#indiewebcamp 2013-12-04

2013-12-04 UTC
dentonjacobs, KartikPrabhu and KevinMarks joined the channel
#
bear I tried but ran into the same issue
#
bear and ended up just using Persona
#
skinny bear: if you have any persona questions, i'm happy to help
#
bear no questions, Persona is working just fine
#
skinny that's what i like to hear
#
bear tho, I would love to be able to have a single persona account but with multiple emails
#
skinny tell me more
#
bear I have 3 persona accounts now
#
bear bear42@gmail.com bear@code-bear.com and bear@bear.im
#
skinny do you see the email picker?
#
bear and I can only link my Mozillians id to one of them
#
bear right, "add another email address"
#
bear but that only allows me to switch which of the 3 emails to use
KevinMarks joined the channel
#
skinny bear: you want to be able to sign into mozillians with any of them?
#
bear yes
#
bear sorry, this may be a Mozillians use of Persona question and not a Persona login question
#
skinny gotcha. why for?
snarfed joined the channel
#
bear when I left mozilla I wanted my original Mozillians account to flip to my personal email
#
bear instead I had to create a new one
#
skinny bear: yah, that's a site issue, not a persona issue. This comes up all the time for Mozillians.
#
bear yea, so I just use a different email for each persona enabled site :)
#
skinny really? how so?
#
bear like right now, wanted to login to indieauth using my bear@code-bear.com, changed the email, clicked on the link in email to sign-in
#
bear got a password error
#
bear but if I go to the Mozillians site, select that email, password works
#
skinny why were you changing the email?
#
bear just to test if IndieAuth allowed me to use both of my persona identities
#
skinny indieauth is where my knowledge runs out. maybe someone else round here can help.
#
bear ah, but i'm not even on the indieauth site
#
bear this is Persona trying to get me to sign-in so it can verify that the email I gave is me
#
skinny are you having to verify the email address more than once?
#
bear can't get past the first one, password keeps getting bounced as wrong
#
bear persona.org sent me a confirmation email for indieauth.com
#
bear I click on it
#
bear the "Last step!" web page comes up
#
bear I enter bear@code-bear and my Persona password for that email
#
bear and it replies password incorrect
#
skinny leeme try. one sec
bnvk and friedcell joined the channel
#
skinny !tell tantek Getting this going on new site... not sure what's wrong here. Does error msg mean twitter doesn't have a link to personandplace or that personandplace doesn't have a correct rel=me? Big difference. http://cl.ly/image/3V3e3V1F2P1j
#
Loqi Ok, I'll tell him that when I see him next
lukebrooker and bnvk joined the channel
#
bret skinny: you need to add a rel=me to your twitter link on http://personandplace.com/
#
skinny bret: I did!
#
bret hrmm... didn't seee it in the footer
#
skinny bret: line 108, not in the footer
#
bret one sec
#
bret oh there it is
#
skinny fixed to https... changed twitter profile to point to personandplace.com...
#
bret rescanned and all that?
#
bret in indieauth
#
skinny rescanned a bunch
#
skinny same schmame
#
bret hrmm
bnvk joined the channel
#
skinny the error message seem to indicate that neither is correct
#
bret not sure right now and im getting distracted with dinner i can help later
#
skinny kk. thx!
#
bret you may want to try an alternate auth provider, but its weird its not seeing your rel me links for twitter
bnvk, snarfed, skinny, fritzy, josephboyle, hober2, saurik_, realzies, tpinto, lukebrooker, josephboyle1, neuro`_, dietrich_, KartikPrabhu1, Jeena_, lukebrooker_, acegiak_ and realz joined the channel
#
Loqi [mention] http://kartikprabhu.com/article/no-comment linked to http://indiewebcamp.com/webmention (webmention)
#
Loqi [mention] http://kartikprabhu.com/article/no-comment linked to http://indiewebcamp.com/principles (webmention)
#
KartikPrabhu1 auto-sending webmentions! w00t!!
KartikPrabhu and poppy joined the channel
#
kartikprabhu.com edited /webmention (+496) "/* IndieWeb implementations */" (view diff)
#
KartikPrabhu anyone played with the Firefox reader mode?
melvster, rtaibah, cweiske, friedcell, shaners, netweb and skinny joined the channel
#
brianhendrickson.com edited /events/2013-12-04-homebrew-website-club (+98) "/* RSVP */" (view diff)
#
KartikPrabhu anyone have some insight into using lxml vs beautifulsoup for parsing html?
#
aaronpk KartikPrabhu: nice post!
#
KartikPrabhu which one?
#
aaronpk http://kartikprabhu.com/article/no-comment
#
KartikPrabhu oh i see! thanks :)
#
KartikPrabhu I used it to test auto-sending webmentions :D
friedcell joined the channel
#
aaronpk seems to have worked!
#
Loqi [mention] http://werd.io/2013/im-looking-forward-to-attending-homebrew-website-club-again linked to http://indiewebcamp.com/events/2013-12-04-homebrew-website-club (pingback)
LauraJ joined the channel
#
Loqi [mention] http://aaronparecki.com/replies/2013/12/03/2/hwc linked to http://indiewebcamp.com/events/2013-12-04-homebrew-website-club (webmention)
#
aaronpk oh funny, p3k thinks calagator is the canonical URL for the event because it's listed as a u-url on the indiewebcamp page
#
aaronpk I wonder if that link should have been a u-syndication instead of u-url
Jihaisse joined the channel
#
KartikPrabhu interesting. is it really a syndication though?
