#indiewebcamp 2013-04-21

2013-04-21 UTC
caseorganic, spinnerin, duckbillp and tantek joined the channel
#
tantek.com
edited /How_the_Indie_Web_Hooks_into_Hosted_Communities (+7) "/* Discussion of Share Buttons */ linkup"
(view diff)
duckbillp joined the channel
#
tantek.com
edited /IndieAuth (+12) "linkup"
(view diff)
#
tantek.com
edited /Template:indieauth-status (+18) "Twitter has been consistently up for a while"
(view diff)
duckbillp, xtof, caseorganic and b0bg0d joined the channel
#
@juanmaleon
RT @t: Have your own #indieweb site &amp
#
Loqi
want to know how to best post/POSSE/accept/display comments?
heathjs joined the channel
#
@yunnnnno37
Have your own #indieweb site &amp
#
Loqi
want to know how to best post/POSSE/accept/display comments? See: http://indiewebcamp.com/comment (ttk.me t4PZ2)
b0bg0d, deusx, catsup, bnvk, seyz, duckbillp, scor, xtof and barnabywalters joined the channel
#
@BarnabyWalters
Just hooked up my #indieweb notes to @brennannovak’s rather awesome @emoome sentiment analysis API! So all my n… http://waterpigs.co.uk/notes/1329/
#
tantek.com
edited /IndieMark (+311) "linkup, add post types, indieweb comments"
(view diff)
#
@emoome
RT @BarnabyWalters: Just hooked up my #indieweb notes to @brennannovak’s rather awesome @emoome sentiment analysis API! So all my n… http:/…
#
aaronpk
barnabywalters: what was your thought on rel="syndication"?
#
barnabywalters
aaronpk: I’m not actually using it yet, I’m using class=u-alternate u-as-downstream-duplicate
#
aaronpk
I notice your notes' links to twitter have u-alternate and u-as-downstream-duplicate, but not rel-syndication
#
aaronpk
is there a reason you aren't using it other than you haven't gotten around to it yet?
#
barnabywalters
I should add it, there’s no particular reason not to
#
barnabywalters
thanks Loqi
#
Loqi
you're welcome
#
aaronpk
oh and u-syndication apparently http://indiewebcamp.com/rel-syndication
#
tantek
aaronpk, previously barnaby had asked me what he should use, and we hadn't really thought it all through, so came up with an extension for now (u-alternate u-as-downstream-duplicate) back then
#
tantek
then when I finally worked through the full indieweb with POSSE to indieweb with POSSE comment flow, then I figured out the use of rel=syndication and class=u-syndication
#
tantek
so the current thinking/design is an iteration on what barnaby and I initially brainstormed before (when we didn't have enough information to come up with a better solution)
#
aaronpk
gotcha
spinnerin, b0bg0d, bnvk, dietrich, seyz and deusx joined the channel
#
aaronpk
is hanging out at Epicenter Cafe today if anybody is around
tantek and eschnou joined the channel
#
aaronpk
planning on hacking some in-reply-to goodness today at epicenter cafe
#
tantek
ooh! awesome!
barnabywalters joined the channel
#
aaronpk
barnabywalters: i'm playing with your php-mf2 library now!
#
barnabywalters
aaronpk: oh great! any feedback greatly appreciated :)
#
aaronpk
so far I ran into the issue of it using the short array syntax and i'm still on php 5.3
#
aaronpk
only in like 3 places though, was an easy fix. I'll send a PR for it.
#
aaronpk
actually one question I had is what happens when it doesn't find any microformat data on the page? does it return an empty items array or no 'items' element, or false?
#
barnabywalters
aaronpk: considering that 5.3 is actually at end-of-life now I’m casually against supporting it, but just changing the array syntax should be okay — a PR would be great :)
#
tantek
if no microformat items are found, I think it should still return a empty items array
#
barnabywalters
it should ALWAYS return an items key
#
aaronpk
that would be wonderful if that were documented in the readme :)
#
barnabywalters
not sure if that’s in the µf parsing spec, but it seemed right. I’ll add it now, thanks
#
aaronpk
barnabywalters: wow I didn't realize 5.3 was EOL already... I have a feeling it's going to stick around on a ton of hosts for a while though
#
tantek
I'll make sure the spec says so
#
aaronpk
hm, a new 5.3 was released 10 days ago
#
aaronpk
oh strange
#
barnabywalters
it’s a tricky one — I’m all for BC, but on the other hand that is one of the main reasons PHP has such a bad reputation
#
aaronpk
yea it's a tough one
#
barnabywalters
and 5.4 has approx. 1/2 the memory footprint in my experience
#
tantek
barnabywalters - the parsing spec is actually clear on this already: http://microformats.org/wiki/microformats2-parsing#parse_a_document_for_microformats
#
tantek
"start with an empty JSON items array"
#
aaronpk
i think the code does always return an empty array, just wasn't documented
#
barnabywalters
tantek: I think I started php-mf2 before that page was written :) must have missed the addition
#
barnabywalters
okay, clarified
#
barnabywalters
I made some progress with the iframe+postMessage stuff today — might try applying it to web actions over the next week
#
barnabywalters
so instead of a button you can opt to have an inline dialog
#
barnabywalters
for commenting in context
#
tantek
barnabywalters - that's awesome!
