#microformats 2013-09-17

2013-09-17 UTC
tantek joined the channel
#
tantek edited /h-event (+77) "fix typo, add subsections, move processing tip for iCalendar in backcompat section" (view diff)
nonge, shaners, chiui, xid_, tantek and Loqi joined the channel
#
tantek there he is
#
tantek hey barnabywalters - just had a discussion about solving the e-content problem
#
tantek updated proposal in the Etherpad
#
gjones yes your right you do need always check for string vs object
#
tantek barnabywalters - look for "Proposal 2 (2013-260)"
#
tantek in https://etherpad.mozilla.org/microformats2parsing
#
gjones I like the solution because the type is implied by the structure rather than through a naming conversion on the properties, seems a better fit to the current JSON design
#
tantek agreed
#
tantek we have to be very careful with naming conventions on the properties
#
tantek we only have naming conventions on the properties when they're in the HTML
#
tantek and when there - those naming conventions only mean specific parsing instructions
#
tantek by the time the properties make it to the JSON output - there is no more naming convention - and that's deliberate
#
tantek ok I'm going to leave the solution in the Etherpad for review
#
tantek before adding it to the wiki
#
tantek gjones I added an O
#
tantek Opinions: section underneath proposal 2
#
tantek if you're happy with it, please +1 it
#
tantek and same with you too barnabywalters and tommorris when you get a chance to review it
#
@akokubo 名刺みたいなデータのマイクロフォーマットのhCardをユーザーアカウントの設計の参考にしようかと思ったけど、仕様が細かくてやり過ぎな感じが…。 https://developer.mozilla.org/ja/docs/The_hCard_microformat http://en.wikipedia.org/wiki/VCard#hCard_1.0 (twitter.com/_/status/379919461366513665)
#
tantek cool
#
gjones The last bit you added about class="p-content e-content" will be a real pain. Processing rule like this meaning holding the type of each property found in memory and post processing to remove duplicates.
#
tantek added some mediawiki headings markup to the etherpad to hopefully make it easier to navigate (and eventually copy to the wiki and maintain structure)
#
tantek gjones - should it produce a duplicate instead?
#
tantek would that be easier?
#
barnabywalters morning all
#
barnabywalters +1 on tanteks suggestion
#
barnabywalters +1 for gjones’s complaint
#
tantek barnabywalters - add it to the etherpad :)
#
barnabywalters that resolves the issues I had with the naming convention approach
#
tantek barnabywalters - yes it was a very serious issue - we discussed it a lot during the meetup
#
tantek barnabywalters - what should happen with class="p-content e-content" then?
#
barnabywalters tantek: the same thing which would happen if you used multiple copies of the same property name with different prefixes anywhere else
#
barnabywalters producing a duplicate
#
tantek oh that's interesting, there's no explicit definition of what happens for say "p-foo u-foo"
#
barnabywalters then consumers search the property arrays for a hash if they’re expecting there to be e- content
#
tantek you're saying each produces a value for "foo" which seems reasonable
#
tantek ok
#
tantek I'm going to take that part out
#
barnabywalters okay, that’s a +1 overall from me then
#
tantek good catch gjones and barnabywalters
#
tantek does the new example make sense?
#
tantek for "p-content e-content"
#
barnabywalters makes sense to me
#
tantek ok cool
#
tantek If that looks good to you guys, go ahead and code and I'll wikify
#
gjones yes the output make sense to
#
gjones I am working on the parser at the moment, I will add this to my todo list and let you known when I have a working example
#
barnabywalters raising as issue on php-mf2
#
barnabywalters https://github.com/indieweb/php-mf2/issues/31
#
tantek thanks guys
pfefferle, eschnou and tantek joined the channel
#
gjones barnabywalters quick question, when you parse http://tantek.com/ how come your parse does not show one h-feed inside another h-feed?
#
gjones there is hfeed on the body and a h-feed on the ol within the body
#
barnabywalters I don’t convert old classnames to new ones by default
#
barnabywalters I’ve actually seen this problem arise a few times — people add mf2 markup, and then add legacy markup in a completely different place
#
barnabywalters in at least one example it was because of the advice tantek gave in a talk to put mf2 markup in your page and then add a legacy classname to the body element
#
gjones Thats it then, I think tantek needs to move the class together on to one element
#
barnabywalters the other thing I was considering doing to try to avoid such problems was to detect whether or not there is any mf2 markup, and if there is never perform conversion
#
barnabywalters yeah
tobyink joined the channel
#
gjones That’s a bit of a messy problem, we need to point out this carefully when telling people to mix and match v1 and v2 markup
#
barnabywalters yeah, thinking about it we probably need to solve it at parser level
#
barnabywalters in fact, isn’t it already mentioned in the parsing spec?
tantek joined the channel
#
gjones I will have a quick look
#
barnabywalters so the parsing spec says things like “If an "h-entry" is found, don't look for an "hentry" on the same element. ”
#
barnabywalters which *doesn’t* solve this particular problem
#
hieule edited /start-simple (+6) (view diff)
#
barnabywalters of legacy classnames and mf2 classnames being on different elements
pfefferle joined the channel
#
gjones Yes I am already doing this but only if the two classes are on one element
#
hieule edited /start-simple (-1) "/* solve simpler problems first */" (view diff)
tantek_ and tom_thumb joined the channel
#
@StepUpWeb hCard — формат, предназначенный для публикации контактной информации http://www.step-up-web.ru/hcard/ (twitter.com/_/status/379956852458414081)
#
tantek edited /start-simple (-5) "Reverted edits by [[Special:Contributions/Hieule|Hieule]] ([[User talk:Hieule|Talk]]) to last version by [[User:Tantek|Tantek]]" (view diff)
TallTed joined the channel
#
tantek edited /simple (+5) "update redirect" (view diff)
#
tommorris apologises for his failure in breaking the wiki. :(
#
tommorris tantek: can you email me the LocalSettings and I'll have a look
tantek joined the channel
#
tommorris tantek: will have a look at the LocalSettings shortly
#
tantek thanks tommorris
jschweinsberg joined the channel
#
tommorris tantek: found the problem. Am making a few other changes too.
#
tantek thanks tommorris - something obvious?
#
tommorris you just need to be explicit about permissions with wgGroupPermission
#
tommorris tantek: sent.
#
tantek tommorris - I agree with autopatrolling admin edits, but not autoconfirmed
#
tantek because the bar for autoconfirmed is still pretty low
#
tantek edited /start-simple (+3) "remove use of trim shortener" (view diff)
#
tantek edited /start-simple (+23) "entry-title" (view diff)
#
tantek edited /tr.im (+53) "add simple - start-simple" (view diff)
#
tantek tommorris - remainder of edits made and I can create pages again
#
tommorris tantek: admin edits are autopatrolled by default. but best make it explicit in the file anyway. ;)
eschnou joined the channel
#
tantek ok
chiui, shaners, tobyink and tantek joined the channel
#
JonathanNeal How are international numbers handled in microformats, or are they? e.g. https://gist.github.com/jonathantneal/bb09f665b07d9952f509
#
tantek edited /h-entry (+592) "add p-location re-use from h-event, drop p-geo p-latitude p-longitude p-altitude as no one has ever used them directly inside an h-entry nor hentry. use nested h-geo on a p-location instead. add FAQs." (view diff)
#
tantek edited /h-event (-248) "drop p-geo p-latitude p-longitude p-altitude as no one has ever used them directly inside an h-event nor vevent. use nested h-geo inside p-location instead." (view diff)
tantek and xid joined the channel