#
aaronpk yeah kinda
KartikPrabhu1 and KartikPrabhu joined the channel
#
bear KartikPrabhu - i'm using beautifulsoup4 in my webmentions python lib
#
bear it's working nicely
#
KartikPrabhu that is what I have at the moment too. but was just wondering. lxml is supposed to be faster but has C-dependencies
Jihaisse joined the channel
#
bear my thought was that these are all people triggered events, so time to process is much lower on the scale than easy to read/learn/tweak code
#
bear if I was processing twitter/facebook scale events, sure - but beautifulsoup4 can handle 1000s per minute and that is fast enough
#
KartikPrabhu yes. that was my conclusion too. I might not be an issue for low traffic sites
andreypopp, glennjones, rtaibah, pfenwick, poppy, Jihaisse, ozten, aaronpk, b0bg0d_ and edsu_ joined the channel
#
Jihaisse Hello
michel_v, jonnybarnes, marjolein, BjornW, glennjones, josephboyle, pfenwick, abrereton and tpinto joined the channel
#
@jontangerine @chloeweil R.I.P. Mildred but all this indieweb real name stuff is rather ace. (twitter.com/_/status/408178704833982464)
rtaibah, adactio, melvster, glennjones, skinny, dentonjacobs, pfenwick and netweb joined the channel
#
@AbolarinAyobami RT @t: nice #IndieWebCamp Hollywood @Farmhouse. discussed: * composite feeds * pinned posts * travel data and more: http://tantek.com/2013/307/t3/indiewebcamp-hollywood-composite-pinned-posts (twitter.com/_/status/408233525260058624)
tilgovi, pfefferle, barnabywalters, HereticLocke, adactio, snarfed, eschnou, eschenal and KartikPrabhu joined the channel
#
KartikPrabhu hello!
#
aaronpk good morning!
KevinMarks joined the channel
#
KevinMarks I talked to Australian radio about indieweb a while ago, they've just posted it http://www.abc.net.au/radionational/programs/futuretense/a-healthy-digital-ecology/5131856
#
aaronpk oh neat!
#
Jihaisse Excellent !
#
Jihaisse that remember me, that I talked about it this saturday : http://philippe-couzon.com/2013/12/03/interviews-quest-ce-qui-vous-passionne-dans-le-web-wia9/
#
Jihaisse it's in french… :/
#
aaronpk very cool! we have a strong french contingent here :)
#
Jihaisse :)
#
aaronpk ...and with no explanation or changes on my end, indieauth is no longer getting SSL errors talking to Twitter
#
KartikPrabhu oh computers! gotta love them
#
aaronpk I did add an issue to give better feedback when that happens again https://github.com/aaronpk/IndieAuth/issues/40
melvster and KevinMarks joined the channel
#
aaronpk KevinMarks: is there a recording linked there? I can't find it
#
KartikPrabhu adactio: how are you parsing webmention author names on adactio.com?
#
adactio KartikPrabhu: if it's marked up in h-entry, I grab the author value
#
KartikPrabhu and do you store mentions and parse them asynchronously?
#
bear @aaronpk twitter has been rolling out new certs on their api servers
#
bear so you may have just caught them with some bad provisioning
#
aaronpk interesting. this is twitter.com btw, not the API
#
bear ah, then ignore the above :/
#
bear (unless one of the twitter ops folk fixed something they discovered as they went along with the other changes)
#
aaronpk likely
#
KevinMarks I'm on roaming mobile in London at the moment, so hard for me to tell
KevinMarks joined the channel
#
KevinMarks The producer said it would be broadcast on Sunday and on the website before then
#
aaronpk oh! didn't even notice the date on that article is in the future
#
KevinMarks Weren't you and Tantek working on future posts?
#
aaronpk tantek already implemented it!
#
aaronpk so he can create a post, but it doesn't appear until the post date
#
barnabywalters there’s probably good cause to add a banner to future posts saying “this doesn’t officially exist yet”
#
aaronparecki.com edited /videos_about_the_indieweb (+372) (view diff)
#
barnabywalters I used to have that show up when people tried to see notes from the future
#
aaronpk I don't have future posts, but I have draft posts, with a similar banner
andreypopp joined the channel
#
barnabywalters so, for future posts, should they show up if you visit their permalink or not?
#
aaronpk barnabywalters: that is exactly what tantek used it for
#
aaronpk if/when I get around to it, I probably will not have them show up on the permalink until the post date, because I already have a "draft post" feature
#
aaronpk you can only see drafts if you're logged in, so I can share drafts with people that way
fritzy, npdoty and tantek joined the channel
#
Loqi tantek: skinny left you a message on 12/3 at 5:14pm: Getting this going on new site... not sure what's wrong here. Does error msg mean twitter doesn't have a link to personandplace or that personandplace doesn't have a correct rel=me? Big difference. http://cl.ly/image/3V3e3V1F2P1j
_6a68, fmarier, bnvk, ryana, npdoty, snarfed, caseorganic, josephboyle and andreypopp joined the channel
#
aaronparecki.com edited /p3k (+241) "/* Publishing Other Content */ pushups" (view diff)
#
tantek.com edited /events/2013-12-04-homebrew-website-club (+10) "add u-uid to canonical event URL to see if that helps p3k (cc: aaronpk) p.s. canonical URL algo: 1st u-uid if any, else 1st u-url" (view diff)
fritzy, tantek, shaners, barnabywalters and ryana joined the channel
#
bret Does anyone know a good self hosted IRCCloud like program?