#
tantek
btw - just checked the uf2 parsing spec history, the bit about start with an empty items array has been there since 2012-10-15 - >6mo. ago :) http://microformats.org/wiki/index.php?title=microformats2-parsing&oldid=48186
#
tantek
I added the explicit JSON in case that prose is not clear
#
barnabywalters
tantek: ah, my mistake :)
#
barnabywalters
also had an interesting discussion about the semantics/pros+cons of using an iframe for this with brennan, which I need to solidify onto the wiki somewhere
#
aaronpk
i would be curious to hear how that discussion went
#
tantek
was the discussion here in IRC?
#
aaronpk
oh! did you meet up with him in iceland?
#
barnabywalters
no, in person :)
#
barnabywalters
yeah, he’s working for Brian too
#
tantek
whoa - awesome!
#
barnabywalters
aaronpk: brilliant, thanks!
#
aaronpk
I may have another one coming to handle failure cases better :)
#
barnabywalters
ooh, my first ever pull request! I get to press the shiny green button ;)
#
barnabywalters
also of potential interest from today: Brennan was showing me emoome, and I asked if I could just pipe my notes into it and have the sentiment analysis done on those instead of having to enter data there. Turns out that’s not possible yet, but there is a text -> sentiment analysis data API, so I hooked them up
#
barnabywalters
first sentiment-analysed note: also had an interesting discussion about the semantics/pros+cons of using an iframe for this with brennan,
#
barnabywalters
ooh, wrong paste :(
eschnou joined the channel
#
aaronpk
tantek: what should a microformat parser do if it's given an empty document?
#
aaronpk
barnabywalters: that is nice JSON. is that your storage format, or is that generated/
#
barnabywalters
aaronpk: generated, but when I move to mostly flat files, my storage format will look a lot like that
#
barnabywalters
with more microformats semantics in the author field — currently using activitystreams for some reason and it’s ugly
#
eschnou
hello barnabywalters, did you have a chance to look at your pingbacks ? let me know if you want to try to interop !
#
barnabywalters
eschnou: not had a chance to fix it yet, might have a go this evening — I’d send you a pingback to ask if you want to test pingbacks, but that may not end well ;)
#
eschnou
barnabywalters, :-)
deusx joined the channel
#
tantek
aaronpk - if you follow the algorithm, a parser parsing an empty document would exit after starting with an empty JSON items array: http://microformats.org/wiki/microformats2-parsing#parse_a_document_for_microformats
#
tantek
but I might as well be explicit about returning the JSON items array
#
aaronpk
oh now here's an interesting one...
#
aaronpk
multiple in-reply-to
#
tantek
aaronpk - in which direction?
#
tantek
one reply post that replies to multiple *other* posts?
#
aaronpk
I am replying to 4 posts with a single post on my site
#
tantek
start with the UI / presentation question
#
tantek
how do you want to present your reply on your own site? and what is your expected presentation of your reply on those other sites?
#
tantek
and do you have an existing example of replying to 4 posts with a single post on your site?
#
tantek
or is this for a post you're writing currently?
#
aaronpk
I'm about to have an example
#
tantek
ah, ok in that case, sketch out how you'd like to display it on your own site (do you plan on showing any of the context that you're replying to? how much of how many of those 4 posts?)
#
aaronpk
well considering i'm not yet displaying content of any replies, for now I will just have links
deusx joined the channel
#
barnabywalters
aaronpk: thanks for the pull request — were you experiencing problems?
#
barnabywalters
also, did you run the tests? If not I’ll do it locally
#
barnabywalters
to better phrase the first question — this does actually fix something, right? it’s always handled empty strings fine
danbri and deusx joined the channel
#
tantek
aaronpk - note that Twitter only allows a single status id for setting in-reply-to-status-id - so at some point in your publishing flow you're going to have to decide which is the *primary* post that you're replying to.
#
barnabywalters
they seem a little over confident about the permanence of the storage there…
#
barnabywalters
and don’t seem to be affiliated with internet archive
#
barnabywalters
also the fact there is no name associated with it (just generic “webmaster@archive.is”) makes me less confident in it
#
tantek
"that will always be online even if the original page disappears" - lol
#
tantek
until archive.is is sold, shutdown etc.
#
tantek
I wonder why it's either not open source, or why they don't just use the internet archive warc format and the archiveteam warc tools.
#
barnabywalters
I came across it via wikileaks, so it may be that it is run by someone with a tinfoil hat
danbri and caseorganic joined the channel
#
aaronpk
barnabywalters: thanks for the updates!
