# 21:52 tantek erlehmann, josephg, Federated Wiki? LMGTFY ;)
# 21:54 tantek welcome morrocco_mole - you're #70 in the room (for today)
# 21:56 morrocco_mole tantek: thanks
# 21:58 Loqi your site is in front of a lod balancer or a proxy
# 21:59 bret also, in terms of return codes, tantek did you spec out some html tags that, while under supported, should allow people not in charge of their server return codes to return codes inside the html?
# 21:59 bret would that work with more JS heavy sites?
# 22:00 erlehmann ahaha oh wow
# 22:00 erlehmann return codes inside of html
# 22:00 erlehmann meaning every http consumer cannot trust 200 ok
# 22:01 erlehmann bret, what if i send a HEAD request?
# 22:02 erlehmann i have a solution for http 410, if you can have .htaccess
# 22:02 tantek you can try, create one, e.g. reply to one benwerd's posts, and the delete it and provide the meta http-equiv
# 22:02 erlehmann RewriteCond %{REQUEST_FILENAME}
-f
# 22:02 erlehmann RewriteCond %{REQUEST_FILENAME}
!-s
# 22:02 erlehmann RewriteRule ^.*$ - [G,L]
# 22:02 bret erlehmann: I was just trying to brainstorm how to handle not being in charge of my return codes using free github pages hosting
# 22:02 erlehmann that means, empty files are 410 by default
# 22:03 tantek and now click on your nick in that list and a simple user page.
# 22:03 erlehmann bret, stop doing that. paying for your own web space is cheaper.
# 22:03 erlehmann bret, your brain power is limited.
# 22:03 tantek erlehmann, be nice - that sounded like a put down.
# 22:03 bret erlehmann: I do, but I was just experimenting with the space I was already in
# 22:03 erlehmann tantek, what is a ”žput down“ ?
# 22:03 pius bret: what codes did you need, just out of curiosity?
# 22:04 erlehmann tantek, bret, i just think that the problem domain is so narrow that under most circumstances it is just cognitively cheaper to provision a web server.
# 22:04 bret I kind of like the idea that the HTML is in charge rather than the server
# 22:05 erlehmann i hate that idea. if it would catch on, every http receiver would have to parse html.
# 22:05 erlehmann (sorry, not a native english speaker)
# 22:05 bret pius: 410 is the proposed return code for accessing an indieweb comment that has been deleted
# 22:06 erlehmann bret, indeed's.
# 22:06 josephg with all the NSA nonsense, I kind of want the web to be new and small again
# 22:06 erlehmann bret, tantek, i think ”žhow can i get my free github to do this and that“
# 22:07 erlehmann is a waste of resources, if for minimal power more, you can get a full solution
# 22:07 tantek 1 hour until erlehmann: how's the h-entry or hAtom markup going? have you tested it with http://pin13.net/ to see if you post information comes through in the JSON?
# 22:08 Loqi I added a countdown scheduled for 8/15 4:08pm (#5238)
# 22:08 erlehmann a friend of mine once basically rewrote some of the GNU coreutils for windows.
# 22:08 tantek josephg - indeed - when aaronpk and I decided on "indieweb" as a community following FSWS2010, be focused it on builders/creators
# 22:08 erlehmann tantek, i stopped doing it after the conversation with hixie. i have no reason to believe that the atom feed i am linking is inaccessible for people.
# 22:08 tantek people actively building/creating, on their own site, for their own site
# 22:08 bret erlehmann: i have nginx running on a VM, but so far I have not had the time to learn about it
# 22:09 tantek erlehmann, if you came here to interoperate with the indieweb, you need to implement h-entry of hAtom
# 22:09 Loqi tantek meant to say: erlehmann, if you came here to interoperate with the indieweb, you need to implement h-entry or hAtom
# 22:09 josephg tantek: yeah, I think thats the interesting segment of the community (in my totally biased opinion)
# 22:10 pius erlehmann: just as some consumers don't have the resources to run javascript, some creators don't have the resources to run their own server
# 22:11 josephg pius: isn't dreamhost or something only a few dollars per month?
# 22:11 erlehmann pius, then enable those to run a server. provide hosting.
# 22:13 tantek running your own server is more of a time/admin cost than $ cost
# 22:14 pius josephg: but to your point, any monetary barrier is still a barrier. per capita income in India, for example, is about $4/day. may not be relevant depending on your audience, but is a thing for a lot of folks.