#
bret browser based
tpinto and tilgovi joined the channel
#
aaronpk bret: you'd need a bouncer and a web IRC client
#
aaronpk there are a bunch of open source bouncers you can install really quickly
#
aaronpk this is a reasonable browser interface for an IRC client https://kiwiirc.com/
#
bret i have a bouncer already
#
bret ill check it out
#
aaronpk hm what ar eyou looking for in a browser IRC client then?
bretwi joined the channel
#
bret just curious really
#
bretwi testing 123
#
bear i'm liking the changes that ircrelay folks have made recently - they now recognize multiple device connections and will send history to both instead of just the first one to connect
#
bear ircrelay is a hosted bouncer service
#
bret i do ZNC, its surprisingly trivial to set up
#
npdoty I've had trouble with getting bouncers installed, but will check out ircrelay
#
bret I can help anyone out with ZNC if they are interested
#
bear ircrelay is using a custom ZNC
#
tantek good morning #indiewebcamp!
#
tantek reminder: Homebrew Website Club meeting(s) tonight! http://indiewebcamp.com/events/2013-12-04-homebrew-website-club
#
tantek aaronpk - presumably you saw skinny's feedback about IndieAuth http://cl.ly/image/3V3e3V1F2P1j ?
#
tantek would it be possible to parameterize that error message?
#
tantek e.g. instead of "… link on this site pointing to your website"
#
tantek how about "… link on twitter.com/skinny pointing to http://personandplace.com"
#
tantek I think that's what skinny is asking for clarity regarding: "Does error msg mean twitter doesn't have a link to personandplace or that personandplace doesn't have a correct rel=me? Big difference."
#
tantek !tell skinny follow-up re: IndieAuth error message clarity: https://indiewebcamp.com/irc/2013-12-04#t1386184156 - just confirming that's what you're suggesting.
#
Loqi Ok, I'll tell them that when I see them next
fritzy and KartikPrabhu joined the channel
#
aaronpk tantek: yeah that seems reasonable
josephboyle joined the channel
#
aaronpk bret: https://github.com/aaronpk/p3k/issues/1
#
bret cool thanks aaronpk
#
bret How do you like working with YAML headers? Anything you dont like? I've started work to move off of jekyll and was planning on keeping them around like you did
#
aaronpk I kind of like it. I know tantek probably doesn't :)
#
aaronpk i'll add another example of something more complicated
abrereton joined the channel
#
bret I'm thinking of flat files, yaml headers, markdown posts, and dom templating with https://github.com/tmpvar/weld
#
tantek aaronpk I'm undecided
#
aaronpk interesting
#
barnabywalters fwiw, I’ve found that storing data in YAML + markdown quite bearable provided the vocabulary used is solid, i.e. mf2 vocabularies
#
tantek it's a tough balance between ease of reading/debugging the raw text of the storage file(s), and the ability to quick view it in a semi-WYSIWYG manner in a browser.
#
tantek (Falcon storage is currently biased towards the latter)
#
aaronpk luckily yaml is pretty close to a plaintext format as in it's very readable and editable in its raw form compared to JSON
#
tantek right - that approach is biased more towards the former
#
tantek as is using markdown for authoring
#
tantek plain text based authoring does feel easier / less cumbersome. I often write my blog posts out in plain text + <a href> tags first, and then when they feel "90% done" start adding markup (including links)
#
tantek the process of doing so causes me to re-read all the text several times and fill in details (figure out where screenshots should go etc.)
#
bret I find that many markdown implementations are not expressive enough, and dont allow for easy access to a lot of html tags that would be nice
#
barnabywalters tantek: sounds like content-out content authoring :)
#
bear markdown works for me only because I normally write really simple posts :)
#
bret it would be nice to be able to easily extend the markdown processor too
#
barnabywalters start with the most basic, most important part of the content and progressively enhance it
#
barnabywalters plaintext + links + paragraph structure
#
tantek yeah
#
tantek paragraph structure is already implied by double line breaks
#
aaronpk bret: markdown technically allows you to add arbitrary html tags
#
tantek as are headings by single lines without a terminating period
#
aaronpk also I've found it relatively easy to add extensions to markdown parsers: http://aaronparecki.com/articles/2012/09/01/1/some-enhancements-to-markdown
#
tantek I've had trouble getting over the warts and ugly hacks in markdown
eschenal joined the channel
#
aaronpk i break out into HTML often, but I prefer things like ![:youtube 600x400](G-M7ECt3-zY) over a full youtube embed code
#
tantek I'll likely end up forking markdown pretty seriously at some point.
#
aaronpk a full fork vs just a bit of pre-processing? I'm curious what you'd change
#
tantek more strictly following it's fundamental premises
#
tantek yes a full fork because even the "original" baseline had too many ugly divergences from it's initially well-founded clean methodology
#
KartikPrabhu how do you guys insert mf2 classes into something written in markdown?
#
KartikPrabhu or any such markup?
#
bret KartikPrabhu, thats a question I struggle with too
#
bret in jekyll sans plugins
#
barnabywalters KartikPrabhu: on the wiki, typically by manually typing HTML tags
#
tantek aaronpk - have I not shared this with you before? thoughts in progress: http://tantek.com/w/Markdown
#
aaronpk KartikPrabhu: my event posts for example are rendered by p3k, I store the raw data in YAML, so I'm not actually writing HTML for event posts: https://github.com/aaronpk/p3k/issues/1#issuecomment-29837806
#
aaronpk tantek: oh cool, no I hadn't seen that before
#
tantek KartikPrabhu - <span class=""> should work ;)
#
barnabywalters on personal sites, typically manual or auto-linking HTML
#
tantek wow it's been over a year since I wrote that up
eschnou joined the channel
#
barnabywalters and auto-microformatting e.g. @name -> <a class="h-card" href="…
#
KartikPrabhu tantek: yes but then you are just writing raw HTML instead of Markdown
#
tantek KartikPrabhu - no, see what aaronpk said above: markdown technically allows you to add arbitrary html tags
#
aaronpk tantek: I like your "....." and ". . . ." for additional header levels
#
tantek aaronpk - that's pretty funny - upon re-reading I'm consider even those a bit unnatural ;)
#
aaronpk oh really? They seem to do a pretty good job of indicating smaller header levels
#
tantek yes they're an incremental improvement on the current state
#
aaronpk ====== ------ ...... . . . . looks like a natural progression
#
tantek however I want to keep considering simpler (more natural to type) alternatives
#
KartikPrabhu tantek: I know that you can write arbitrary HTML tags in Markdown and they'll come out right. But that seems a bit dirty, to insert HTML markup into Markdown files
#
iamshane.com edited /icon (+833) "added me to /icon examples. since: 2012-11-18" (view diff)
#
bret looking forward to tonight :)
#
bret Whos is in PDX right now?>
#
aaronpk tantek: well anyway, I could get behind forking markdown w/you and writing a PHP parser
#
tantek ooh - that's motivation to keep updating that page :)
#
aaronpk I wonder if there's a near-plaintext version of events that we could add to the parser to output a proper h-event
#
tantek I'd start simpler (or more frequently used), e.g. mentions of people
#
aaronpk sure
#
aaronpk well I already do @person for that
#
tantek I'd also try to re-use prior art from http://microformats.org/wiki/picoformats
#
aaronpk great!
