#microformats 2018-03-18

2018-03-18 UTC
#
KartikPrabhu
mf2py already adds elements to fudge the rel-category to p-category parsing
#
KartikPrabhu
ok now I am very confused
#
Loqi
h-review
#
KartikPrabhu
is that not equivalent to just replacing the "fn" class with "p-item h-item p-name" ?
#
KartikPrabhu
or is it "fn" is parsed as a p-name but the resulting value goes into the p-item > h-item > p-name nested property?
#
Zegnat
<span class="fn">Thing being reviewed</span> → <span class="p-item h-item"><span class="p-name">Thing being reviewed</span></span>
#
Zegnat
Is how I read that
#
tantek
Zegnat is correct
#
KartikPrabhu
<sigh> this is going to be very complicated
#
tantek
KartikPrabhu: note that class names are unordered in a class attribute
#
KartikPrabhu
tantek: yes that is what alerted me
#
tantek
therefore your example of "p-item h-item p-name" is identical in meaning to "p-item p-name h-item"
#
gRegorLove
Correct, otherwise you'd end up with something like this: http://pin13.net/mf2/?id=20180318004531321
#
tantek
which may help better explain how it works
#
KartikPrabhu
gRegorLove: yeah mf2py simply replaces the class now too
#
KartikPrabhu
tantek: would be great if backcompats had examples of expected output
#
KartikPrabhu
tantek: Zegnat: however multiple properties "fn" and "url" all map to the same "h-item" right?
#
tantek
KartikPrabhu: TBH a lot of that is for backward compat with very old examples
#
tantek
so it is likely worth re-researching actual usage
#
KartikPrabhu
ok hreview is the only very complicated one so would be good to simplify it
#
Zegnat
Hard to say without looking at old usage, KartikPrabhu. If there is presedence for multiple fn’s meaning multiple items, then no. But I would think multiple items just hard require the `item` class to have been used in the mf1.
#
KartikPrabhu
for example what happens to <span class="fn">Thing being reviewed</span> <a class="url" href="example.com">URL</a> ?
#
KartikPrabhu
inside hreview
#
Zegnat
So probably safe to put all fn and url that are direct descendents of the hreview on the same item
#
tantek
hReview both in theory and practice only allowed one item
#
tantek
so before you work too hard on backcompat for that, might try going through /hreview-examples-in-wild
#
Zegnat
Looking at http://microformats.org/wiki/hReview fn was never supposed to be a direct child to hreview, always needed an item in between.
#
Loqi
hReview 0.4 (in progress)
#
gRegorLove
KartikPrabhu, did you find this bug via the test cases, or some real hreview example?
#
KartikPrabhu
this is going to be very very messy for both mf2py and php-mf2
#
KartikPrabhu
and hreview is the only special case I think
#
KartikPrabhu
anyway gtg will catch up on logs later
#
tantek
my guess is that most / many hReview in the wild are going to be from that WordPress plugin that gets tweeted like mad
#
Zegnat
Have to leave too, soon to be 2am and my bed is calling
#
gRegorLove
Ah, interesting catch, Zegnat.
#
gRegorLove
G'night!
#
KartikPrabhu
gRegorLove: writing the backcompat-rules. I am much more happy if real examples actually simplify this
#
gRegorLove
"However, when using item info subproperties ("fn", "url", "photo"), they must be nested inside the item element." http://microformats.org/wiki/hReview##However,+when+using
#
Loqi
hReview 0.4 (in progress)
Garbee joined the channel
#
gRegorLove
Looks like the h-review compatibility section started with the fn, url, photo properties but not 'item', so maybe an oversight that has stayed in that section. http://microformats.org/wiki/index.php?title=h-review&diff=prev&oldid=54814#Parser_Compatibility
#
Loqi
h-review
#
gRegorLove
I *think* we can remove fn, photo, and url from that compatibility section, leaving the item.fn, item.photo, and item.url properties.
#
gRegorLove
Does that sound right, tantek?
tantek joined the channel
#
gRegorLove
Final one for me tonight: I think the rel-tag backcompat text here: http://microformats.org/wiki/h-review#Parser_Compatibility needs to be updated to match hentry's: http://microformats.org/wiki/h-entry#Parser_Compatibility
#
Loqi
h-review
#
tantek
likely
#
tantek
gRegorLove: there's mostly hReview 0.3 out there (published) which had direct uses of fn url photo
#
tantek
without "item"
#
tantek
hence why I suggested you check real world examples before attempting to simplify or otherwise re-write / update any of the backcompat
#
tantek
at this point, who knows how real world examples have aged, and they should be the source of informing what we need backcompat for
#
gRegorLove
I hadn't noticed the version numbers on them previously. Interesting.
#
tantek
we don't need to invent new simplified backcompat for anyone because no one will *start* publishing backcompat
#
tantek
the only backcompat we need for hReview is the minimum necessary to capture actual real world examples
#
tantek
so before you propose removing / adding / changing anything - check actual published examples
#
gRegorLove
Checking now
#
gRegorLove
Hah http://bordeaux.pikadom.fr/ "Apache is functioning normally"
#
tantek
yeah, go ahead and unlink (with <nowiki>) any dead examples and move them to a "Historical Examples" section
#
tantek
they were all valid as of when they were added
#
tantek
(rather than deleting)
#
gregorlove
edited /hreview-examples-in-wild (+317) "starting Historical Examples section"
(view diff)
#
gregorlove
edited /hreview-examples-in-wild (+262) "more historical examples, +nowiki"
(view diff)
#
gregorlove
edited /hreview-examples-in-wild (+87) "confirmed Hardcore Epoxy has hreview"
(view diff)
[kevinmarks] joined the channel
#
[kevinmarks]
Maybe link to Internet archive for when the example was valid?
#
gRegorLove
Good idea
#
gregorlove
edited /hreview-examples-in-wild (+460) "more historical examples, internet archive link"
(view diff)
#
gRegorLove
Oh yeah, we're going to need to do something about the ReCAPTCHA version on the wiki. v1 API shuts down March 31.
tantek joined the channel
#
gregorlove
edited /hreview-examples-in-wild (+971) "/* Historical Examples */ Internet Archive links"
(view diff)
tantek joined the channel
#
gregorlove
edited /hreview-examples-in-wild (+343) "mv more historical examples"
(view diff)
tantek joined the channel
#
gregorlove
edited /hreview-examples-in-wild (+1071) "more reviewed/historical examples"
(view diff)
[eddie], nitot, voxpelli, sebsel, [gerwitz], barpthewire, Garbee, [kevinmarks] and nitot_ joined the channel
#
sknebel
^^^ spam
barpthewire and [eddie] joined the channel
#
Loqi
[kartikprabhu] Proposal for properties that need to be nested in other things for backcompat maybe use the following example from `hreview["url"]` ## old format ```json "url": [ "p-item", "h-item", "u-url" ...
[snarfed], barpthewire and [eddie] joined the channel
#
@cote
Man, then Web 2.0 chickens sure is coming home to roost. If only we’d just stick to RDF and microformats.
(twitter.com/_/status/975438307709636611)
[pfefferle] and [cleverdevil] joined the channel
#
gRegorLove
KartikPrabhu: tantek is suggesting we verify there's current examples of hreview published that way (without the 'item' intervening). If not, we probably don't need to worry about it.
#
Loqi
hReview Examples in the wild
[eddie] joined the channel
#
KartikPrabhu
gRegorLove: awesome! let me know what you find and I can update the backcompat-rules accordingly
[kevinmarks], KartikPrabhu, [cleverdevil] and Stickers666 joined the channel