#microformats 2015-06-10

2015-06-10 UTC
KartikPrabhu, elux, fuzzyhorns and tantek joined the channel
#
ben_thatmustbeme is still lost at times on why p- and not u-
#
ben_thatmustbeme <a class="p-author h-card" href="http://martin.example.org/">Martin</a>
#
ben_thatmustbeme why wouldn't that be u-author ?
#
aaronpk it appears to have the same parsed result either way, not sure
#
tantek aaronpk: only because the "u-* h-*" parsing fix hasn't happened yet in implementations except for mf2py
#
tantek see http://microformats.org/wiki/microformats2-parsing-brainstorming#Nested_h-.2A_objects.27_.22value.22_property
#
tantek ben_thatmustbeme: to answer your question, we may end up allowing either and then expanding the authorship algorithm to handle both
#
tantek which would then put a slight preference on u-author since by getting the author's URL you can likely get more information than just their name
#
ben_thatmustbeme so the non-technical answer is really what data you are focusing on as the primary item
#
aaronpk at some point i need to see some sort of written sumary about what's changed in microformats parsing
#
aaronpk hard to follow the IRC chatter
dym_cx, eschnou, tantek, KevinMarks_ and kez joined the channel
#
hendrick edited /hcalendar-authoring (+198) "/* Related Pages */" (view diff)
kez, Zegnat, ChiefRA, csarven, eschnou, chiui, pfefferle, pfefferle_, KevinMarks__, Left_Turn and KevinMarks_ joined the channel
#
csarven Dear LazyMF , Are @rel self and bookmark changed in mf2?
adactio and KevinMarks_ joined the channel
#
@wpscouts Author hReview : Ajouter une Evaluation étoilée á WordPress et des Témoignages d'Utilisateurs http://wpscouts.com/author-hreview/ via @wpscouts (twitter.com/_/status/608575044797739008)
pfefferle, KevinMarks__ and glennjones joined the channel
#
KevinMarks csarven: you mean rel=me ?
#
KevinMarks rel=self is a weird Atom thing
glennjones, eschnou and pfefferle joined the channel
#
@hirameki (掘り返し) h：1130 ひらめいったーをAutoPagerizeのMicroformatsに対応させる。させろ from id:fuba http://ryogrid.net/idea/twit/1130 (twitter.com/_/status/608596980428738563)
#
csarven KevinMarks IIRC from mf1, "self bookmark" was used as a permalink
#
csarven I suppose in mf2, its equivalent is u-url
#
csarven Yeap: http://microformats.org/wiki/hEntry#What_about_rel_bookmark
netweb, glennjones, KartikPrabhu, pfefferle, TallTed and eschnou joined the channel
#
ben_thatmustbeme do any implementations of the parser support the p-audio yet?
#
ben_thatmustbeme or rather u-audio
#
kylewm csarven: rel-self is an Atom thing, not part of microformats
#
kylewm rel-bookmark is equivalent of u-url, yes, though there is proposed backcompat parsing for it (supported in mf2py only right afaik)
#
kylewm right now* afaik
#
kylewm (oh sorry, I didn't see KevinMarks already replied about re-self)
#
csarven I was talking about self *and* bookmark together. It is about the current document's permalink.
#
csarven "bookmark external" would be for.. external links
#
kylewm there's no rel-self defined for html at all
#
kylewm with or without bookmark
eschnou, pfefferle and fuzzyhorns joined the channel
#
csarven kylewm OKie.. then I'm mistaken. I don't remember why I had self in the first place. Probably pre-2007 stuff.
#
kylewm csarven: possibly for Pubsubhubbub?
#
csarven I really can't remember. No big deal :)
KevinMarks_ and KartikPrabhu joined the channel
#
kylewm KevinMarks_: hullo, seeing a different encoding problem now. when the test case doesn't define a charset, BS4 guesses, and in the case of Tantek Çelik it guesses windows-1252
#
kylewm "If you happen to know a document’s encoding ahead of time, you can avoid mistakes and delays by passing it to the BeautifulSoup constructor as from_encoding."
#
kylewm do you think we should pass in utf-8 always? maybe just for all tests?
KartikPrabhu and KevinMarks__ joined the channel
#
KevinMarks__ Will that override actual encoding? There are lots of sites that use non utf8 encodings, especially Chinese and Japanese
#
KevinMarks__ If it just changes default that's good.
#
kylewm do you think we should pass in utf-8 always? maybe just for all tests?