#
aaronpk oh funny.. ajturner is on that page!
#
aaronpk small world
#
tantek I think plain text *mobile* authoring is now a much more important design consideration than plain text *desktop/laptop* authoring which is what Markdown was designed for/around/based on
#
tantek so another approach would be to redo markdown *completely* based on mobile text authoring considerations / patterns / usage patterns
#
aaronpk plaintext mobile event posting would be <3 for me
#
aaronpk there have been already a couple cases I've wanted to do that
#
tantek right! so add to the picoformats page (stuff you've seen/used in the past) and perhaps we need a picoformats-brainstorming page
#
tantek (for new things we want)
#
barnabywalters new versions of taproot running at http://test.waterpigs.co.uk/ if anyone wants to try to break it
#
tantek so perhaps fixing markdown is not enough
#
tantek perhaps we should recreate a mobile-first version of markdown
#
barnabywalters http://test.waterpigs.co.uk/notes/new may also be of interest
#
tantek ooh that's cool barnbaywalters!
#
tantek perhaps add that to /Taproot on the wiki?
#
barnabywalters tantek: doing
fritzy joined the channel
#
barnabywalters I finally caved and added a homepage feed
#
barnabywalters starting to build a feed reader-type thing was the last straw :)
#
bear tantek - if you get a picoformat version of something that allows me to use it instead of markdown… I'll write the python parser and be giving you hugs for months
#
barnabywalters I also found this which is nice guidance on what file permissions stuff should have: http://serverfault.com/questions/357108/what-are-the-best-linux-permissions-to-use-for-my-website
#
tantek bear - wow ok
#
barnabywalters not sure where to reference it on the wiki — /deployment maybe?
#
aaronpk bear hugs!
#
tantek barnaby - makes sense
#
bear I so want a markdown spirited markup that allows for contacts and links and replies and … and … - yes!
#
waterpigs.co.uk edited /Taproot (+76) "/* Roadmap */ moved rewrite, removal of flaky framework nonsense to done, added link to test deployment" (view diff)
#
tantek also relevant to mobile-first txt authoring http://tantek.com/w/TxtShorthands (these days I'd like to auto-expand these, rather than actually publish them)
fritzy joined the channel
#
barnabywalters argh unix permissions — what should I set my umask to so that new files are 660 by default?
#
barnabywalters 007?
#
barnabywalters oh wow it worked
#
snarfed insert james bond joke here
#
waterpigs.co.uk created /Apache (+1133) "dumped stub with anecdotes+links about file permissions" (view diff)
#
barnabywalters ^^^ dumped a bunch of stuff about apache above, feel free to clarify
poppy and tpinto joined the channel
#
aaronpk interestingly, the convention of a YAML header block at the top of a file is eerily similar to email encoding
#
aaronpk email + http actually. header block at the top, two newlines, then contents of the email or HTTP document
julian` joined the channel
#
julian` hey guys ive been given assignment to make a mock fansite of a television show and ill be marked on my semantics and structure and the use of new html5 elements, can you tell me how good those things are and note anything i could improve on please i'd be hugely greatful. url is http://julian.so/simpsons
KartikPrabhu joined the channel
#
KartikPrabhu julian`: It seems that on the Home page you have a header with no headings in it. In fact the heading of the page 'Home' appears to be in a section. That might be a problem if you're worried about semantics and such
#
barnabywalters does using <section> have any practical benefit, ever?
#
barnabywalters AFAIK it does two things: scopes <header>, <footer>, <address> elements, and resets the outline algorithm
#
barnabywalters so h1 > h2 and h1 > section > h1 in theory mean the same thing
bnvk joined the channel
#
barnabywalters but are extremely tricky to select with CSS, and other than that seem to have no practical benefit
#
barnabywalters apart from maybe including one document inside another without messing up the theoretical outline
#
tantek I have found <section> more trouble than it is worth and now recommend against using it.
#
tantek have been documenting fairly decent consensus about its troubles: http://microformats.org/wiki/section
#
barnabywalters in many cases it’s becoming the new <div> — giving web devs a generic container element with no meaning, but letting people claim they’re using semantic markup
julian` joined the channel
#
julian` sorry d/ced
#
barnabywalters tantek: didn’t know about that page — useful, thanks!
#
julian` do you recommend i take out the <section> tags then?
#
tantek I recommend taking out any tags that do not provide a discernible benefit for your use-cases.