# 22:14 bret I already pay for two different hosting. One shared, one a full VM on my own. GH-pages, with all its flaws, is all I could get running thus far. I want to get down to one system, but I just have not had time
# 22:15 josephg I think the situation with encryption is the same - usability is a bigger barrier to security than the actual security flaws in products imo
# 22:15 pius bret: that constraint breeds creativity … keep running with it. :)
# 22:16 tantek josephg - huge focus on UX/usability here in #indiewebcamp
# 22:16 tantek happens mostly as a side-effect of selfdogfooding
# 22:17 pius tantek: speaking of the perils of servers, I think your site is down
indiewebcamp-vis joined the channel
# 22:20 erlehmann bret, look at hipster news
# 22:27 bret I'm looking, I don't see anything different
# 22:28 tantek bret - have you tried implementing a deleted comment on bret.io?
# 22:28 bret heh, no not yet. I should though! Wouldn't be that much work
# 22:30 tantek josephg - all content federation uses h-entry
# 22:30 tantek RSS/Atom are only interesting if you have specific legacy use cases you're worried about
# 22:31 tantek josephg - that above getting started link included reasons why
# 22:31 tantek h-entry is the semantic building block of the indieweb.
# 22:32 tantek am copying the why to the h-entry page itself
# 22:34 josephg there's obviously a bootstrapping problem, but I'm curious what tools are written to use that
# 22:34 bret erlehmann left without even saying goodby! :p
warden_ joined the channel
# 22:47 josephg tantek: ok - I've put in h-entry, e-content and p-name
lukecyca joined the channel
# 22:48 tantek josephg - how does the result look from pin13.net?
# 22:50 tantek looks like you can add class="u-url u-uid" to your <a href= inside the headings
# 22:50 tantek e.g. <h1 class='p-name'><a class='u-url u-uid' href=
# 22:52 tantek e.g. <span class='p-author h-card'>Joseph Gentle</span>
# 22:52 tantek and <time class='dt-published' datetime='2013-07-30'> Tue Jul 30 2013</time>
# 22:53 tantek you can also leave off u-uid if there's only one u-url
# 22:53 tantek and only one place you serve your permalinks from
# 22:54 tantek it's a way of saying this is the canonical URL for this post
# 22:54 tantek syndication, indexing, referring in other posts, etc.
# 22:55 melvster tantek: sure ... that's the one I use for this one ... I have a personal blog too ... but this is for people interested in the read/write web and standards
# 22:56 tantek melvster - more interested in the blog/site you use to represent yourself on the web - hence "indie web"
# 22:57 melvster tantek: yeah, have one of those too on wordpress ... im trying to reignite my microblog presence tho
gravel joined the channel
# 22:58 josephg thats not quite right - pin13 doesn't resolve the url -> absolute
# 22:58 aaronpk there is a problem with the url resolver in pin13 right now
# 22:58 tantek aaronpk - but pasting html won't help with resolving relative URLs either
# 22:58 tantek since you need the original document URL to do that ;)
# 22:59 josephg hahaha I was like "uh, this isn't working either!"
# 23:02 josephg also speaking of 'paste the html', curl josephg.com | pbcopy
# 23:04 josephg so if I'm reading this right, h-entry is aiming to be the new atom/rss?
# 23:04 tantek josephg - more like the new <entry> / <item> :)
# 23:08 Loqi erlehmann: how's the h-entry or hAtom markup going? have you tested it with http://pin13.net/ to see if you post information comes through in the JSON?
# 23:08 Loqi Countdown set by tantek on 8/15/13 at 3:08pm
# 23:16 tantek benwerd - certainly better over dinner/drinks - speaking of
# 23:17 tantek anyone SF/bayarea-ish up for an informal indieweb meetup/dinner/drinks tonight?
# 23:17 benwerd_ alas, I can't do tonight - am free from Saturday night onwards
bnvk, indiewebcamp-vis and swinerg joined the channel
beredon joined the channel
# 23:22 pius busy tonight, but will try to drop in for the next one
# 23:22 josephg yeah I'd like to meet you guys & chat, but I'm busy tonight too
# 23:27 benwerd_ how would Sunday / Monday / next Friday work as possibilities?
# 23:28 aaronpk so close, I will be in Santa Monica next weekend but not quite SF!
state joined the channel
# 23:45 neuro` A question for people who implemented replies for Twitter messages: do you store the message information or do you fetch it everytime you generate the page?
# 23:46 aaronpk the key is the URL, and I store an HTML and parsed JSON version
# 23:46 tantek neuro` what do you mean by "message information" ?
# 23:46 neuro` aaronpk: thank you. That's where I was heading too.
# 23:48 neuro` tantek: sender (Twitter account + screen name + photo), message and date. This is everything I need to display the message I'm replying nicely.
# 23:50 tantek neuro` in that case, be sure to store the information independent of twitter
# 23:50 tantek i.e. rather than "Twitter account", store "author URL"
# 23:51 tantek this will make it easier when you move to supporting indieweb-to-indieweb reply contexts beyond just twitter-to-indieweb
# 23:51 neuro` tantek: that too. Think I'll store the whole json first and play with this.
# 23:51 neuro` My first concern was "what if people change their screen name / photo"
# 23:52 tantek neuro` indeed - you can always get their new screen name by following the tweet permalink and seeing if it redirects to a new screen name
# 23:53 neuro` Wil store everything first, just in case Twitter shuts down its API. Would be suicidal, but may happen.
# 23:55 neuro` Would I trust them, I'd probably not be here coding at 2AM (OK, insomnia didn't help either)
# 23:56 tantek pius, bret - FYI my site is up again. appears to have been down for about an hour due to server install software issues!
# 23:56 tantek (which fortunately all I had to do was file a trouble ticket about, rather than debug myself)