#indiewebcamp 2014-05-17

2014-05-17 UTC
caseorganic and tantek joined the channel
#
aaronpk sometimes I wish I was better at CSS
brianloveswords joined the channel
#
snarfed i mostly wish i didn't need to use CSS
#
kbs heh
#
sparverius i use php
#
sparverius form your opinions now
#
tantek what is php?
#
Loqi PHP is a programming language and web server runtime environment used for many IndieWeb projects http://indiewebcamp.com/PHP
#
tantek :)
#
tantek snarfed - your CSS is quite decent!
#
snarfed tantek: thanks!
#
kbs my opinion is that tantek is linguidexterous(?) when it comes to php
tgbrun joined the channel
#
kbs ah, how interesting - polyglot seems to be the better term for it...
#
tantek.com edited /events/2014-05-21-homebrew-website-club (+14) "/* RSVP */ just show up" (view diff)
#
tantek kbs - or biglot if it's just two languages ;)
#
kbs :) biglot is a neat word - not to mention it is itself scannable in two ways (big-lot)
brianloveswords joined the channel
#
bret aaronpk: any advice on how to get a test token from https://tokens.oauth.net/token ?
#
bret oh wait nm
jedahan, emmak, brianloveswords, KartikPrabhu, snarfed, tantek, bshambaugh and JasonO joined the channel
#
@joeyalkes Charting country singer EILEEN CAREY @ the GRAND OLE’ ECHO SUNDAY MAY 18 @ 5pm https://www.facebook.com/events/730273137030970/730279923696958/?comment_id=730285013696449&notif_t=event_mall_comment #radio #indieweb #AtwaterVillage (twitter.com/_/status/467503018926026752)
#
tantek wat
#
KartikPrabhu haha talk about spam
tgbrun and tantek joined the channel
#
tgbrun quit()
tgbrun joined the channel
#
tgbrun quit
#
aaronpk bret: what do you mean "test token"?
snarfed and indie-visitor joined the channel
#
Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
#
KartikPrabhu hey NewYorker!
#
NewYorker Hi Kartik
#
KartikPrabhu new here? welcome!
#
NewYorker I just stumbled onto indiewebcamp while looking into web mentions.
#
KartikPrabhu that's how I ended up here as well :)
#
NewYorker I'm sort of in roll my own cmd mode.
#
NewYorker *cms
#
KartikPrabhu you mean webmentions/.
#
NewYorker Although I might go with Jekyll and just flesh it out with some custom code on top.
#
KartikPrabhu easiest to not build your own CMS in the beginning
#
NewYorker Yeah, I have a lot of ruby code for different components I want, but I did get it all glued together in working form.
#
NewYorker *never did
#
NewYorker That's why I'm thinking it might make more sense to focus on an existing system that can take plug ins
#
NewYorker Do you know of any that implement Purple Numbers?
#
KartikPrabhu A lot of people here build their own systems. But yeah using Jekyll and adding to it sounds like a good idea. I did that with Django
#
KartikPrabhu what is Purple Numbers?
#
NewYorker Oh wow, you are in for a treat then. They go back to the Mother of All Demos.
#
KartikPrabhu Is this the Doug Englebart's thing to link to paragraphs?
#
NewYorker Yup
#
NewYorker it comes from a paper call "Authorship Provisions in Augment"
#
aaronpk http://en.wikipedia.org/wiki/Purple_Numbers
#
KartikPrabhu oh you might be interested in http://indiewebcamp.com/fragmention which does something similar. You can link to arbitrary phrases!
JasonO joined the channel
#
JonathanNeal 👍
#
NewYorker oh cool
#
aaronpk http://eekim.com/software/purple/purple.html
#
NewYorker Thanks for the links Aaron!
#
aaronpk wow I had no idea...
#
KartikPrabhu cooked up by KevinMarks_ and scripts by JonathanNeal for fragmentions
#
aaronpk an attempt in 2001 to solve the exact same problem as fragmentions
#
NewYorker Here is Doug's paper: http://dougengelbart.org/pubs/seminars/sembinder1992nov/Q.pdf
#
aaronpk "HTML allows you to link to parts of a document if the target is defined (using the <a name=""> tag). However, if no target is defined, then it is impossible to address parts of an HTML document." http://eekim.com/software/purple/purple.html#hid2A
#
KartikPrabhu aaronpk: yeah interesting... lokos like something Medium does though. Assign a name="" to each paragraph
#
NewYorker The idea in Augment was to have structural anchors related to the document outline and also to have GUID's for each statement with the option of giving them nemonic keys as well
#
NewYorker The neat thing was that the Statement IDs would be preserved across edits restructuring the document
#
KartikPrabhu Newyorker: the neat thing about fragmentions is that you don't have to reply on the author having inserted the "name" and "id"
#
KartikPrabhu s/reply/rely
#
Loqi KartikPrabhu meant to say: Newyorker: the neat thing about fragmentions is that you don't have to rely on the author having inserted the "name" and "id"
JasonO joined the channel
#
snarfed ah, JonathanNeal's around!
#
snarfed JonathanNeal: wonder if you saw https://github.com/chapmanu/fragmentions/issues/9 . apologies for the nudge.
#
NewYorker Ah, that sounds like some of the work on Open Hypermedia System that can take external link servers to annotate closed documents
JasonO joined the channel
#
NewYorker The real challenge is getting round trip editing when you want to preserve Augment like Statement IDs
#
NewYorker in my code I used data- attributes and a rub goldberg assemblage of Ruby Gems
#
NewYorker *rube
#
KartikPrabhu Newyorker: that is one aspect of this thing that I don't understand. Even URLs are not preserved if the source changes so why do people expect pointers to parts of documents to be preserved
#
NewYorker There was an MIT Press book From Web to Workplace that had a lot of discussion of that
#
NewYorker It really is important if you are dealing with versioning for something like a legal doucment
#
NewYorker where you are interested in how one part might have changed between versions even if it was moved to a completely different part of the document
#
KartikPrabhu so is the idea to go above and beyond normal fragments like http://example.com/article#fragment ?
#
NewYorker Yeah, you want a fine grain to address elements down to the phrase level without having to manually create a gooleplex of anchor tags
#
NewYorker and you want to cover both the hierarchical dimension of the document layout as well as the semantic dimension of identifiable content chunks
#
NewYorker so you can either address the second subsection of section 3 or WHAT was at that location in the document at a given time
#
NewYorker The extra meta data provided by the dual ID models might also help in implementing Selective Undo
#
KartikPrabhu hmm what happens when I replace all the text at a URL?
#
aaronpk maybe the real question to answer is how to link to and store previous versions of documents
#
KartikPrabhu aaronpk: if you can reference some datetime (for versions) and fragmentions, then it should be all ok
#
NewYorker I think DARCS' Theory of Patches might help there
#
KartikPrabhu but you still reply on authors being consistent and all that
#
NewYorker Ever use DARCS?
#
JonathanNeal snarfed: I’m around.
#
JonathanNeal Sorry for the delay. We are hours or days from having our second baby.
#
NewYorker Dave's Advanced Revision Control System
#
JonathanNeal I thought, man better check on indie web.
#
snarfed JonathanNeal: 👍
#
bret aaronpk: just a valid token I can play around with
#
bret aaronpk: i grabbed the token that ownyourgram generataed
#
snarfed and congrats! we're expecting our first, due in about 16 weeks
#
JonathanNeal Don’t apologize for bumps either. I am really terrible at project maintenance and want to get better.
#
NewYorker It is a Haskell based package that treats versions as collections of discrete edit operations
#
JonathanNeal snarfed: congrats!
#
aaronpk bret: oh interesting
#
JonathanNeal I was unable to replicate #9. Want to give me more details? snarfed
#
snarfed JonathanNeal: sure, looking
#
NewYorker and the system can re-order the patches to consolidate them and selectively roll back orthogonal changes
#
aaronpk twitter just emailed me to tell me it's been 7 years since I created my twitter account
#
NewYorker Another idea I was kicking around was using an MD5 hash of an element's content as its anchor ID.
#
KartikPrabhu i don't know. All these sound much more complicated. Fragmentions sound so much simpler
#
NewYorker Then the link would only be valid if it pointed to the same text that you wanted to link to, so you couldn't have a scenario of giving someone a link to what you think is a positive book review and having them follow the link later on only to find an edited version of the review panning the book.
#
NewYorker I wound up with three different sets of purple number anchors for all of my document chunks to cover linking to this section, linking to this statement, and linking to this exact text
#
NewYorker The Fragmentations look really intriguing though.
#
NewYorker Has anyone thought about using word count offsets from the nearest defined anchor?
#
aaronpk yeah
#
aaronpk it's not on the wiki page tho
#
aaronpk http://indiewebcamp.com/irc/2014-04-22/line/1398233800
#
aaronpk http://www.kevinmarks.com/poemfragmentions.html##What+if+the+linked-to+text+occurs+more+than+once
#
NewYorker Thanks for those pointers!
#
NewYorker Could we use any exotic Unicode arrow glyphs in our urls?
Kopfstein joined the channel
#
aaronpk i believe one of the examples used stars ##★★★★☆
#
aaronpk https://github.com/chapmanu/fragmentions##In+another+example
#
aaronpk or better, https://github.com/chapmanu/fragmentions##★★★★☆
#
aaronpk (you'll need to install the browser extension for that one)
#
NewYorker I have got to get this stuff working on my site!