#
barnabywalters julian`: unless you can demonstrate to yourself some practical benefit, I would recommend removing them
#
tantek and if you *can* document practical benefit for <section> - *that* would be interesting to know too
#
julian` i was thinking of them as splitting up the articles
#
KartikPrabhu i usually use sections for marking up parts of an article that are in some sense independent like chapters in a book
#
tantek <div> can do that already
#
barnabywalters <article> is genuinely useful as it makes authors think about syndication
#
julian` but i was told by my tutor to use these other ones and div as a last case scenario
#
tantek KartikPrabhu - if something is already "in some sense independent" then it's worthy of being an article
#
barnabywalters “can this content be taken out of context?”
#
tantek julian` ask your tutor *why*
#
barnabywalters “what content needs to be under this element for it to be reusable?”
#
KartikPrabhu tantek: they are not independent of the article itself but of each other. Maybe like sections of one chapter in a textbook or soemthign
#
julian` if you guys arent sure if it should be used or not how am i meant to be haha
#
Loqi nice
#
tantek julian` challenge your tutor to come up with practical benefits for <section> over <div> that can even outweigh just the simple cost of typing 4 more characters "section" vs "div
#
tantek " for each start/end tag :P
#
tantek and show them http://microformats.org/wiki/section
#
KartikPrabhu tantek: when you say 'practical use' do you mean in terms of syndication and styling or something? Or in a semantic sense?
#
barnabywalters KartikPrabhu: how about something which effects users in the remotest way?
#
tantek practical - what does it do for the user? what's the use-case that it supports?
#
barnabywalters I have not seen any tool which uses <section> to expose useful functionality to actual real people
#
tantek similarly, the outline algorithm (I think it was a huge complicated mistake and should be dropped)
#
barnabywalters or, rather, functionality which couldn’t be provided with <div>, i.e. styling
#
KartikPrabhu I can think of one. If I were to write a book/paper with sections I could generate a table of contents using sections.
#
barnabywalters KartikPrabhu: software has been doing that with h1-6 for years
#
barnabywalters e.g. mediawiki
#
tantek indeed
#
barnabywalters and “if I were to write” is theoretical — who is *actually doing this*
caseorganic joined the channel
#
KartikPrabhu so the assumption is that anything between one h2 and another h2 is a section?
#
tantek KartikPrabhu - or just use a div again
#
tantek so far you haven't mentioned any benefits of using section *above and beyond* just using a div.
#
aaronpk asciidoc basically does that, there's header tags and everything between the header tags is part of that chapter
#
aaronpk chapter or section of a chapter
#
tantek does markdown? (it could)
#
KartikPrabhu tantek: div can be used for everything. If I considered all divs as sections that would be a nightmare.
#
tantek no one said anything about considering all divs as sections
#
tantek but rather vice-versa, just use div where you need a section
#
KartikPrabhu and use h1-6 to delinate actual parts of an article?
#
KartikPrabhu delineate*
#
tantek (they're two different statements, just like "all squares are rectangles" and "all rectangles are squares" are two different statements, one of which is true and the other false)
#
julian` think im gonna stick to section cos i know he will mark me down for using divs
#
tantek julian` <section class="usless-html5-bling"> ?
#
julian` lool
#
barnabywalters so, looks like o'reilly are actually using section for real-world book authoring: https://github.com/oreillymedia/HTMLBook/blob/master/schema/htmlbook.xsd#L511
#
julian` it may give some functionality in the future as opposed to a div
#
julian` jump to sections on a website etc
#
tantek <section class="because-div-would-get-me-marked-down">
#
julian` LOL
#
tantek "**may** give some functionality in the **future**" = premature markup.
#
barnabywalters use-case: (allows for modularization of HTML files at the chapter level),
#
julian` that'd be his answer if i asked
#
KartikPrabhu barnabywalters: that is sort of what I was going for
#
KartikPrabhu except I was thinking of writing scientific papers
#
tantek barnabywalters you did say *practical* - when was the last time you marked up *chapters* of something?
#
aaronpk LOL! "Read the Working Draft of the HTMLBook Specification here: specification.asciidoc"
#
barnabywalters tantek: I don’t write much with chapters, but o'reilly authors do
#
tantek so then tunnelvisioning on "chapter" isn't the right approach either
#
aaronpk self-dogfood fail
#
tantek but rather, to say, how would you markup a whole book
#
tantek aaronpk LOL
#
barnabywalters point it, someone is at least using for something other than me-too “semantic” HTML bandwagon stuff
#
aaronpk I mean I know asciidoc is the standard format at o'reilly, but come on!
#
barnabywalters i.e. have written software which takes it and does things with it
#
barnabywalters it might not necessarily be a better choice or a good approach
anselm joined the channel
#
barnabywalters htmlbook in general looks over-engineered
#
aaronpk books are hard
#
tantek barnabywalters welcome to modern markup proposals
#
aaronpk cause you're talking about a markup that can be used to produce a document in html, e-reader proprietary formats, PDF for print (with appropriate spacing/page divisions), etc
#
julian` why no <content> to replace article and sections
#
barnabywalters aaronpk: I’ve been wrestling with HTML -> PDF conversion at work. very messy business
#
julian` and then use <h*>'s to divide content into importance
#
barnabywalters julian`: what would it do?
#
barnabywalters what functionality/experience/benefit would <content> offer readers/users?
#
tantek also - having written a book about HTML5, *in* HTML5, I did use <section> for the chapters (as well as nesting that inside an <ol start="[chapter-number]"><li>) and with hindsight, the <section> added *nothing*. It was a waste.
#
julian` tell the browser where the content of the page starts and ends
#
tantek julian` - <content> is called <div>
#
barnabywalters julian`: sounds like <main> to me
#
barnabywalters which, again, I have not seen any useful tools for yet
#
julian` hmm
#
julian` so what does article do
#
tantek barnabywalters - I believe accessibility tools (screenreaders?) have some support for <main> now
#
barnabywalters julian`: not browsers, *users*, specifically. the *humans* who are reading the content
#
julian` nothing
#
aaronpk <why><not><just><have><every><word><be><its><own><tag></tag></own></its></be></word></every></have></just></not></why>
#
barnabywalters tantek: oh cool!
