#microformats 2017-03-06

2017-03-06 UTC
[dgold] and nitot joined the channel
#
gRegorLove Backcompat question for: http://pin13.net/mf2/?id=20170306004752354
#
gRegorLove Per http://microformats.org/wiki/h-adr#Backward_Compatiblity "If an "h-adr" is found, don't look for an "adr" on the same element."
#
Loqi h-adr
#
gRegorLove I think that means the h-adr should show up as a child, not as a property of the h-card. Looking more like: https://gregorlove.com/php-mf2/test.php
#
aaronpk can't we just simplify the backcompat rules and say if there are any mf2 on the page then don't do any mf1 backcompat?
#
aaronpk IMO this is all very messy and doesn't actually provide useful results
#
gRegorLove Following the backcompat algorithm (in the work I'm doing now) is actually giving much better results than latest php-mf2
#
aaronpk the useful backcompat stuff is when a publisher does not have mf2 but still has mf1 like reviews or recipes
#
gRegorLove Yes, that's the things I'm working on.
#
aaronpk i just don't see the use in trying to solve mixed mf2+mf1
#
gRegorLove The current approach is "replace all mf1 class names with their mf2 versions" which causes the messes we've run into. Actually following the algorithm, taking into account whether the parent element has an mf2, works much better.
#
aaronpk well that's good at least
#
gRegorLove Note in my link there that the <div class="locality">MF1</div> didn't get upgraded to p-locality
#
gRegorLove Because it's within an mf2 element
#
gRegorLove So my current understanding of the algorithm seems to contradict http://microformats.org/wiki/microformats2-parsing-issues#uf2_children_on_backcompat_properties, which is why I'm asking with that specific example
#
Loqi microformats2-parsing-issues
#
gRegorLove If that was just a <div class="adr"> then the improved backcompat would make it <div class="adr p-adr h-adr"> and it would parse as a property, not child
#
aaronpk huh i see what you mean
#
aaronpk it looks like the intent is to have that h-card be the value of the "adr" property
nitot joined the channel
#
gRegorLove Agreed. I think http://microformats.org/wiki/microformats2-parsing#algorithmi is more up-to-date
#
Loqi [Tantek Çelik] microformats2 parsing specification
#
gRegorLove "if that child element itself has a microformat ("h-*" or backcompat roots) and is a property element, add it into the array of values for that property..."
#
gRegorLove If there's any hReview examples that php-mf2 is handling badly now, let me know and I'll try them with this. So far using old examples on the wiki
#
aaronpk here's one https://github.com/indieweb/php-mf2/issues/107
#
gRegorLove Oh yeah :)
#
aaronpk it's a complete disaster lol
#
aaronpk https://w7apk.com/kenwood-th-d74a
#
Loqi [Aaron Parecki (W7APK)] Kenwood TH-D74a 144/220/430 MHz APRS D-Star Handheld Transceiver
#
gRegorLove I used that in earlier tests, but a lot has changed since. Let me try again.
#
gRegorLove https://gregorlove.com/php-mf2/test.php?id=2
#
Loqi [Aaron Parecki (W7APK)] An amazing radio with tons of features, but battery life is unacceptable.
#
aaronpk much better!
#
aaronpk isn't the name property there supposed to be summary?
#
gRegorLove summary parses as p-name, according to h-review backcompat. Is that what you mean?
#
aaronpk that's odd. why?
#
aaronpk i would have thought summary would be summary
#
gRegorLove Not sure. http://microformats.org/wiki/h-review#Backward_Compatiblity
#
Loqi h-review
#
aaronpk huh
#
gRegorLove Maybe an oversight. I don't see p-summary
#
gRegorLove Other than that is anything wrong?
#
aaronpk oh... there is no "name" in hReview
#
aaronpk that's why
#
aaronpk and everything in mf2 has to have a name
#
gRegorLove Ah
#
aaronpk hm i think then it should actually put it in both mf2 name and summary
#
aaronpk that looks correct though!
#
aaronpk wanna try one with mixed mf1 and mf2? https://aaronparecki.com/2016/12/15/11/bose-solo-15-sound-system-review
#
Loqi [Aaron Parecki] Bose Solo 15 Sound System
#
gRegorLove Easy to add. Makes sense to me, though maybe p-summary should be added to the properties for h-review
#
aaronpk yes, let's open an issue on h-review then
#
gRegorLove Refresh the id=2 test above; has p-summary now
#
aaronpk cool
#
gRegorLove Mixed: https://gregorlove.com/php-mf2/test.php?id=3
#
gRegorLove Hm
#
aaronpk it's definitely an improvement
#
gRegorLove The h-review itself looks good. Some other issues though
#
gRegorLove Ah, <span class="reviewer vcard"> won't be upgraded because it's within an mf2 h-* root
#
aaronpk i guess i could put p-author h-card on that
nitot, KevinMarks and [kevinmarks] joined the channel
#
[kevinmarks] How does unmung.com do with these examples?