#
aaronpk fragmentions is easy! just drop in the javascript file!
#
aaronparecki.com edited /fragmention (+608) "add section about linking to the nth occurence" (view diff)
#
JonathanNeal Yea, fragmentions are fun ;)
#
JonathanNeal Also, contractions are 15 minutes apart.
#
aaronpk !
#
NewYorker Say, does anyone know of a tool to build an inverted file full text index of a static site that can be searched client side after you deploy to your server?
v0 joined the channel
#
KartikPrabhu Is anyone using the <picture> element stuff yet?
#
NewYorker never used it so far
#
KartikPrabhu just read about it on Smashing Magazine (http://www.smashingmagazine.com/2014/05/14/responsive-images-done-right-guide-picture-srcset/) and boy is it complicated!
#
NewYorker It looks it
#
KartikPrabhu if the smashing article looks complicated then using it will surely be more than that
#
aaronpk browsers should just implement content-aware scaling
#
aaronpk http://digital-photography-school.com/wp-content/uploads/2009/06/ps-before-after-tm.jpg
#
JonathanNeal KartikPrabhu: srcset is where it’s at.
#
aaronpk http://helpx.adobe.com/en/photoshop/using/content-aware-scaling.html
#
NewYorker My gut says to use PHP to detect a phone and swap images server side.
#
JonathanNeal It’s fun to realize that SVGs encapsulating images have been able to do this for years.
#
KartikPrabhu aaronpk: I looked at that way back when it was just a PhD thesis by the 'inventors'. it is pretty neat but doesn't work as well for large scalings
#
NewYorker But you only want to send as much image data as the client needs over the wire.
#
aaronpk KartikPrabhu: yeah, i've been following that for a long time. I forgot how old it is now actually.
#
KartikPrabhu atleast a 7 years old! I was in my third year in college!
#
NewYorker Wasn't there some sort of a MIP map encoding for scalable raster images?
#
KartikPrabhu people have been talking about just more "responsive" image formats that can be served appropriately by servers but don't know what happened there
#
KartikPrabhu in any case curious how this affect mf-parsing...
#
NewYorker Well time to call it a night, thanks for all of the research pointers everyone!
#
KartikPrabhu good night NewYorker! come back again :P
#
Loqi see you in the morning!
#
aaronpk so optimistic Loqi
#
Loqi who, me?
#
JonathanNeal These are the 4 “element query” techniques I’ve heard of. Am I missing others? https://gist.github.com/jonathantneal/d4e27ac9a65a426e6870
#
kartikprabhu.com edited /2014/Guest_List (+417) "/* East */ added myself" (view diff)
#
KartikPrabhu JonathanNeal: isn't there an element query polyfill?
#
JonathanNeal KartikPrabhu: yes, several, and I’m suggesting that these are the four ways they accomplish the element sizing listeners.
#
KartikPrabhu oh I see. I misinterpreted that part
#
KartikPrabhu JonathanNeal: what is your take on the picture-element?
snarfed joined the channel
#
JonathanNeal KartikPrabhu: it’s great.
#
JonathanNeal It’s really useful for art direction, responsive design, and even bandwidth efficiency.
#
KartikPrabhu I feel it is sort of annoying that it is tied to viewport size instead of something like element query
#
JonathanNeal I’m not sure it isn’t tied to element size.
#
KartikPrabhu from all that I've heard the "min-width" is the viewport
tantek and snarfed joined the channel
#
KartikPrabhu snarfed: any good guide to having submodules?
snarfed1 joined the channel
#
aaronpk !tell cweiske just for you, I added keyboard shortcuts to indieauth.com. You can now select a provider with j/k and press enter to select!
#
Loqi Ok, I'll tell them that when I see them next
#
aaronpk http://aaronparecki.com/notes/2014/05/16/2/indieauth
#
snarfed1 aaronpk++, thank you!
#
Loqi aaronpk has 463 karma
snarfed joined the channel
#
@kevinmarks "@ashedryden: How open company policies can harm marginalized users: http://t.co/6fazaBo7Lv" #indieweb (twitter.com/_/status/467558797334966272)
#
@nickreynoldsatw RT @kevinmarks: "@ashedryden: How open company policies can harm marginalized users: http://t.co/6fazaBo7Lv" #indieweb (twitter.com/_/status/467563089224040448)
tpinto and cweiske joined the channel
#
Loqi cweiske: aaronpk left you a message 48 minutes ago: just for you, I added keyboard shortcuts to indieauth.com. You can now select a provider with j/k and press enter to select!
#
cweiske !tell aaronpk thanks - now I have to press j and then return to select the single provider I have. at least I can use the keyboard exclusively now
#
Loqi Ok, I'll tell them that when I see them next
glennjones, kerozene_, petermolnar, chrissaad and jonnybarnes joined the channel
#
Loqi jonnybarnes: barnabywalters left you a message on 5/16 at 8:25am: RE comment looking weird — thanks, you found a bug with mf-cleaner! Fixed, regression test added and updating now.
tgbrun, petermolnar, tpinto and chrissaad joined the channel
#
@jordan_web This is a great idea, - Indie Box: Let's Bring Our Data Home - watch the second video. #indieweb #indiebox (twitter.com/_/status/467641910786805760)
petermolnar, KevinMarks, tgbrun and snarfed joined the channel
#
snarfed !tell bnvk hey, saw you tried bridgy publish! and i saw the facebook failure, looks like you hit https://github.com/snarfed/bridgy/issues/123 . sorry about that :/
#
Loqi Ok, I'll tell them that when I see them next
eschnou and carlo_au joined the channel
#
kylewm snarfed: do you still think it's a problem with bridgy's app key?
#
snarfed kylewm: i'd guess so, since it's the same behavior
#
snarfed silly :/
#
snarfed and i can't easily change the app key because all of the existing users' access tokens were generated with it
#
snarfed i guess i could use a different one just for fb publish, and require fb publish users to re-auth
#
kylewm hmm, well it does say "This app can post on your behalf"
#
snarfed the auth process? yeah, it's asking for the right oauth scopes
#
snarfed since the exact same code works with a different app id
#
snarfed if you're playing with this, https://developers.facebook.com/tools/debug/accesstoken is useful for debugging
#
snarfed (you can pull your access token out of the FacebookAuth datastore entity)
#
snarfed it also only happens for posts. i use bridgy publish all the time to posse fb comments, rsvps, and likes, and they all work fine
#
snarfed :/
#
kylewm rsvps work great for me too
#
kylewm oh, even you get the post error?
#
snarfed not sure, i rarely (never?) try to use it for posts
#
kylewm oh interesting, you know how you can only post for yourself when you're in developer mode or whatever
#
kylewm i'd assumed it was something like that
#
snarfed i wish!
#
snarfed i'll check again, but that at least usually gives a reasonable error code and msg
#
snarfed sucks that this is an opaque 500
#
snarfed kylewm: wow, i must not use fb often enough. i can't figure out how to add you as a friend
#
kylewm ha
#
snarfed hrug
#
snarfed er shrug
#
snarfed kylewm: added you to the bridgy facebook app
#
kylewm oh rad, another set of eyes might help (not optimistic tho)
#
snarfed oh man, maybe you're right about the developer mode thing. looking at https://developers.facebook.com/apps/256884317673197/review-status/ , publish_actions isn't on the approved list yet
#
snarfed asking for it now
#
kylewm i tried making a post through apigee with the bridgy access token and it worked!, wonder if that means it'll work from bridgy
#
snarfed i'm guessing no. i think the approval is handled manually, and i haven't even finished it yet. they want a lot of details, whee
#
kylewm oh, how strange i wonder why it worked from the api console thingy
#
kylewm it said published by Brid.gy and everything
#
snarfed maybe it works for you now that you're a developer
#
snarfed ah, maybe not, looks like it still doesn't work for you through bridgy itself
#
aaronpk gooooood morning
#
Loqi aaronpk: cweiske left you a message 7 hours, 33 minutes ago: thanks - now I have to press j and then return to select the single provider I have. at least I can use the keyboard exclusively now
#
aaronpk I suppose I could make "enter" just select the first one if none are selected. I didn't want to automatically highlight the first one cause that would look weird.
#
aaronpk mostly because the current implementation of "highlight" is just setting the background of the row to yellow, which is not the prettiest thing
jonnybarnes joined the channel
#
snarfed morning aaronpk
#
aaronpk !tell cweiske ok you're right, now you can just press "enter" and it'll sign in with the first
#
Loqi Ok, I'll tell them that when I see them next
#
snarfed hey aaronpk, have you ever gone through facebook's permissions review process for an app of yours?
#
aaronpk no, what's that?
#
snarfed about to start it now, but worried about the attention it'll bring
#
snarfed go to the fb dev site for an app of yours, then click Status & Review on the left side
#
aaronpk uh, crap they moved the link again
#
aaronpk wtf how do i get to the developer site from facebook.com
#
snarfed got me, just go to developers.facebook.com
#
aaronpk weird ok
#
snarfed or, this is what i'm talking about: https://snarfed.org/fb.png
#
kylewm it's in my left sidebar, under "Pages"
#
aaronpk kylewm: how'd you get it there? that sounds useful
#
kylewm dunno, I didn't change it from the default (I meant to say "Developer" is below "Pages")
#
aaronpk snarfed: crazy, never seen that before, is it new?
#
snarfed aaronpk: i think so. definitely new to me
#
snarfed about to apply for post perms for bridgy since they're not listed as approved, which i think might be causing https://github.com/snarfed/bridgy/issues/123
#
aaronpk wait you can't post by default now?