#
tantek julian` - article indicates that chunk of content can be independently syndicated
#
barnabywalters stomps off to make a tool which removes all content on a page other than the <article> to help people test it’s correct usage
#
tantek that's both useful, and used (e.g. Readability parsers for <article> for this reason)
#
tantek heh
#
julian` so you guys dont use section?
#
tantek julian` I specifically *stopped* using section
#
barnabywalters julian`: I used to, then realised it was pretty much pointless
#
barnabywalters much of upgrading the markup on the latest version of my personal site has been ripping out <section> elements ;)
#
barnabywalters with *no* change to functionality
#
barnabywalters apart from a small speed boost ;)
#
julian` oh ok gonna put a div in and if he tells me i shouldve used a section ill tell him @t told me it was pointless
#
barnabywalters less bytes and fewer DOM nodes
#
julian` let me show u a pdf he uploaded
fritzy and andreypopp joined the channel
#
KartikPrabhu so far what I have gathered is 'just use divs to partition an article. WIth heading levels to denote new parts or subparts'. Is that the general feeling here?
#
julian` http://html5doctor.com/downloads/h5d-sectioning-flowchart.pdf
#
julian` so should a heading go inside an article or outside
#
tantek julian` you can even cite this IRC conversation :)
abrereton joined the channel
#
julian` when grouping images together
#
julian` should <figure> be used
#
aaronpk possibly the shortest TTSD by Yahoo yet? http://blog.ptch.com/post/68919650988/ptch-joins-yahoo
#
tantek julian` - I tried using <figure> / figcaption and it was too cumbersome so gave up on that too
pfenwick joined the channel
#
aaronparecki.com edited /site-deaths (+169) "/* Upcoming */ Ptch.com" (view diff)
#
tantek wow a PDF to explain which semantic *HTML* element to use - there's some irony
#
tantek julian` re: "logical to add a heading", a div with a first child that's a <hn> element works fine
#
tantek and in fact I have yet to find a reason to cluster stuff between subheads in my articles (blog posts)
#
tantek just using <h1> for the name of the post, and then <h2> for subheads (and rarely <h3> for subsubheads) has worked just fine
#
tantek no benefit/need found to explicit markup sections
#
tantek even with just a <div>
#
bear the reason I avoid using <h..> for articles is to avoid having to deal with different ages of my articles and their varying use of h1 - h5
#
tantek so again I'll push back with, if it doesn't provide a concrete user-observable benefit, the markup should be left out
#
bear <article> and <section> just make it so easy to find the true edges of a post
#
tantek bear, I do take advantage of <article>'s heading scoping and starting with <h1> inside each <article>
#
tantek that seems to work fine
#
bear oh, so the above (which I haven't fully digested yet) was geared towards removing <section> ?
#
bear just goes to read scrollback
#
tantek yes - it can't be justified based on any practical usefulness
#
bear I only really used it when I was doing markup back in the sgml days for long collections of xml documents
_6a68 joined the channel
#
barnabywalters okay, I made this: http://test.waterpigs.co.uk/services/test-article/
#
barnabywalters if it finds an <article> element, it spits out just that HTML, no CSS or surrounding HTML
#
barnabywalters to help people evaluate whether or not the element does indeed contain enough information for it to be syndicated alone
#
aaronpk should the header be included inside the article then?
#
bear nicely done!
#
aaronpk also holy crap
#
aaronpk barnabywalters++
#
Loqi barnabywalters has 21 karma
#
bear barnabywalters++
#
Loqi barnabywalters has 22 karma
#
barnabywalters YAY INTERNET POINTS
#
Loqi does a happy dance!
#
barnabywalters thanks :)
#
KartikPrabhu barnabywalters++
#
Loqi dude
#
KartikPrabhu looks like mine passed!
#
barnabywalters KartikPrabhu: yay!
#
Loqi woot
#
barnabywalters send me the URL, I’ll add it as an examile
#
barnabywalters s/examile/example
#
Loqi barnabywalters meant to say: send me the URL, I’ll add it as an example
#
barnabywalters also going to add a note about marking up with h-entry for *real* reusability
#
KartikPrabhu http://kartikprabhu.com/article/no-comment you might want to check if it passes by your reckoning
#
barnabywalters <article> is a nice official hook for getting people asking questions about standalone syndication, h-entry makes it practical
#
tantek well put
#
KartikPrabhu on a related note. has anyone used 'reader mode' in FF? It also seems to do similar things.
#
tantek that's a good summary to add to the wiki
#
barnabywalters KartikPrabhu: your article looks excellent standalone, sans-css!
#
KartikPrabhu boom! :D
#
aaronpk hm I apparently do not have <article> tags on my articles
#
KartikPrabhu barnabywalters: which is why I am curious about FF reader mode. It seems to remove my 'further reading' section :(
#
KartikPrabhu and I could not find any documentation on it
#
barnabywalters KartikPrabhu: they might be reusing the old readability code, that’s open source IIRC
scor joined the channel
#
barnabywalters safari + mobilesafari use a fork of it for their reader mode
#
KartikPrabhu thanks. will look into it
squeakytoy joined the channel
#
barnabywalters aaronpk: how do you manage file permissions on your server when deploying with git?