#
gRegorLove afaict, mf2py backcompat is based on php-mf2
#
gRegorLove I don't know python super well, but it looks very similar
KevinMarks joined the channel
#
[kevinmarks] Hm. I should look at the test cases again, I did spend some time with them a while back.
nitot, KevinMarks_, KevinMarks, tantek, timo, [dgold], nitot_, pfefferle, GK_1wm___SU and [eddie] joined the channel
#
@bloggermin Blogger SEO: Microformats http://www.bloggermin.cashmore.com.br/2012/07/seo-blogger-microformats.html https://twitter.com/bloggermin/status/838781356406616064/photo/1 (twitter.com/_/status/838781356406616064)
[kevinmarks], nitot, KevinMarks and [keithjgrant] joined the channel
#
@rbairwell Just realising what a rabbit hole Microformats/microdata is.... Looks like the next few days are going to be "interesting".... (twitter.com/_/status/838801868541345792)
tantek, bwt, gRegorLove and nitot joined the channel
#
@SebastianX I admit that ranting against totally useless microformats like no-follow & Obama's failed foreign policy is a threat to national security. (twitter.com/_/status/838830497665011713)
#
tantek wat
#
aaronpk markov bot?
[eddie] joined the channel
#
[eddie] like the store?
#
[eddie] lol the really funny thing is it tweeted a thread. One of the tweets in the thread says "I'll just ignore a failed banana republic."
#
[eddie] lol
#
sknebel pretty sure the term is more than half a century older than the store
#
gRegorLove Hey tantek, could you take a look at yesterday's convo and let me know your thoughts: https://chat.indieweb.org/microformats/2017-03-06#t1488761293131000
#
[eddie] Oh, interesting. I’ve actually never heard of that as a term outside of a noun for the store
#
sknebel it's a degratory term for a state dominated by corruption supporting powerful interests (originally, banana export companies basically ruling latin american states)
#
sknebel but we really are getting off-topic
#
[eddie] We are, so we’ll let the convo die. But thanks for the term expansion. Always helpful to learn new things ?
#
tantek re: yesterday and back compat
#
tantek that?
#
tantek not sure which exactly after reading the thread
#
tantek I think the key use-cases are:
#
tantek 1. sites that haven't been updated with mf2 yet (e.g. lots of WordPress hentry out there, let's make that work "out of the box" with mf2 parsers)
#
tantek 2. sites that MUST keep some classic mf for search engines etc. at least for the root object, but want to add more details with mf2 inside, WITHOUT losing the root object, and maybe they can't update the root object (WordPress, again, because class="hentry" is in core, not themes)
#
tantek 3. sites that use a classic mf + mf2 on the same root object, then only mf2 inside (this should "just work" now across parsers)
#
tantek aaronpk, gRegorLove does that make sense to you (especially 1 & 2)
#
aaronpk re: 2, I think we're on the same page, but just want to clarify that "want to add more details with mf2 inside" does not preclude adding "details" with mf1
#
aaronpk my main use case for that being I want my reviews to show up in google with the little stars but i also want them to have mf2
#
tantek right now that is the case yes, because there are WordPress pages with multiple hentry objects on them with properties and such, and we want to keep supporting that
#
tantek your main use case for reviews sounds like 3
#
aaronpk "then only mf2 inside"?
#
tantek right, the only child mf objects should be mf2
#
aaronpk oh... child *objects*
#
tantek unless I'm missing something, that should work fine for Rich Snippets (the little stars)
#
tantek as in a h-something inside an h-root of some sort
#
gRegorLove tantek: my main question came from this line on h-adr: If an "h-adr" is found, don't look for an "adr" on the same element.
#
tantek gRegorLove: that's the same as nearly every other vocab spec backcompat
#
tantek so not specific to adr
#
gRegorLove Right
#
tantek e.g. h-card and h-entry same the same thing
#
gRegorLove Just citing it for this specific example.
#
tantek so what's the question that applies across them
#
gRegorLove It seems to conflict with the microformats2-parsing-issues example I linked, where <div class="adr h-adr"> is upgraded to <div class="adr p-adr h-adr">. The node parser does that, making the adr the property of the h-card, not a child
#
tantek ah - not quite. adr is a special case
#
tantek as a class name
#
tantek because it is *both* a property class name of hCard
#
tantek *and* it is a root class name in and of itself
#
gRegorLove Ahh
#
tantek "geo" is similar I think
#
tantek yup: http://microformats.org/wiki/Main_Page#Classic_Microformats
#
Loqi Welcome to the microformats wiki!
#
gRegorLove Thanks, Loqi!
#
tantek those are the only two classic microformats that do not start with "h*"
#
Loqi you're welcome, gRegorLove
#
tantek and also happen to be properties of hCard
#
gRegorLove Is it correct to say then: If you're inside a mf1 root class, 'adr' should be considered a property for backcompat and upgraded to 'adr p-adr'?
#
gRegorLove I think that's correct based on http://microformats.org/wiki/microformats2-parsing##if+that+child+element+itself+has+a+microformat
#
Loqi [Tantek Çelik] microformats2 parsing specification
#
tantek gRegorLove: not quite, because we can be more restrictive for adr, since AFAIK it was only a property class name in the context of hCard
#
tantek and frankly for backcompat we should be as restrictive as possible, to minimize burden on parsers, and minimize false positives
KartikPrabhu joined the channel
#
gRegorLove I'm not sure I follow
#
gRegorLove I thought adr / h-adr could be used to mark up addresses, doesn't have to be inside an hCard / h-card
#
gRegorLove Oh, I think I get it. I misspoke. Yes, 'adr' is only treated as a property inside hCard by the backcompat code -- not inside just *any* mf1 root.
tantek joined the channel
#
@snowcloneminute In microformats, no one can hear you dissert (twitter.com/_/status/838876924663500800)
[terminalpixel], [eddie] and [dgold] joined the channel