#
snarfed aaronpk: i think you still can, but fb always 500s when bridgy's app id tries
#
snarfed exact same bridgy code works w/different app id. silly.
#
aaronpk facebook is returning a 500 error? weird
#
snarfed aaronpk: eh, i get a low steady background hum of 500s from fb
#
snarfed but this one for publishing posts is 100%. odd because commenting, liking, and rsvping works. just posting always 500s
#
bret.io edited /auth-brainstorming (+1) "/* Using the Token Endpoint */" (view diff)
#
bret aaronpk: can you double check that edit?
tantek, brianloveswords, KevinMarks and snarfed joined the channel
#
aaronpk bret: no that should be GET
#
aaronpk if you're just verifying an access token
brianloveswords joined the channel
#
bret funny you should say that the exact moment i see my mistake
#
bret ill fix it
#
jonnybarnes font awesome has been updated for anyone that uses it
#
bret.io edited /login-brainstorming (-1) "/* Verifying the authorization code */ revert back to GET" (view diff)
#
aaronpk wait that's a different page
#
bret shit
#
Loqi lolz
emmak joined the channel
#
bret.io edited /login-brainstorming (+1) "/* Verifying the authorization code */ I'm dumb" (view diff)
#
aaronpk the good news is minor edits don't get sent out in the weekly summary
#
bret on the auth brainstorming page it says "he micropub endpoint makes a POST request" when the code block said GET..... is that a mistake?
#
bret.io edited /auth-brainstorming (-1) "/* Using the Token Endpoint */ Rolling back to post" (view diff)
#
rascul i was thinking about indieauth lately
#
bret the*
#
aaronpk bret: yes, the wording is wrong, i'll fix it
#
rascul i was thinking that it could be convenient to have a default provider that it automatically chooses to attempt login
#
bret ok
#
aaronparecki.com edited /auth-brainstorming (+9) "/* Using the Token Endpoint */ fix wording" (view diff)
#
rascul i was also thinking of linking some avatar service (ie libravatar) with indieauth, but then i thought it would be better for something like http://yourdomain.com/avatar.png
#
aaronpk rascul: i've thought about that, but doing that brings up a whole bunch of edge cases I'd have to handle
#
rascul aaronpk yeah i thought writing the code for it might get silly enough not to bother
#
aaronpk and I just didn't want to diagram all those out right now. especially now with keyboard shortcuts for signing in.
#
rascul i didn't think much on it, it was just a fleeting thought like 99.9% of my thoughts
#
bret yay everything works now
#
Loqi does a happy dance!
#
aaronpk yay
#
aaronpk rascul: someone was talking about setting up a service that would return well-formatted profile info given a domain, it would handle all the h-card parsing and such
#
rascul interesting, that
#
aaronpk like *vatar.com/aaronparecki.com.jpg would fetch my home page and look for an h-card and return the photo
paulcp joined the channel
#
aaronpk i don't really need more projects, otherwise I would totally do that
#
rascul it also might make sense just to do the simple thing and specify that avatar should be at /avatar.png or so
#
rascul skip all the parsing and complexity!
#
aaronpk yeah but there's also things like full name
#
aaronpk which is what h-card is for
#
rascul the h-card idea is a better one i think
#
rascul my thoughts that led me to that were thinking that if i were displaying someone else's comments, it would be nice to show avatar or something there with their name
#
aaronpk indeed, most of us do that already
#
aaronpk since the avatar often is already in the post
#
rascul now i'm thinking it might be better to grab the h-card like you were saying and put the whole h-card next to the comment or so
#
rascul i was just pondering this though, i'm not close to implementing that sorta thing yet
warden joined the channel
#
emmak when i first started using microformats, i assumed that commenters' avatars would be searched for in an h-card on their front page, but it doesn't appear to be the convention
#
emmak it adds complexity, but i think would give people more freedom with how they want to format their site
#
KartikPrabhu agreed. I don't have an avatar on my h-card on the homepage either
#
aaronpk emmak: you mean as opposed to including an avatar on each post page?
#
emmak aaronpk: exactly
#
KartikPrabhu aaronpk: yeah. Also other h-card details that people might choose to use in the future
#
aaronpk i think in practice most people have chosen to include an avatar on each post page, but it certainly isn't expected
#
aaronpk and because most people do, most people writing parsers only look there
#
aaronpk also it's a lot of extra work to fetch the home page and look there, also possibly fetching a /contact page linked off the home page for the people who don't want their face on their home page
#
KartikPrabhu Isn't http://microformats.org/wiki/representative-hcard-parsing supposed to be for those use cases?
#
aaronpk yeah
#
aaronpk its just a lot of work
#
emmak KartikPrabhu: i wasn't aware of that page, thanks for the link
#
aaronpk which is why i woulndt mind using that algorthm as a service
#
KartikPrabhu I made a python implementation that I use while parsing comments: https://github.com/kartikprabhu/indietools/blob/master/rep_h_card.py
#
emmak what if the algorithm was rolled into mf2 parsers?
#
aaronpk the problem is it requires making http requests
tpinto and caseorganic joined the channel
#
emmak i am thinking it might be useful to make a library that sits on top of the mf2 parser, that handles common indieweb tasks that require http requests
#
emmak fetching replies, and avatars
#
emmak you could also use it to make something like an rss reader
#
emmak has anyone made something like this?
#
aaronpk I think barnaby may have
#
aaronpk Check https://github.com/taproot
#
aaronpk And https://github.com/indieweb?tab=repositories
#
bret lol my micropub endpoint is already getteing requests for GET /wp/wp-login.php and /user/password
barnabywalters joined the channel
#
KartikPrabhu emmak: that is my aim with the "indietools" repo. to collect python implementations of common indieweb tasks that sit on mf2py parser
warden_ and caseorganic joined the channel
#
Loqi [@silverton] I stand by every tweet
#
Loqi all available from the firehose & my weblogs. AND i retain full editorial rights One more reason @indiewebcamp wins (http://twtr.io/mQfjzkf0dQ)
caseorganic joined the channel
#
snarfed KartikPrabhu: interesting! sounds similar to https://github.com/kylewm/mf2util …?
#
KartikPrabhu somewhat it seems. at the moment they do diff things
#
KartikPrabhu they might have similar goals though
#
aaronpk bret: lol yeah, so many bots
#
bret jerk bots
#
aaronpk i'm gonna try to write an upload-by-email bridge for the wiki since there's a couple photos on my phone I want to upload
#
barnabywalters aaronpk: what’s stopping you from uploading phone photos via the browser?
#
aaronpk so much effort
#
aaronpk and tiny buttons
#
aaronpk speaking of photos... I forgot about including the photos in the wiki backup
#
snarfed shh aaronpk, don't let tantek hear you say mobile web isn't as good as native apps!
#
snarfed :P
#
aaronpk heh, I never said that :P
#
aaronpk it's more an issue with mediawiki on mobile
#
snarfed heh. i'll say it if you don't.
#
snarfed that complaint is definitely the norm w/mobile web. mediawiki isn't the exception
#
aaronpk likely an artifact of the fact that web has been around a lot longer than mobile so there's lots of baggage
#
aaronpk it's entirely possible to make a reasonably good experience if you design for mobile first on the web
#
snarfed sure! still, that's unusual
caseorganic joined the channel
#
snarfed common case today is, when both are available, the native app is a way better experience
#
snarfed i don't necessarily like it, and i happily grant that history is the cause, but it doesn't really matter
#
snarfed anyway. no clue why i'm picking this fight. sorry. :P
#
aaronpk hehe
#
bret aaronpk: what would be the best way to get a token from the endpoint that is valid for example.com?
#
bret to test agaist valid tokens that are not authorized
#
bret on my endpoint
#
aaronpk hmmm
snarfed joined the channel
#
bret dont worry about it if it isnt trivial
#
aaronpk can't think of a quick way
#
bret what is the proper resposne code if its unauthorized?
#
aaronpk for the response from your micropub endpoint?
#
bret yeah
#
aaronpk should be at the bottom here http://ownyourgram.com/creating-a-micropub-endpoint##error+code
#
bret oh man, looking that the wiki let me go through that
#
kylewm snarfed: good luck on the facebook app review! assuming that is the problem, it's pretty dumb they return Internal Server Error"
#
snarfed kylewm: seriously!
#
snarfed thanks
#
kylewm aaronpk: have you and/or bret thought about what micropub would send if the source text is markdown? just content=raw markdown?
#
kylewm like if the composition UI was for writing markdown
#
bret kylewm: not at all. I just assumed the client would render it
#
bret errr actually wait maybe that wouldnt work
#
bret because people store markdown, so they would want that
#
kylewm I'm still a little confused about which side is the "client" :P
#
kylewm bret: maybe worthwhile to send both? so that people who store html could compose in markdown and publish in html?
#
aaronpk in order to be useful for writing full posts, an interface also needs a way to fetch the raw source of an article
#
aaronpk which may mean something like a GET request to the micropub endpoint with a post URL
#
aaronpk where the response would be the source of the post
#
kylewm aaronpk: that is for editing?
#
barnabywalters aaronpk: not sure what you mean by “writing full posts”
#
aaronpk more than plaintext posts
#
barnabywalters aaronpk: why does a client need to be able to fetch the source for an article for that?
#
aaronpk because the rendered HTML is likely not the thing the user wrote
#
barnabywalters aaronpk: still confused — what UI is this require to implement?