#
aaronpk barnabywalters: unfortunately I have all files owned by the web server right now, and the web server user pulls down from git
#
barnabywalters I’m having no end of trouble, with newly created local files not being readable for some reason
#
aaronpk I ran into too many problems with file permissions and umask trying to have both my user and the web server create files
#
barnabywalters aaronpk: at the moment I just set everything to 777 everywhere
#
aaronpk hah ugh
#
barnabywalters trying to move away from that with the new version
#
bear barnabywalters I use a script to export the git files to a tarball and then push that to the server
npdoty joined the channel
#
barnabywalters bear: how does that handle changes on both the server and the dev machine?
#
aaronpk barnabywalters: the problem I had was that both the web server user wanted to write files on disk, and also the user that ran background jobs and git pulls, and that got complicated
#
barnabywalters i.e. I’ve made code changes locally and have published several new notes remotely, and it all needs to sync up everywhere
#
bear push to branch for dev deploy
#
bear I always treat the server as the published copy
#
barnabywalters and yes, I know I should have code and data in different places ;)
#
bear :)
#
aaronpk barnabywalters: meh. my site is a pile of code and data
#
barnabywalters bear: can you document your approach on http://indiewebcamp.com/deployment?
#
barnabywalters maybe as a sub-section of Git?
#
aaronpk I thought about separating it, but it would have involved a lot of unnecessary generalization of layouts/templates and such
#
bear barnabywalters - yep, but until later tonight
#
barnabywalters bear: no problem, thanks :)
#
barnabywalters aaronpk: indeed, unnecessary+premature generalisation of taproot has cost me days of development time
#
bear for me everything is processed on my local copy and I only push the generated files (right now)
#
barnabywalters and has constituted the larges amount of code I’ve ripped out in the new version
#
aaronpk heh
#
bear what I have to change is how a webmention gets updated - and that i'm working on
#
barnabywalters I think I’m going to stick with my two repos, each with a master and deploy branch
#
bear i'm toying with the idea of mentions being their own file that is pushed to the repo
pfenwick joined the channel
#
barnabywalters I just need to figure out sensible permissions
#
barnabywalters maybe a post-deploy hook which runs chgrp -R www-data?
#
aaronpk bear: fwiw I started by storing mentions inside the post file and am moving to storing mentions in their own little subfolder
#
barnabywalters on the deploy machine
#
barnabywalters aaronpk: bear: likewise
#
bear barnabywalters - yep, that is a sane first step
#
barnabywalters bear: so if you’re saying that’s sane, and it’s fixing my current problems, I’m just going to go ahead and do that ;)
#
bear +1
#
bear just remember to remove the .git dir if you can - or at least make sure it's protected
#
barnabywalters I need the git dir there so I can commit remote-created files — so I need to protect that from the group changes?
#
bear no - protect it from web access
#
barnabywalters or just enforce stricter permissions
#
barnabywalters ah okay
#
barnabywalters it’s out of my document root
#
bear cool
#
barnabywalters does it still matter whether or not www-data can get to it?
#
bear no
#
barnabywalters cool
#
bear just that some folks mistakenly allow it to be accessed and then wonder how their ssh config or passwords get found
#
barnabywalters bear: indeed :) I’m fairly cautious about things I understand, like what files get web access. much hazier on unix permissions and groups and what needs access to what
#
bear all www-data needs is group +r
#
barnabywalters s/web access/served by apache
#
Loqi barnabywalters meant to say: bear: indeed :) I’m fairly cautious about things I understand, like what files get served by apache. much hazier on unix permissions and groups and what needs access to what
#
bear and +x
#
barnabywalters and +w to the things it writes to, right?
#
bear yes
snarfed joined the channel
#
bear but I try to keep those files/dirs very specific
#
bear i.e. only make the ones that *have* to be written to +w
#
barnabywalters indeed — limiting it to my “data” folder and web/img currently, which are the only places taproot is allowed to change files or create new ones
#
bear cool
#
barnabywalters so what do I need to do to allow my user to be able to chgrp without sudo?
#
barnabywalters (and sorry for all the questions)
#
bear it should always require sudo imo
#
barnabywalters ah — so putting it in a post-commit hook actually isn’t a good plan?
#
bear no, you can sudo access the one command for that one user
#
aaronpk well you could always whitelist that one sudo command
#
bear so put the command into a shell script and create a sudoers entry for that command for that user
#
barnabywalters okay — wow this got complicated fast ;)
#
bear good security tends to get that way
#
barnabywalters hopefully documenting it all on the wiki will help clear it up for me and guide others
#
barnabywalters so I need a line like this:
#
bear you want a sudoers line that reads something like: myuser ALL=(ALL:ALL) NOPASSWD: /usr/local/bin/foo.sh
#
barnabywalters me ALL=(ALL) NOPASSWD: /path/to/post-commit-hook
#
bear yes
#
barnabywalters cool, thanks
snarfed joined the channel
#
bear and the really paranoid would change the ALL part to have that specific hostname
#
aaronpk barnabywalters: no that would mean the post-commit-hook can be run as root, you probably want just the one specific command like bear has
#
bear ah - aaronpk makes a good subtle point
#
barnabywalters aaronpk: uh which bit does that?
#
barnabywalters the ALL:ALL?
#
barnabywalters instead of ALL?
#
bear the /path/to/post
#
bear instead of pointing to your post commit hook
#
aaronpk make a script that just does the chgrp, and save it something like /usr/local/bin/foo.sh
#
bear make a utility script that your post commit hook *calls*
#
barnabywalters so, basically wrap chgrp
#
barnabywalters or make a specific version of chgrp
#
bear so you isolate what actually needs root
#
aaronpk whee security
#
bear :)
#
barnabywalters it all makes sense, it’s just so obscure and difficult to learn without helpful people like you around :)
#
bear I have an in-house security team, so this stuff is what I deal with every day as they try to break my servers
#
bear so yea, glad to help spread the word and answer questions
andreypopp, _6a68, bnvk, fritzy and snarfed joined the channel
#
barnabywalters okay, only one task left before test.waterpigs.co.uk can go properly live — h-entry to ATOM converter
caseorganic joined the channel
#
snarfed it looks great, barnabywalters!