#
barnabywalters
no problem :)
#
aaronpk
i was probably passing null as the input
#
barnabywalters
so, ruby has a really strong culture of testing right? Is it to the point that it’s expected that if you contribute to a project you add tests to cover the stuff you did?
#
barnabywalters
because on the one hand I think that is a really good thing and should be done, but on the other hand I don’t want to come across as a dick who demands people write tests for the code they’re contributing
#
aaronpk
i suppose that is true
#
aaronpk
i should have updated the tests :)
#
Loqi
fo sho
#
barnabywalters
hah, no problem :) In this case a test highlighting exactly what the problem solved is would have been the biggest benefit — as I was a little confused as to exactly what problem you were having/fixing
#
aaronpk
very true
#
aaronpk
what is the "name" of an h-entry?
#
barnabywalters
aaronpk: semantically equal to the older p-entry-title
#
barnabywalters
but more consistent with implied parsing and other microformats 2 schemas
#
aaronpk
oh... it's finding the p-name of my h-card inside the entry
#
barnabywalters
hm, weird, php-mf2 should do a depth first search and then ignore any elements which were parsed
#
aaronpk
not sure if my markup is messed up or the parser is, heh
#
barnabywalters
aaronpk: I am seeing the correct result — the content of the note via implied parsing
#
barnabywalters
php-mf2 was the first project I used composer+packagis+tagging for, and I did have some weird versioning issues earlier
#
aaronpk
check out the "name" field onder properties
#
aaronpk
there's a lot of stuff under there that shouldn't be
deusx joined the channel
#
barnabywalters
aaronpk: what should be there? I don’t think you’re explicitly specifying a p-name for the note
#
aaronpk
i'm not, it's finding the p-name of my author entry but showing it as the p-name of the note
#
aaronpk
i would have expected that to be blank
#
aaronpk
also why is it not finding my author block?
#
barnabywalters
aaronpk: not finding author block because you’re not using p-author, just classic µf author
#
barnabywalters
so it pops up under children rather than author
#
aaronpk
ok there it goes now
#
barnabywalters
the name property of a microformat is never blank unless the element is blank, or a blank name is declared
#
barnabywalters
thanks to implied p-name parsing
#
aaronpk
ok so I think the problem is it's using the name of the author as the name of the entry since there isn't an explicit name record for the entry
#
barnabywalters
can you drop the parsed name value which is wrong in here/pastebin or something? so I can verify we’re both looking at the same thing
#
barnabywalters
and that is the name of the h-entry?
#
aaronpk
that is showing up under "properties" of the h-entry
#
barnabywalters
that is the correct behaviour — look at the full line ;) It’s the textContent of the h-entry element
#
barnabywalters
because you haven’t given the h-entry an explicit p-name, the parser uses the textContent of the root element
#
barnabywalters
this is how we get <span class="h-card">Aaron</span>
#
barnabywalters
but it applies to everything
#
aaronpk
huh.. ok
#
aaronpk
what is the normal name of a note? the full note text?
#
barnabywalters
aaronpk: yeah, or perhaps the truncated text? not sure
#
barnabywalters
I did actually raise this tendency of the implied rule to produce confusing output as an issue with tantek, I think we decided it was probably a theoretical problem
#
aaronpk
well it's no longer theoretical :)
#
barnabywalters
this is true
#
barnabywalters
I did have an idea for an addition to the implied parsing rule, where if there’s no explicit p-name, it looks for the first <hX> element before resorting to the entire textContent
#
aaronpk
reads up on mf2 prefixes
deusx joined the channel
#
barnabywalters
as the first heading child of an element is probably a safe bet for that element’s title
#
barnabywalters
at least, more so than the full textcontent
#
aaronpk
ok adding p-name on the e-content element works now
#
barnabywalters
aaronpk: cool :) tantek: I’d be interested to hear thoughts on adding a check for a child H element for implied p-name parsing
#
aaronpk
btw having this parser makes this stuff wayyy easier to see what's going on
#
barnabywalters
aaronpk: glad it could be of use, and thanks again for the PRs and feedback :)
#
barnabywalters
off to bed — goodnight all
#
aaronpk
barnabywalters: not sure I agree with the H element tho... shouldn't it be p element?
#
Loqi
sweet dreams
#
tantek
scrolls up
#
aaronpk
kind of got de-railed and is now trying to add mf2 to his posts
#
tantek
oh awesome!
#
aaronpk
i'm thinking I can make my site response to urls with .json on the end by running the corresponding page through the mf2 parser
#
tantek
exactly!
#
aaronpk
shoot... I just realized a major problem with the way i've got this structured
#
aaronpk
my sidebar is outside the h-entry element, but that's where the syndicated links are
caseorganic joined the channel
#
aaronpk
so that's going to be a problme for notes as well as articles
#
aaronpk
for example, http://aaronparecki.com/articles/2013/03/28/1/an-open-challenge-to-app-net the tags appear in the sidebar, so they won't be part of the h-entry