#
barnabywalters s/require/reduired
#
Loqi barnabywalters meant to say: aaronpk: still confused — what UI is this reduired to implement?
#
barnabywalters argh
#
barnabywalters a live preview UI?
#
aaronpk an editing ui
#
barnabywalters ah, okay
#
rascul http://hallojs.org/demo/markdown/
#
kylewm editing is kind of out of scope for what's been spec'd so far, right?
#
aaronpk yeah
#
rascul that's what i plan to implement for myself, or something similar
#
aaronpk but I want to get there eventually
#
kylewm rascul: ha, I am aware I'm writing the 3000th dillinger.io clone :)
#
rascul i like dillinger but i like hallo better
#
rascul dillinger has the live preview thing, but hallo has both wysiwyg editor and source editor in
#
kylewm aaronpk: do you have an intuition on whether you'd prefer to send the raw text as content vs. a separate field (e.g., content_raw)?
#
kylewm rascul: was there more of that sentence?
#
rascul there may originally have been more of that sentence
#
rascul your guess about what it was is as good as mine though
#
rascul i have a peanut sized brain with the mental capacity of a chipmunk :)
#
kylewm heh
warden__ and grantmacken joined the channel
#
@ablaze @dbreunig #indieweb (twitter.com/_/status/467742245530398720)
#
KartikPrabhu is indieweb becoming synonymous with progressive enhancement?
#
aaronpk dammit gmail
#
aaronpk maybe time to install mailpile
#
aaronpk this sucks. now I can't even forward emails to myself
#
aaronpk "Our system has detected an unusual rate of unsolicited mail originating from your IP address. To protect our users from spam, mail sent from your IP address has been temporarily rate limited."
#
barnabywalters aaronpk: that sort of thing happend to a friend of mine just as he was sending out a whole load of important business emails
#
barnabywalters google shut down access to his entire account for a few days
#
aaronpk wow
#
aaronpk the problem is that I run a mail server for all my domains, but then that mail server pretty much forwards everything to my gmail account
#
aaronpk so my domains get a bunch of spam which gets forwarded to gmail, making it look like my IP address is a giant spammer
#
KartikPrabhu aaronpk: does that also happen if you setup to send mail as that address too?
#
aaronpk that's actually what triggered this
#
aaronpk i just configured my phone to send mail from my own mail server, and gmail rejected the mail i just sent
#
KartikPrabhu no. i meant using Gmail's "send as" set to your domain email. Basically identifying it as you
#
aaronpk yeah I don't think it's getting as far as hitting my gmail account before gmail's smtp servers reject it
#
KartikPrabhu oh... yeah that is weird.
tgbrun and chrissaad joined the channel
#
Jeena I've been working on the homepage for my website last week and I think now I'm quite satisfied https://jeena.net/ Now that I don't only have the blog but also notes and photos I wanted to point that fact out to people who surf to the homepage.
#
barnabywalters Jeena: that’s looking great!
#
Jeena thanks, it is the third itteration, it isn't that easy as I thought ^^
snarfed, paulcp and tantek joined the channel
#
aaronpk ack
#
@vasilis Uh oh, the @indiewebcamp site is 502ing http://indiewebcamp.com/fragmention (twitter.com/_/status/467751107482370050)
#
aaronpk dammit
#
aaronpk it's up now, was only a few seconds!
#
Jeena hehe
#
aaronpk oh he deleted the tweet already
#
aaronpk i'll reply anyway :)
#
@aaronpk @vasilis sorry about that, was a misconfiguration on a system update to the indiewebcamp server! Was only down for... http://aaronparecki.com/notes/2014/05/17/1/ (twitter.com/_/status/467751672442138624)
ttepasse joined the channel
#
tantek aaronpk - did you cache his tweet before he deleted it? I noticed it didn't show up in your post's reply-context
#
aaronpk no, I guess I should have replied to the wiki line
kbs joined the channel
#
aaronpk k fixed it a little http://aaron.pk/r4W71
#
tantek whoa the permalink on the reply-context says /Array!
#
aaronpk yeah, i think that's a problem with the irc log markup
#
tantek <a href="Array"
#
tantek !!!
#
aaronpk this looks wrong http://pin13.net/mf2/?url=http%3A%2F%2Findiewebcamp.com%2Firc%2F2014-05-17%2Fline%2F1400355467
#
tantek which part?
#
aaronpk oh there's a bad p-url
#
aaronpk p-author h-card
#
aaronpk ah much better
paulcp joined the channel
#
aaronpk ok I also added # anchors for each line in case you want to grab a link to the logs in context of the day
#
tantek yay! I've kept hacking it by hand ;)
#
Loqi giggles
#
aaronpk me too :)
#
tantek aaronpk, how do you get the reply-context to update?
#
barnabywalters aaronpk: thanks! much appreciated :)
#
aaronpk i think i'm going to have to delete the cached file
#
tantek I tried sending a webmention :)
#
tantek (which is the update protocol for reply-contexts ;)
#
aaronpk i thought that was the update protocol for comments
#
tantek (yes we have an update protocol for reply-contexts :D )
#
tantek it is for both!
#
tantek bi-directional webmentions!
#
aaronpk how does that work for reply context?
#
aaronpk reading http://indiewebcamp.com/reply-context#Update
#
barnabywalters tantek: I only refresh reply contexts manually, because IMO it’s a different thing
#
aaronpk what! where did that come from
#
barnabywalters a reply context is something I’ve chosen to cache
#
barnabywalters a comment is unsolicited
#
tantek a writes a reply to b, a sends webmention to b, b includes a's reply as a comment. b later updates post, sends a webmention to all the comments, a receives webmention from b, notices that the the source is the reply-context of the target, and optionally updates the reply-context
#
tantek it's how the whole distributed system keeps everything notified of updates of all the various pieces that are cached!
#
tantek pure notification baby, zero polling :D
#
barnabywalters tantek: who has implemented that?
#
tantek barnabywalters: AFAIK no one has implemented reply-context CRUD. Hence why it's in the http://indiewebcamp.com/reply-context#Brainstorming section ;)
#
barnabywalters (see also http://indiewebcamp.com/reply-context##temporal+context)
#
tantek let's see if I can determine from the history what inspired me to write that up
#
barnabywalters fragmentions ftw
#
aaronpk interestingly this was an exact case where the temporal context was important, because he deleted the tweet
#
tantek aaronpk - deletes are documented too:
#
aaronpk still looks prety uncertain for deletes
#
tantek http://indiewebcamp.com/reply-context#Delete
#
tantek aaronpk - more like, open to further exploration, rather than "uncertain"
#
aaronpk sure
#
tantek hence: " This may be something we need to do individual UI experimentation on to get a feel for what are good options to consider. "
#
tantek there was no obvious "right" answer when I wrote it up (or still)
#
tantek however, now we have a real world example!
#
tantek of a reply-context to a deleted original post!
#
tantek fascinating he even replied to your reply to his deleted tweet!
#
aaronparecki.com edited /reply-context (+98) "/* Delete */ add example of a reply to a deleted post" (view diff)
#
aaronpk feedback on the new color requested: http://indiewebcamp.com/irc/2014-05-17#t1400355364
#
tantek what changed?
#
aaronpk tweets are bluer now, and the full line is blue not just the text
#
barnabywalters aaronpk: much better
#
tantek ooh
#
tantek so here's a dumb question, are there such search/streaming APIs for mentions of words/hashtags in public posts on G+ or FB?
#
tantek snarfed ^^^
#
barnabywalters aaronpk: maybe make the per-line hash links the same colour (or even fainter) than the timestamps
#
aaronpk oh thanks yeah
#
snarfed tantek: maybe!
#
snarfed not sure
#
snarfed if they exist, bridgy doesn't use them
#
snarfed tantek: https://developers.facebook.com/docs/graph-api/real-time-updates/v2.0
#
aaronpk can only subscribe to pages and users
#
tantek snarfed, I was wondering if Loqi would use them to bring in such posts the way he does for Twitter.
#
Loqi :)
#
tantek good Loqi, pat pat.
#
aaronpk i don't even think there's a search api for hashtags on facbeook
#
kbs maybe just google search with suitable time-range check? :) https://www.google.com/search?q=%22%23indiewebcamp%22+site%3Afacebook.com ?
#
aaronpk i can't believe that works kbs
#
kbs probably :) dunno
#
barnabywalters kbs: ha ha ha that’s hilarious
#
aaronpk I want a realtime technorati search for #indieweb
#
barnabywalters works for indiewebcamp too! https://www.google.com/search?q=%22%23indiewebcamp%22+site%3Afacebook.com#q=%22%23indiewebcamp%22+site:plus.google.com
#
tantek aaronpk, yeah :/
#
Loqi gives aaronpk a realtime technorati search for
#
barnabywalters aaronpk: you could easily add support for superfeedr’s keyword tracking
#
barnabywalters that’s probably the closest you’re going to get
#
aaronpk oh yea true
#
tantek kbs - except that search isn't time ordered
#
tantek first result is 11 months old
#
barnabywalters I have a IRC room where with notifix subscribed to “indieweb”, which sometimes surfaces some interesting stuff
#
kbs tantek: yea - not time-ordered. Probably closest [if there's no alternative] might be to hack something witha suitable time-range query
#
tantek check THIS out: https://www.google.com/search?as_sitesearch=facebook.com&tbs=sbd%3A1%2Ccdr%3A1%2Ccd_min%3A1%2F1%2F1970&q=indieweb+OR+indiewebcamp
#
aaronpk adding quotes filters out some junk
#
aaronpk https://www.google.com/search?as_sitesearch=facebook.com&tbs=sbd%3A1%2Ccdr%3A1%2Ccd_min%3A1%2F1%2F1970&q=%22indieweb%22+OR+%22indiewebcamp%22
#
tantek and -application.php
#
tantek https://www.google.com/search?as_sitesearch=facebook.com&tbs=sbd%3A1%2Ccdr%3A1%2Ccd_min%3A1%2F1%2F1970&q=%22indieweb%22+OR+%22indiewebcamp%22+-%22application.php%22
#
kbs cool :)
#
kbs aaronpk: nice addition with the search on the irc-logs page, very handy
#
tantek except now that's only *seven* results
#
aaronpk seven very accurate results :)
#
tantek but not often enough to be worth coding support for streaming
#
aaronpk barnabywalters: could you make notifix send stuff into here?