#
snarfed nice that it's responsive
#
barnabywalters snarfed: thanks! still working on improving the responsiveness, feedback appreciated
#
snarfed minor nit: on small screens, you might want to set a left margin or padding on your h-card
#
snarfed right now it's flush left
#
barnabywalters especially as I don’t use a smartphone every day, and so don’t notice these itches
#
barnabywalters snarfed: thanks! will fix
snarfed joined the channel
#
tantek barnabywalters - I recommend something cheap like a low end iPod5Touch.
#
tantek or even iPod4Touch
#
snarfed or for testing, just resize your browser :P
#
snarfed i usually use thin browser windows so that's how i noticed
#
tantek yes that too :)
#
barnabywalters I tend to use iOS simulator when I want to do iPhone-size testing
#
tantek it's different seeing it in your hand at the actual (physical) size too
#
barnabywalters but nothing replaces actually using the site daily on a small screen for understanding the design problems which need solving
#
tantek holding it at handheld distance etc.
#
barnabywalters hence why my site has always been optimised for iPad and 13" macbook usage ;)
#
barnabywalters unwittingly, of course, but inevitable
#
barnabywalters tantek: I have an iphone-only browser app on my iPad which gets part of the way — the device still feels very different and is used differently though
#
KartikPrabhu barnabywalters: you might want to set a margin/padding overall for small-screens. The text sits flush against the edges which looks a bit awkward. I like your menu changes!
#
barnabywalters KartikPrabhu: thanks for the feedback!
#
bear wanders to the pub
#
KartikPrabhu barnabywalters: I just happened to wander to http://test.waterpigs.co.uk/services and your Django is showing! :P
#
barnabywalters KartikPrabhu: not django, but thanks for the tip-off :)
#
KartikPrabhu aah ok. It sounded good though ;)
#
barnabywalters I actually don’t use that templating language myself — it’s part of the default error page which I haven’t overridden yet
lukebrooker and caseorganic joined the channel
#
tantek has anyone looked at or used refresh.io? thinking about how an indieweb-focused version of that would work
#
barnabywalters tantek: oh good to know someone’s building that — I have a theory that when we have seamless brain-computer interfaces this sort of thing will eradicate vapid small talk, leaving only productive discussion
#
barnabywalters isn’t the indieweb version just looking at the person’s homepage?
#
barnabywalters that could be a good framing device for homepage design
#
snarfed barnabywalters: agreed, looking at their homepage is half of it
#
snarfed the other half is relationship-specific notes
#
snarfed kind of like personal CRM
#
barnabywalters snarfed: indeed — highlight the things you and this person have in common
#
barnabywalters specifically social connections
#
snarfed oh, no, i was thinking notes about things you've actually talked about or done together
#
barnabywalters snarfed: ah okay
#
snarfed inferences too though
#
snarfed while i was at google, i started a few times to build the obvious facial recognizer for glass
#
snarfed as a side project
#
snarfed but they very clearly forbid it, even internally
#
snarfed …which is understandable :PP
#
tantek yes - the connections come from aggregations that can do queries
#
tantek which is more than you can get from just viewing someone's personal site
#
barnabywalters snarfed: that’s the basis of my brain-computer interface theory. on meeting someone, facial recognition resolves their face into a URL, fetches it and finds their profile information and lifestream, which is seamlessly injected into your memory
#
snarfed damn
#
snarfed awesome, and incredibly creepy
#
snarfed like lots of bleeding edge social tech
#
Loqi hehe
#
snarfed heh, i wonder what triggered that laugh
#
barnabywalters potentially not creepy provided only public information is used
#
snarfed sadly, that doesn't prevent creepiness
#
snarfed social norms and built in feelings like that are a little rational, but only a little, not a lot
#
barnabywalters yeah, it would be difficult to get right
#
barnabywalters fortunately it doesn’t exist yet :)
#
snarfed well, the memory injection part doesn't
#
snarfed the rest is mostly here!
#
barnabywalters true
#
barnabywalters resolving faces into URLs?
#
barnabywalters only with centralised providers like facebook
#
snarfed sure
#
barnabywalters does fb have an API where you upload a photo and it gives you back profile data?
#
snarfed but you're conflating "doable" with "indie"
#
snarfed an unofficial one, maybe
#
snarfed it can suggest people to tag in pictures
#
snarfed g+ may too
#
barnabywalters snarfed: oh sure, this is obv. a pipe dream at the moment :) but there is realistic potential to do indie social graph analysis
#
snarfed sure!
#
aaronpk face.com was exactly that, but facebook bought them
#
aaronpk (face recognition API)
caseorganic, fritzy, andreypopp, KartikPrabhu, npdoty and skinny joined the channel
#
Loqi skinny: tantek left you a message 4 hours, 16 minutes ago: follow-up re: IndieAuth error message clarity: https://indiewebcamp.com/irc/2013-12-04#t1386184156 - just confirming that's what you're suggesting.
#
aaronpk wow, medium redesign, and keyboard scrolling with up/down keys broke https://medium.com/beautiful-stories/8d615d86ac04
#
snarfed oh man, that's horrible
#
snarfed i can't stand webapps that break basic keyboard functionality
caseorganic joined the channel
#
snarfed …hmm, up and down keys work ok for me
#
aaronpk i'm on chrome
#
barnabywalters works on ff nightly
#
barnabywalters goodnight all
#
Loqi ciao
#
snarfed yeah, looks like it's a chrome/webkit bug, not ff
caseorganic joined the channel