#
KevinMarks__ The Unicode dammit stuff tried to deal with guessing encoding
#
kylewm blargh
#
kylewm no, it overrides
#
kylewm yeah but it guesses wrong on these short samples
#
kylewm even with chardet installed, it guessed wrong for t's name
#
kylewm >>> soup.original_encoding
#
kylewm 'ISO-8859-2'
#
KevinMarks__ We could just add encoding to the test
#
kylewm yeah that's what i will propose
#
kylewm <meta charset="utf-8"> fixes in
#
KevinMarks__ I mean the meta tag for it, not a change to the test runner
#
KevinMarks__ Yes
#
kylewm collapsing whitespace and ignoring extraneous keys didn't reduce the number of failing tests as much as I'd hoped
#
KevinMarks__ The real scary case is when the encoding varies within a pagw
#
KevinMarks__ This used to be a big problem with blogs copy and pasting quotes
#
KevinMarks__ And if they had a windows nonbreaking space char in 0x80 the utf8 decoder would throw an exception
#
kylewm UnicodeDammit.detwingle :)
ben_thatmustbeme, KevinMarks_ and tantek joined the channel
#
ben_thatmustbeme so in going through as2 there are a couple of things that i think make sense, but just had questions on
#
ben_thatmustbeme first was supporting multiple sources for audio / video / picture tags
#
tantek depends on *why* you're publishing multiple sources
#
ben_thatmustbeme right now they would all get clumped under just "audio" correct?
#
ben_thatmustbeme well for video its about browser support
#
ben_thatmustbeme a reader would need both if they want to re-include
#
tantek if you can answer that question, then we can look at the proper HTML markup, whether it's different formats for different UA support, or different resolutions for devices, or different sizes for bandwidths
#
tantek nope, for video it's all 3
#
tantek and that's where we start asking/looking for real world publishing examples of multiple sources of audio / video / picture
#
tantek to determine how examples should mark them up
#
tantek or rather, to determine *what* specific problems the examples should show solutions for
#
tantek rather than attempting to solve an arbitrary m x n x o cubespace of possibilities
#
ben_thatmustbeme this is where I was testing pin13's support
#
ben_thatmustbeme https://ben.thatmustbe.me/static/test.html
#
ben_thatmustbeme more of told me that it doesn't support u-audio grabbing src= value
#
tantek more of?
#
ben_thatmustbeme s/more of told me that it/actually it just confirmed that pin13s parser/
#
Loqi ben_thatmustbeme meant to say: actually it just confirmed that pin13s parser doesn't support u-audio grabbing src= value
#
ben_thatmustbeme http://pin13.net/mf2/?url=https%3A%2F%2Fben.thatmustbe.me%2Fstatic%2Ftest.html
#
KevinMarks_ Looks like mf2py isn't normalising urls http://www.unmung.com/mf2?url=https%3A%2F%2Fben.thatmustbe.me%2Fstatic%2Ftest.html&html=&pretty=on
#
ben_thatmustbeme but that shows properly parsed values. the question is how to group the multiple audio files which are just alternates
#
ben_thatmustbeme that looks like 4 seperate files in the parsed version
KevinMarks__ joined the channel
#
kylewm KevinMarks_: it normalises them for me https://kylewm.com/services/mf2?url=https%3A%2F%2Fben.thatmustbe.me%2Fstatic%2Ftest.html
#
ben_thatmustbeme added a p-name too so it has something better in there
#
kylewm possibly you aren't passing url to the parser?
#
tantek ben_thatmustbeme: "just alternates" is ambiguous
#
tantek alternates in what dimension(s)?
#
ben_thatmustbeme alternate encodings of the same video content, they are grouped in an <audio> tag which is intended for that
#
tantek is anyone publishing that?
#
ben_thatmustbeme audio tag i am already using though i don't rewrite code for multiple versions yet
#
ben_thatmustbeme in indieweb not sure anyone is, outside of indieweb it shouldn't be too hard to find some samples
#
KevinMarks__ It's less common with audio, as mp3 is ubiquitous. Rarer to see m4a these days
#
KevinMarks__ Also rarer to see different quality choices for audio based on bandwidth
#
aaronpk silo example of different encodings of the same video content: youtube
#
tantek right this is my point - we need examples of multiple sources on real world sites
#
tantek in order to come up with *practical* examples for such
#
tantek in my experience I've seen more JS randomness for doing weird format detection stuff
#
tantek than use of multie <source> audio or video :/
#
tantek hence the challenge to actually show a real world example with *markup* for multiple sources
#
aaronpk i actually used to publish multiple encodings of my videos on my site, but can't find the examples anymore
#
aaronpk but that was pre-HD days, so it was like 320p vs 480p
#
ben_thatmustbeme I only do one right now because i haven't built that part out yet. Wasn't going to bother until i fixed up mobilepub
#
tantek people also do weird UA-detection stuff on the server and then only serve one source etc.