#
barnabywalters aaronpk: I’ll have a go
#
aaronpk I still haven't gone to the trouble of setting up a PuSH subscriber for loqi
#
aaronpk it looks like you can just subscribe to this feed http://superfeedr.com/track?include=indieweb
#
tantek hmm - I don't see any results
#
snarfed superfeedr++
#
Loqi superfeedr has 1 karma
notifix joined the channel
#
aaronpk yeah it's a weird fake feed that is just there to enable the PuSH subscription
#
barnabywalters +track indieweb
#
notifix You are now successfully tracking indieweb
#
barnabywalters okay, that should do it
#
aaronpk welcome notifix
#
tantek you know what's worse than Google's incomplete search results for Facebook? *Facebook's* incomplete search results for *Facebook*: https://www.facebook.com/hashtag/indieweb (3) or https://www.facebook.com/hashtag/indiewebcamp (0)
#
bret tantek: you have to pay to have everything show up
#
bret as the published ;)
#
bret publisher*
#
tantek search is hard, let's go social
#
aaronpk thoughts on whether I should auto-link things like /post in the IRC logs to wiki pages?
#
tantek seems useful and mostly harmless
tilgovi joined the channel
#
aaronpk yeah, likely to generate a few false positives but not the end of the world
#
tantek wonders if we should create stub pages for /nick /join /query
#
aaronpk ha
#
aaronpk and /quit
#
tantek false positive = excuse to create a stub
#
bret aaronpk++
#
Loqi aaronpk has 464 karma
#
tantek oh well not many plus.google.com results either
#
tantek so perhaps Twitter search is the only one that's statistically relevant now
#
tantek (most of the G+ or FB results were also available on Twitter too)
#
rascul if you connect to irc in some way so that the reverse dns points to your domain it'll show up when people /whois you
#
tantek that's a neat trick!
#
rascul /whois rascul
#
rascul for example
#
tantek first thing I tried ;)
#
rascul i figured :)
#
rascul i run znc on rascul.io and connect to it through my tinc vpn
#
rascul also the channel founder(s) could talk to freenode guys and maybe give out indiewebcamp cloaks
#
rascul then for /whois the people with the cloak will be nick@indiewebcamp/user
#
aaronpk i was reading about that, I don't think we qualify for some reason
#
bret rules--
#
Loqi rules has -1 karma
#
aaronpk alright, now all your blabber on IRC commands is linked to non-existent wiki pages http://indiewebcamp.com/irc/2014-05-17#t1400358846
#
bret :D
#
rascul hrm don't qualify? too bad that
tilgovi joined the channel
#
tantek barnabywalters: hey this "add .json onto the end" doesn't work anymore: http://waterpigs.co.uk/notes/1329/
#
tantek i.e. http://waterpigs.co.uk/notes/1329/.json
#
tantek or http://waterpigs.co.uk/notes/1329.json for that matter
#
barnabywalters tantek: that is true. I ditched content negotiation because it turned out to be a huge maintentance chore of very little actual value
#
aaronpk huge maintenance chore? how?
#
tantek *now* that is worth documenting somewhere
#
tantek rather, now *that*
#
barnabywalters aaronpk: because the .json view was essentially the same as the parsed microformat view, and I never used it so it kept breaking
#
barnabywalters typicaly invisible data problems
#
aaronpk huh
#
aaronpk my .json view is *literally* the parsed microformat view
#
aaronpk as in, I render the HTML as normal, then pass it through mf2-php before responding
BigBear joined the channel
#
aaronpk it's like getting content negotiation for free
#
barnabywalters aaronpk: content negotiation for one specific format
#
aaronpk i said "like" :)
#
barnabywalters I had a bunch of code which was devoted to being really flexible because I thought it was going to be useful
#
barnabywalters then it wasn’t
#
barnabywalters aaronpk: yeah, that’s a much more sustainable approach
#
tantek.com edited /reply-context-examples (+433) "various types of aaronpk's replies, subheads for easier discovery" (view diff)
bshambaugh joined the channel
#
tantek aaronpk, barnabywalters mine copy/pasting some of those experiences / support to: http://indiewebcamp.com/json ?
#
tantek s/mine/mind
#
Loqi tantek meant to say: aaronpk, barnabywalters mind copy/pasting some of those experiences / support to: http://indiewebcamp.com/json ?
#
tantek would be useful to document both what's working (aaronpk), and what was tried and abandoned and why (barnabywalters )
#
aaronpk k
#
tantek.com edited /IRC (+273) "Personal domain whois - brief documentation from rascul's notes in IRC" (view diff)
#
tantek rascul - any additional details you could add to http://indiewebcamp.com/IRC#Personal_domain_whois would be greatly appreciated!
bshambaugh joined the channel
#
aaronpk that /IRC page is getting really long... worth breaking out into other pages?
#
tantek aaronpk - probably worth some redesigning / rethinking
#
aaronpk also my attempt at indexing the logs so that calendar would render faster failed miserably yesterday
#
aaronpk so might be a good idea to move the calendar to a separate page
#
rascul ooo i'm famous!
#
tantek.com created /your_domain (+29) "r" (view diff)
#
tantek thanks rascul! appreciated in advance! :)
#
tantek aaronpk, any chance we can make http://indiewebcamp.com/irc (without trailing slash) redirect to http://indiewebcamp.com/IRC instead of http://indiewebcamp.com/irc/ ?
#
bret down to the last mile here. transforming an authorized micropub post req to a fie :)
#
bret file*
#
aaronpk bret: yay!
#
Loqi :D
#
aaronpk tantek: sure..let me see what's going on
#
aaronparecki.com created /irc (+16) "r" (view diff)
#
rascul the irc page is indeed pretty long
#
aaronpk i think the web server is grabbing the redirect since there's a folder named /irc on the server, i'll have to override it
#
aaronpk ah I see yeah...
#
tantek I would start moving the lengthy IRC clients stuff to a different subpage
#
tantek and ONLY keep the clients that folks in the community actually actively use
#
bret this highlight on rollover in the IRC logs is very nice :)
#
tantek i.e. clients people use = clients you can get help easily to setup / troubleshoot
#
tantek a-ha, re: reply-context CRUD: http://indiewebcamp.com/irc/2013-08-13#t1376448809
#
aaronpk ok this might break everything. brace yourselves.
#
aaronpk i got rid of the automatic redirect of />{folder} to /{folder}/ when {folder} exists
#
barnabywalters awesome, indieweb activity from the 60s! http://indiewebcamp.com/irc/1969-12-31
#
tantek I'm ok with that
#
tantek hahaha barnabywalters
#
barnabywalters must have been a challenge what with the web not existing
#
bret groovy man
#
barnabywalters that would explain the lack of very much activity
#
aaronpk and now /irc is a mediawiki redirect to /IRC
#
tantek yay for small victories!
#
Loqi :D
snarfed joined the channel
#
aaronpk i also added stupid caching of the calendar so at least it doesn't regenerate on every requets now
#
aaronpk just every 5 minutes, so if nobody views the page in 5 minutes it'll go regenerate
#
aaronpk actually i can probably set that to a day
#
bret tantek, are you backing up a copy of the wiki?
#
tantek bret no, though I'm interested in figuring out how to do that if it means *caching* a local copy of the wiki I can read offline :D
#
aaronpk there's info here! http://indiewebcamp.com/wiki/backup
#
bret well, you can sort of do that with the text version
#
aaronpk yeah if you don't mind reading mediawiki source
#
bret its an interesting view
#
bret mostly for discovery, i usually go look at the actual wiki if I find something
#
tantek aaronpk - what would you think of moving the "full" IRC logs calendar to its own page, and keeping just a row of the current/past 2months on the main IRC page?
#
aaronpk oh good idea
#
tantek how hard would it be to generate a second IRC mini calendar like that?
#
aaronpk not sure that'll make a difference with the speed, but would be good for the UI anyway
#
aaronpk let me check the plugin
#
aaronpk i'll have to add a parameter, but not a big deal
#
tantek that way we can just link from that section to a {{main|irc/logs}} page or something
#
tantek and a single row of 3 months will look cleaner too
#
tantek or 6 months would be fine too
#
tantek a 3x2 grid has a certain asthetic appeal
notifix joined the channel
#
tantek a/asthetic/aesthetic
#
aaronparecki.com created /IRC/logs (+325) "new page for full calendar view of IRC logs" (view diff)
#
tantek fascinating, the last time we had a whole inactive day in IRC was 2013-01-26!
#
aaronparecki.com edited /IRC (+38) "only show past 6 months in the calendar view" (view diff)
#
aaronpk please excuse me while I hack up this mediawiki extension
#
bret github is promoting https://guides.github.com/activities/citable-code/ Digital object identifers
#
tantek haahaha that's funny. I wonder who talked them into it. DOIs are kind of a joke.
#
tantek Because you, know, URLs.
#
aaronpk question: is the subtle color shading useful? http://indiewebcamp.com/IRC#Logs
#
tantek and this kind of nonsense: "Before Zenodo can issue a DOI for your …" - say what? I have to get some ORG to issue URLs for me? no thanks.
#
tantek I suggest scroll to the bottom, click (*) Not helpful at all, Submit
#
tantek hint: anyone that talks about "Minting a …" is probably doing it wrong and trying to sell you some service for doing the so-called "minting"
#
tantek ok I'm trimming IRC clients from the short list
#
tantek anything that no one speaks up about will be bumped down to the "full" list down at the end (which may go to a new page)
#
tantek anyone use Limechat?
#
aaronpk yes
#
aaronpk mostly ios, sometimes osx
#
tantek is it free?
#
aaronpk not on ios
#
rascul.io edited /IRC (+440) "/* Personal domain whois */ add more scenarios and links" (view diff)
#
aaronpk but it's open source
#
tantek thanks aaronpk
#
tantek next, anyone use XChat?
#
rascul if so they should switch to hexchat
#
tantek on Linux?
#
rascul or windows
#
tantek rascul, do you use hexchat?
#
rascul i don't anymore
#
rascul now i use konversation
paulcp and jonnybarnes joined the channel
#
tantek hmm - then why hexchat instead of konversation?
#
rascul hexchat instead of xchat
#
rascul hexchat forked from xchat awhile ago, now xchat is dead and hexchat is actively maintained
#
bret hexchat is what I as a linux GUI client
#
rascul http://hexchat.github.io/news/announcement.html
#
tantek good to know
#
tantek anyone use CIRC?
snarfed joined the channel
#
rascul real men use shic https://github.com/halhen/shic
#
pdurbin I'm not sure I'd call DOIs a joke. A racket, maybe. :)
#
aaronpk i'm glad you're cleaning these up :)
#
tantek pdurbin lol
#
aaronparecki.com edited /IRC (+1) "/* Logs */ fix month param" (view diff)
#
rascul there's a log from 1969!
#
rascul http://indiewebcamp.com/irc/1969-12-31
#
aaronpk ugh i should just delete those
#
tantek yes
#
aaronpk gone
#
aaronpk ok now the next question is how to get crawlers to stop crawling ALL OF THE DATES since 1969
#
aaronpk you should see the access log on this thing
#
tantek wait what
#
Loqi yea
#
tantek return 404 for days without logs?
#
tantek.com edited /IRC (+51) "clean up IRC applications section, minimize to those either one per platform, or actively in use only" (view diff)
#
snarfed !tell gRegorLove just fyi, i fixed https://github.com/snarfed/bridgy/issues/133 . let me know if it works ok for you!
#
Loqi Ok, I'll tell them that when I see them next
BjornW joined the channel
#
aaronpk okeeee
#
aaronpk days with no logs now return 404
#
tantek hmm - 2nd thought - maybe only days before 2011-02?
#
aaronpk ?
#
tantek realizes maybe it is useful to pull up logs and browse "Next" across logs in the past for months where we were not active EVERY SINGLE DAY?
#
aaronpk next/prev links still work, it just sends back 404
#
tantek oh neat!
#
aaronpk also even though a page is not linked from the main calendar view there may still be activity: e.g. http://indiewebcamp.com/irc/2011-09-27
#
tantek a-ha
#
aaronpk i just specifically made those not be linked on the calendar cause that's not very useful
#
tantek.com edited /IRC (+77) "install, and trying a little re-ordering to help readers first" (view diff)
#
aaronpk for example, first day: http://indiewebcamp.com/irc/2011-02-12
#
aaronpk click "prev" and you can keep going back. but crawlers probably won't
#
tantek ok take a look, I tried to reorder the /IRC page a bit
#
tantek aaronpk - I bet we have client logs from before that date
#
aaronpk how?
#
tantek since those logs are only from since Loqi joined, right?
#
tantek e.g. logs in Colloquy locally
#
tantek or your bouncer
#
aaronpk usually the first thing I do when I join a new channel is add Loqi to it
#
@vasilis I wrote about the ideas behind Voto, my self hosted, IndieWeb inspired photo gallery. http://nerd.vasilis.nl/voto-self-hosted-indieweb-photo-gallery/ (twitter.com/_/status/467783779437015040)
#
tantek aaronpk - however if you read the 2011-02-02 logs it certainly sounds like we've already been chatting in that channel for a while
#
aaronpk found the beginning of my logs, i'll add them
#
tantek oh! cool. so there were more!
#
aaronpk barely
#
aaronpk 20 lines or so
#
tantek really that's it?
#
aaronpk unless I joined from a different IRC client, which is unlikely
#
aaronpk I know you're not supposed to hide text from crawlers, but... I'm thinking about hiding all but the main linked line on IRC log permalinks
#
aaronpk should improve the google results when searching for a specific phrase
#
tantek yeh
#
tantek aaronpk, I see this from you earlier that day in the logs: "Wiki edits sent to this channel? [ ] yes / [ ] no"
#
aaronpk tantek: yeah that's the first thing in my logs
#
tantek I posted this the day before the logs started: https://twitter.com/t/status/35853533420978177
#
@t indiewebcamp.com is up! Are you an #indieweb builder? Sign up and join us: http://indiewebcamp.com/Guest_List http://ttk.me/t4AF2 (twitter.com/_/status/35853533420978177)
#
aaronpk wonder why twitter isn't autolinking those links
#
aaronpk oh because t.co didn't exist
#
rascul i still feel that we need indieweb.camp
#
tantek and because it was before Twitter autolinked!
#
aaronpk no way, twiter autolinked since the beginning i thought
#
aaronpk also that would imply they are caching the rendered HTML of tweets which I seriously doubt
#
tantek nope
#
aaronpk i like my answer better :)
#
tantek they are caching the rendered HTML yes
#
tantek they did not autolink since the beginning
#
tantek in fact, everytime they changed *how* they autolinked, the changes only applied to future tweets
#
tantek e.g. they used to require http(s): for links
#
aaronpk serioulsy? examples?
#
tantek yes, when they stopped requiring http(s) for links
#
tantek then future links without protocol started autolinking
#
tantek while the old ones never got linked up
#
@rossk @dan_munz the attempts to reboot personal publishing ('the indieweb') and privacy (tools like keybase) are interesting. (twitter.com/_/status/467786627352252416)
#
tantek in fact, when I started Falcon, one of the reasons was Twitter didn't autolink
#
tantek e.g. https://twitter.com/t/status/19456845265
#
@t @robdolin thanks! @Falcon project site is http://ttk.me/w/Falcon, no photo/link support yet, but it does auto-link ... http://ttk.me/t46u1 (twitter.com/_/status/19456845265)
#
aaronpk ugh, how do I solve this second-resolution problem with the irc logs
#
aaronpk that was a terrible idea
#
tantek aaronpk, for example I made this change in Falcon to stop Twitter from auto-linking all my permashortlinks: https://twitter.com/t/status/49323828383453184
#
@t New @Falcon permashortlink design: (ttk.me) = citation only, no extra content. explicit http ttk.me URL = more to see. (ttk.me/t4As3) (twitter.com/_/status/49323828383453184)
#
aaronpk oh right!
#
tantek and then they changed it again after that!
#
aaronpk i'm tempted to switch the irc permalinks to use the database id
#
tantek noooooooooo
#
tantek database id = side effect of a cache!
#
aaronpk i know...
#
aaronpk but 1400363946 is not sufficient precision to find a single record
#
aaronpk and 1400363946-1 1400363946-2 seems like a lot of work
#
tantek is this the - multiple statements in the same *second* problem?
#
aaronpk yeah
#
aaronpk it's happened a number of times in normal conversation now, but also happens almost every time loqi prints out a tweet from a link
#
Loqi yeah!
#
aaronpk e.g. http://indiewebcamp.com/irc/2014-05-17/line/1400363938
#
aaronpk click the next line down and it disappears
#
aaronpk i'm unfortunately not storing millisecond precision in the logs, so I can't use that
#
aaronpk but i'm open to suggestions
#
tantek well there are multiple problems
#
tantek first of all - nothing should be disappearing, ever
#
aaronpk it's not "disappearing" it's just unaddressable
#
tantek worst case, the permalink should always show the *first* thing that matches
#
tantek and the rest below
#
aaronpk hm I might be able to solve half the problem
#
tantek I'd say solve that for now
#
aaronpk oh wait crap...