#
tantek point is - I'd say ditch the multi-source examples for now
#
tantek rather, ditch the *theoretical* multi-source examples, and if anyone objects, ask for the same kinds of real world example citations that we're asking for above
#
aaronpk as a publisher, I would much rather publish only full 1080p HD because rendering multiple formats is annoying and takes a long time
#
ben_thatmustbeme agreed, transcoding is annoying
#
ben_thatmustbeme though it can be pretty important for mobile
#
ben_thatmustbeme ideally the format wars can just be relegated to "in the future browser makers will finally just support 1"
KevinMarks_ joined the channel
#
tantek ben_thatmustbeme: agreed that in theory it can be pretty important for mobile - but if it is actually important, then finding real world examples, and thus documenting them, should be easy
#
aaronpk that's becoming less and less true tho, because my phone now gets faster internet than my office
#
ben_thatmustbeme aaronpk: same here, haha, we have DSL in our office
#
ben_thatmustbeme so i can accept that for now, a single <source> is sufficient
#
ben_thatmustbeme until need arises
#
ben_thatmustbeme my second question was on supporting language of an h-entry
#
ben_thatmustbeme though i know tantek already said (other channel) that it can be difficult to convince publishers to include such information
KevinMarks__ joined the channel
#
ben_thatmustbeme My last question thus far will likely get relagated to "real world examples" which is what I will be working on soon. Its on having some way to convey a "type" of h-entry or action taken
#
ben_thatmustbeme I will be publishing any activity i do, as an h-feed
#
tantek ben_thatmustbeme: yes, my point is that there must be specific incentive for authors to do something before they will, and do it right (keep it up to date)
#
ben_thatmustbeme but while posts you can determine things like "like" from the existance of u-like-of, ways to determine create of a post vs edit of a post are not so clear
#
tantek without such incentive, it doesn't matter what others (parser devs, readers, etc.) *want*, the authors will either ignore them, or worse, copy / paste from a template and get it wrong, or neglect it and let it get out of sync
#
ben_thatmustbeme unless there is a bunch of u-edit / u-create etc
#
tantek huh? topic switch? confused
#
ben_thatmustbeme sorry, moved off the second question as it didn't seem to be getting a response
#
tantek sorry, I'm on the CSS telcon so trying to multitask ;)
#
ben_thatmustbeme and the language thing was just more of a curiousity
#
tantek anyway, historically, lang attribute has been pretty crappy
#
tantek especially lang=en, due to aforementioned template/copy/paste problem
#
tantek it's the simple proof that authors won't care or will get wrong, things they don't have sufficient incentive to get right
#
ben_thatmustbeme ahh true, a lot of peopel jsut copy paste that
#
tantek therefore it is better to *not* ask them to do something, then to ask them, and have them get it wrong and provide *noise*
#
tantek something which lots of wishful thinkers continue to get wrong in terms of asking authors to do extra work
#
ben_thatmustbeme switching to other question. I will be publishing data on creating / editing / deleting posts. an h-feed of things done, not actual posts.
#
ben_thatmustbeme i already do this with incoming events (though i have yet to have anyone edit a reply or anything)
#
tantek edits are actual posts
#
tantek also - this is probably more of an #indiewebcamp topic so let's take it there
gRegorLove joined the channel
#
ben_thatmustbeme okay
#
KevinMarks__ Language of an entry I'd look at Stephanie Booth for examples
#
tantek KevinMarks: could you explain author incentive though?
#
KevinMarks__ http://climbtothestars.org/wordpress/basic-bilingual/
#
KevinMarks__ Steph posts in English and French, and posts a summary in the other language
#
KevinMarks__ And wrote a WordPress plugin for other people to use too
KevinMarks_ joined the channel
#
KevinMarks_ She uses both visible [fr] [en] markup and lang on the hentry
chiui joined the channel
#
tantek KevinMarks - but why? what's the incentive to do so besides markup geekery?
#
tantek what code consumes that markup and does anything with it?
#
tantek says as a fellow markup geek
KevinMarks__ joined the channel
#
KevinMarks__ Steph's deeper point is about perception of language use
#
KevinMarks__ We are all imperfectly multilingual in different ways
#
KevinMarks__ Assuming that a text has a single language is an oversimplification
#
KevinMarks__ Which Google insists on, to the point of telling you not to use more than one language as they treat it as a bug
#
KevinMarks__ https://support.google.com/webmasters/answer/182192?hl=en
#
tantek wow
#
tantek classic programmer-think: your natural behavior is too complex for our code, change your natural behavior to conform to our limited world model
#
KevinMarks__ They do use rel=alternate hreflang
#
tantek that's got to be an existing anti-pattern defined somewhere
#
KevinMarks__ https://support.google.com/webmasters/answer/189077?hl=en
#
tantek programmer blinders or something
#
tantek could you document/cite that on microformats.org/wiki/rel-alternate
#
tantek ?
#
KevinMarks__ Stephanie gave a tech talk about it at Google (I invited her to)
#
KevinMarks__ Also programmers assuming there is a simple country:language mapping
#
aaronpk ah if only it were that simple
#
KevinMarks__ Stephanie lives in Switzerland, the French speaking bit
#
KevinMarks__ Many us sites assume that their .ch site should be in German
KevinMarks_ and elf-pavlik joined the channel
#
KevinMarks the hreflang is already documented there
#
tantek is the google citation already there?
#
tantek the support.google citation that is
#
csarven KevinMarks I'm in Bern, CH (mostly German in this Canton). https://developers.google.com/structured-data/testing-tool/ displays content in English (outer) and German (inner).
#
KevinMarks Google thinks that they can do better by algorithm than by explicit setting
#
KevinMarks and a lto of the time they are right
#
KevinMarks but some of the time they have misdefined the problem
#
KevinMarks assuming things have a single language is one of those times
KevinMarks__, chiui, dym_cx, glennjones, KevinMarks_, eschnou, elf-pavlik, iwaim, Erkan_Yilmaz and fuzzyhorns joined the channel