#
tantek so at least there's no apparent data loss in the presentation
#
aaronpk I can't assume database IDs are sequential
#
aaronpk cause I often re-import things later
#
aaronpk it would be very very close though
#
tantek yeah
#
aaronpk hm i have another hacky idea
#
tantek aaronpk, so this is an interesting clue: http://indiewebcamp.com/irc/2011-02-12/line/1297543213
#
barnabywalters aaronpk: potential solution: start storing millisecond-precision datestamps, and go through tacking on .1 second to each duplicate row from the past
#
tantek thus implying, that before that edit, Loqi was sending indiewebcamp edits to ANOTHER channel
#
aaronpk barnabywalters: I kind of have to avoid any solution that involves changes to the existing data, since there is so damn much
#
aaronpk tantek: oh funny... let me check my local logs to see if I notice any other channels
#
tantek grep *all* your local logs for indiewebcamp site wiki edits from Loqi
#
aaronpk ah, microformats
#
aaronpk just one line
#
aaronpk ok
#
aaronpk http://indiewebcamp.com/irc/2014-05-17/line/1400363946
#
aaronpk the line above has the same timestamp, so clicking it just takes you to the same page
#
aaronpk but now nothing is missing
#
tantek.com created /DOI (+1060) "stub with obvious criticism, citation of GitHub silo support in some context" (view diff)
#
tantek aaronpk - much nicer!
#
tantek bret: I started this since you brought it up: http://indiewebcamp.com/DOI
#
bret i have never heard of them before
#
aaronpk not sure if it's worth trying to make the other line have a unique URL
#
tantek bret - that's why I wrote up the page
#
bret seem clear.. reminds me of the UUID stuff in atom feeds
#
tantek there's so much architecture astronomy talk in any use/description of DOIs that you could easily waste hours of your time before it was clear it was a waste of time
#
bret overy complex, not very useful
#
tantek.com created /doi (+17) "r" (view diff)
#
aaronpk what is DOI?
#
Loqi http://indiewebcamp.com/DOI
jedahan joined the channel
#
aaronpk oh good lord
#
bret "Referring to an online document by its DOI provides more stable linking than simply referring to it by its URL, because if its URL changes, the publisher need only update the metadata for the DOI to link to the new URL."
#
pdurbin very common in the academic publishing world
#
bret sooooo, here is an example of a citation backing that statement up: http://www.sciencedirect.com/science/article/pii/S1096751603000897
#
bret and another one: http://www.businessweek.com/stories/2001-07-22/online-extra-how-the-digital-object-identifier-works
#
tantek bret how do those back that up?
#
KartikPrabhu why does the publisher not just replace old URL with new URLs?
#
bret seems kind of like a tautological argument
#
bret tantek ;) a joke
#
tantek lol
#
aaronpk until very recently, github didn't send http redirects for URLs when you'd rename or move a repo
#
bret i used two urls to point to the 2 of the 3 citations to back up the above quote.. one which is a tautoligy of the quote, the other being a paid for article
#
tantek "since we can't trust ourselves to update system A, we've created system B instead!"
#
KartikPrabhu DOIs are very common way to "cite" papers but never got the point
#
KartikPrabhu tantek: yeah that is what it sounds like
#
tantek it's all handwaving in terms how DOI bottlenecks are any more stable than someone running their own domain
#
bret KartikPrabhu, except most PDFs i ever come across dont have the DOI with them! and the DOI is very hard to find even from the metadata in the paper
#
KartikPrabhu bret: agreed!
#
tantek.com edited /DOI (+292) ". and Unobvious time waster" (view diff)
#
bret like if the DOI was the hash of metadata, like newbase60 or something, then it would be interesting
#
tantek what is a doi
#
Loqi http://indiewebcamp.com/DOI
#
aaronparecki.com edited /DOI (+0) ". after citation" (view diff)
#
aaronpk what is a doi?
#
Loqi A DOI is an abbreviation for Digital Object Identifier, "a character string used to uniquely identify an object such as an electronic document"[1] http://indiewebcamp.com/DOI
#
tantek.com edited /DOI (+0) "cite before ." (view diff)
#
tantek what is a doi
#
Loqi A DOI is an abbreviation for Digital Object Identifier, "a character string used to uniquely identify an object such as an electronic document[1]" http://indiewebcamp.com/DOI
#
bret i swear academics are going to sink their own ship
#
tantek bret - they've been sinking their own ships online ever since RDF. XML. etc.
#
tantek people keep giving them money to build new ships to sink
#
bret aaronpk can do as many test posts with ownyourgram once it gets a 201 with a location header?
#
tantek.com edited /DOI (+42) "note obvious problem up front in dfn" (view diff)
#
tantek what is a doi?
#
Loqi A DOI is an abbreviation for Digital Object Identifier, "a character string used to uniquely identify an object such as an electronic document[1]", kind of like a URL with more bureaucracy http://indiewebcamp.com/DOI
#
tantek that's better
#
aaronparecki.com edited /DOI (-2) "remove leading "A" from dfn line" (view diff)
#
aaronpk what are DOIs? <-- should that work?
#
bret can "since we can't trust ourselves to update system A, we've created system B instead!" be added as a pull quote?
#
tantek.com edited /DOI (+4) "linky" (view diff)
#
aaronpk can I cite IRC lines?
kbs joined the channel
#
barnabywalters aaronpk: only if you get ACME Academic Co. to “issue” a DOI for them
#
bret requires a $10,000 subscription each year
#
aaronpk like all of tantek's IRC lines that ended up in that article
#
barnabywalters bret: I was going to say “fax machine” but that works too ;)
#
aaronparecki.com edited /DOI (+112) "cite lines from irc" (view diff)
#
aaronpk like that
#
bret and a FAX MACHINE! >:{D
#
kbs library science meets the web :)
#
barnabywalters argh — anyone know why XMLHttpRequest doesn’t honour Connection: close headers?
#
aaronpk wow, no
#
barnabywalters in responses, that is
#
tantek.com created /architecture_astronomy (+818) "stub with dfns, quote, see also" (view diff)
#
barnabywalters hrm, might be able to shim the required behaviour
#
kbs I guess DOIs really want to be more like a sha than a url
#
tantek fighting patterns with other patterns
#
tantek.com edited /antipatterns (+29) "/* See also */ architecture astronomy" (view diff)
#
tantek what is architecture astronomy?
#
Loqi architecture astronomy is the practice of analyzing problems, seeing patterns, and then generalizing to higher and higher level abstractions on top of those patterns to the point where the abstractions become so general, so vague, so detached from the original problems being analyzed, that they don't mean anything at all http://indiewebcamp.com/architecture_astronomy
#
kbs envisages google indexing documents by signature
#
tantek kbs - we have those that anyone can "issue". They're called permashortcitations
notifix joined the channel
#
tantek what is a permashortcitation?
#
Loqi A permashortcitation (or permashortid) is a short non-hyperlinked citation to a post permalink, in contrast to a permashortlink which does hyperlink to a post http://indiewebcamp.com/permashortcitation
#
kbs one question is whether they are really permanent
#
tantek "really permanent" <--- that's funny
#
tantek like security, it's all about probability, lifetime etc.
#
aaronpk i wonder... will reply context fetching code treat replying to this URL as replying to the wiki page? http://indiewebcamp.com/irc/2014-05-17/line/1400366526
#
aaronpk using the original post discovery rules?
#
aaronpk barnabywalters: this is not returning any results http://waterpigs.co.uk/services/original-post/?url=http%3A%2F%2Findiewebcamp.com%2Firc%2F2014-05-17%2Fline%2F1400366526
#
kbs a sha of a document is more permanently associated with a document than a permalink :)
#
tantek kbs a sha of a document is even more fragile if you don't know precisely where the document starts or ends
#
aaronpk but a url of a document can be used to identify a document just as well as a sha can
#
tantek with copies of documents mutating over time, even in small ways, whitespace etc. any checksum eventually fails, even if the copy "looks" fine.
#
tantek also doesn't allow for minor edits like fixing typos, punctuation etc.
#
tantek that the author has no desire to create new identifiers for
#
kbs aaronpk: hm - as long as the document remains (unmutated I guess) at the url
#
aaronpk kbs: no, see web.archive.org
#
barnabywalters aaronpk: that’s true, indieweb/original-post-discovery isn’t a spec-compliant implementation yet
#
tantek.com edited /IRC (+29) "shorten link a bit" (view diff)
#
kbs aaronpk: well - you do assume then that brewster will continue to keep documents at links for the foreeable future
#
aaronpk web.archive.org being one example
#
tantek kbs, rather, in practice, documents mutate
#
barnabywalters aaronpk: can you raise an issue on https://github.com/indieweb/php-original-post-discovery
#
tantek thus making checksums more fragile / less longterm
#
barnabywalters aaronpk: or add it to https://github.com/indieweb/php-original-post-discovery/issues/2
#
aaronpk i could just as easily ask a network of participating nodes (like DHT) "hey everyone, who knows what used to be at http://example.com/old-page?"
notifix joined the channel
#
aaronpk barnabywalters: ohh your thing only looks for twitter URLs?
#
kbs tantek: correct - so each version would be individually citeable, if that makes sense...
#
tantek everything you expect to do with a DOI, do instead with the URL. e.g. share it, put in a registry, put it in the document itself etc.
#
tantek no DOI needed.
#
barnabywalters aaronpk: I wrote the minimal amount of code required at the time to get twitter OPD working :)
#
aaronpk cool
#
tantek then if the URL breaks, you can always search for it
#
kbs is not making a case for DOIs :)
#
tantek DOI is basically a poor reinvention of URL
#
kbs but I do think that there's a distiction between "associating content with a URL", and "associating content with its checksum"
#
tantek for short-term, I can see uses for checksum. but not long term, because everything violates their checksum in the long term due to mutation (format changes, transforms etc.)
#
tantek thus I dispute this: "a sha of a document is more permanently associated with a document than a permalink"
#
kbs :-)
caseorganic joined the channel
#
kbs only sees citations to versions of documents
#
barnabywalters ah ha! XMLHttpRequest just wanted a Content-length header in order to work correctly
#
tantek in practice, people cite documents in general, e.g. books, rather than any particular printing thereof
#
bret aaronpk, getting stuck with ownyourgram
#
aaronpk bret: how can I help
#
tantek (or rather such versioined citations are the tiny minority 1% of uses)
#
kbs tantek: yes, fair enough - think you are pointing out a different problem - how to group versions of a document under one common pointer
#
tantek kbs - no, that's how people refer to documents
#
tantek they don't think they're doing any grouping
#
kbs they != me, at least :)
#
bret i'm not seeing any parameters in the body of the post request from ownyourgram when I press test post
#
tantek right, I'm putting you in the exceptional class along with people who always cite which specific printing from which publisher of a book that they read
#
tantek instead of just naming the book
#
aaronpk oh..?
#
bret now it could totally be that Im just looking in the wrong place
#
pdurbin I'm not going to argue in favor of DOIs but at least they're recognizable to academics. If they see doi:10.12345/9876 it means something to them. By comparison, URLs seem kinda... random.
#
aaronpk could be not expecting multipart form encoded?
#
aaronpk pdurbin: lol
#
bret this works: curl https://ugitpub.herokuapp.com -d h=entry -d "content=Hello World" -H "Authorization: Bearer XXXX"
#
barnabywalters pdurbin: it is hard to tell if you are being sarcastic
#
bret with the correct token
#
tantek aaronpk, pdurbin, double lol
#
pdurbin :)
#
kbs heh
#
aaronpk well trolled sir
#
pdurbin I'm just sayin'
#
aaronpk bret: so in that case curl is sending a regular x-www-form-encoded post, but ownyourgram.com is sending a multipart post so that it can include a photo
#
aaronpk if you're using a good HTTP parsing library it should handle it transparently, but since you're using node, I can't make any promises
#
tantek to be fair, I think this gets back to - it's too hard to setup your own website, content hosting/editing, and URL "minting" (lol) in a way you feel is dependable, so instead they depend on a bureaucracies (those who hand out DOIs) to do so
#
bret darn i though express would handle that
#
aaronpk express? let me see
#
tantek if we make indieweb easy for academics, plenty would ditch DOI dependency
notifix joined the channel
#
tantek or at least enough of them would, that others would follow
#
pdurbin in bureaucracies we trust
#
tantek said like an academic ;)
#
aaronpk tantek: that is well put. and is also not unique to academics
#
aaronpk see: enterprise software
#
tantek right. or hospitals. other institutions.
#
tantek it's because DIY (indieweb) is hard that they seek "solutions" from "authorities"
#
kbs git commit ids are going to be the next red flag to wave in front of tantek :)
#
tantek kbs - those are just part of URLs right? ;)
#
aaronpk bret: looks like it does handle it, but might need an additional config line http://www.hacksparrow.com/handle-file-uploads-in-express-node-js.html
#
kbs tantek: no urls involved - that's what makes them useful :)
#
pdurbin kbs: those commit IDs will go bad too someday! just like the checksums! trust nothing!
#
tantek aaronpk, caseorganic, where do academics, enterprises, governments fit into in /generations ?
#
kbs pdurbin: indeed, indeed :)
#
bret aaronpk oh this is the crap that changed in express 45
#
bret 4*
#
aaronpk oh
#
tantek kbs - there's the longerterm discussion about the DNS bottleneck problem
#
bret yeah the example i am looking at has middleware I didnt notice
#
tantek which is how git commit ids are different
#
tantek if I understand correctly
#
tantek feel free to start an "Alternatives" section here to at least capture such thoughts: http://indiewebcamp.com/DNS
notifix joined the channel
#
tantek (though it may be a while before we can put them into actual practice)
#
aaronpk hmm, barnabywalters was experimenting with some alternative DNS stuff
#
tantek yeah - gotta document that stuff!
#
barnabywalters indeed
#
aaronpk i'm still more interested in a truly decentralized DNS rather than "alternately-centralized" DNS like blockchain or .onion
#
bret aaronpk yeah express removed bodyParser and replaced it with something that does not handle multipart
#
aaronpk bret: wtf
#
tantek aaronpk, good distinction
#
bret yay, part of my knowledge is already irrelivant
#
aaronpk bret: maybe look at something besides express? i've heard bad things about it
#
bret aaronpk i hate it, but i dropped a bunch of time/effort into it so its my best chance at the moment
#
aaronpk tantek: yeah, .onion is obviously an alternate centralized registry, but things like namecoin are less obviously so
#
aaronpk bret: not true, see: http://messymatters.com/sunk/
#
aaronpk :)
#
bret i like some things about it
notifix joined the channel
#
kbs basically would like to have http://www.textfiles.com/etext/REFERENCE/prime10.txt and http://www.ccl.net/cca/data/prime-numbers/prime10.txt have the same citation :)
#
aaronpk kbs: implying that the document does not have a canonical URL?
#
kbs implying that content > location of content
#
aaronpk if you think of the URL as the name of the document, then the URL > location of the content
#
tantek aaronpk++
#
Loqi aaronpk has 465 karma
#
kbs doesn't think of URLs that way :)
#
tantek kbs - people, companies, and the media does now
brianloveswords joined the channel
#
tantek based on use in/by all three
#
tantek URLs have become names despite so many efforts to disclaim them as such
#
tantek it's gone far beyond tech/academic circles
#
aaronpk also, give me a URL like you just did and even if I don't know about it exactly, I at least can get a sense of what the document probably is, which may trigger my human brain to think of other places to look for it
#
aaronpk whereas you give me any hash and i'm just like wtf numbers
#
kbs I should hasten to point out that just pointing out the issue that I find useful (to me) rather than offer a solution [say, involving shas :)]
#
Loqi fo sho
#
aaronpk in other words, fuzzy matching is possible with URLs
#
kbs to me, again, I find that not everyone keeps consistent information at any given URL (though indiewebcamp clearly generally does)
#
tantek kbs "not everyone does" is insufficient criticism, for the counter is, "it is possible to" which is good enough for an existence proof
#
aaronpk while that may be true, I've found that more often the presentation around the content changes more than the URL does
#
aaronpk like, good luck finding a mechanism for returning a consistent ID for this document on their new and old design http://readwrite.com/2014/05/09/4g-3g-smartphone-data-price-difference
snarfed joined the channel
#
tantek kbs - your hashes would fail your use case of "… and … have the same citation" :)
#
tantek Content-Length: 524655 vs. Content-Length: 596616
#
tantek and thank you for proving the point about hashes not working in practice
#
kbs tantek: ^^ "I should hasten to point out that just pointing out the issue that I find useful (to me) rather than offer a solution [say, involving shas :)]
basal joined the channel
#
tantek kbs - just following up on your concrete example with a concrete flaw.
#
kbs what is the flaw with wanting those two documents to have the same citation?
#
tantek oh that's all, ok here you go:
#
tantek The Project Gutenberg Etext of The First 100,000 Prime Numbers, 1993.
#
aaronparecki.com created /json (+1111) "stub with dfn and notes from IRC conversation" (view diff)
#
aaronparecki.com moved /json to /JSON "proper noun caps"
#
kbs I must be too dumb to follow your argument tantek :) ah well
#
aaronpk i think tantek was giving you a citation that works for both URLs
#
tantek :D
#
kbs so, the idea is that citations should be titles of documents rather than URLs?
#
pdurbin data can be cited as well
#
tantek lacking a canonical URL, people use names of documents, e.g. names of books, movies, songs
#
pdurbin https://www.force11.org/datacitation
#
kbs this is rather circular to me anyway :) if there's a 'canonical URL' - then clearly the url is the "citeable thing". I've always thought citations were more about how to cite things that don't have that property
#
kylewm.com edited /micropub (+21) "added brainstorming section" (view diff)
#
kylewm.com edited /micropub (+528) "/* Brainstorming */ thinking out loud about how to handle markdown input" (view diff)
#
pdurbin kbs: I imagine in a full citation, you'd still want to include the author, title, year, etc. Not just the DOI or URL. For a more complete citation.
#
kbs pdurbin: yea, makes sense :)
#
kbs I think (in the case of github specifically) I've internally just assumed the commit-id as being the 'citation', with additional data (as you indicate) seems appropriate. Zenodo doesn't make sense to me for github, for that reason :) but I'm personally less sure about the 'URLs are sufficient' for all digital citations...
#
kbs also realizes this is a red flag , and starts to regret raising the issue :)
#
pdurbin kbs: what about zenodo doesn't make sense?
#
aaronpk goes back to the upload-by-email script
#
kbs aaronpk++ wise man :)
#
Loqi aaronpk has 466 karma
#
kbs pdurbin: zenode <-> github - associates a fairly random number with a release, and that zenodo stays around for the long term. Compared to - commit-id which is implicitly a 'canonical-id-for' a release, and allows replication of a (verifiably identical) document outside of zenodo
#
pdurbin kbs: would it help if zenodo were open source?
#
kbs pdurbin: guess it's more that I'm not seeing the DOIs that it offers, solve a problem that I find needs solving [specifically as it relates to github]
#
kbs I think long-term archiving of content is a problem that it might be addressing, dunno - but that's different from creating a DOI for a git release...
#
pdurbin ok. makes sense. DOIs are not for everyone :)
#
kbs apologizes if pdurbin is associated with zenodo :)