#microformats 2012-08-27

2012-08-27 UTC
lgedeon, kennyluck, tantek and tantek_ joined the channel
#
tantek aaronpk - re: http://aaronparecki.com/2012/236/article/1/you-should-not-be-displaying-relative-dates nice post - just one nit, you linked to http://microformats.org/wiki/datetime-design-pattern but apparently ignored the big warning at the top ;)
#
tantek though I should now update that warning/advice to use the <time> element instead
#
aaronpk yes, I've been noticing it's hard to tell what is the most current information on the microformats wiki
#
tantek aaronpk - big yellow box at the top makes it hard to tell?!?
#
tantek or do you mean on other pages
#
aaronpk especially since things like this: http://microformats.org/wiki/hAtom#Entry_Published say to use the datetime-design-pattern
#
tantek I means specifically about the page you linked to
#
aaronpk well it just says "please review" and is yellow, it doesn't say "deprecated" and it's not red
#
tantek ah - ok, that's a good point, hAtom needs updating for that.
#
aaronpk I found more things mentioning the datetime-design-pattern than the value-class-pattern
#
tantek now that's a problem
#
aaronpk like i said, it's hard to tell what's the most current information
#
aaronpk there's also the problem of what the site says is the most current vs what google and other search engines actually parse
#
tantek edited /datetime-design-pattern (+298) "add explicit note/recommendation to use HTML5 time element" (view diff)
#
aaronpk ahh much better!
#
tantek edited /datetime-design-pattern (-9) "or a DRYer solution" (view diff)
#
tantek the one thing that bugs me about <time> is the DRY violation it so often requires
#
aaronpk I suppose
#
tantek it bugged me enough that we have a VCP+<time> solution that avoids DRY violation :)
#
tantek my blog uses it
#
tantek in short, you can do <span class="dtpublished"><time class="value">20:40</time> on <time class="value">2012-08-24</time></span>
#
tantek which gives you full DRY principle adherence, as well as the flexibility of reordering time / date for your own readability preference
#
aaronpk I want to do a follow-up post of recommendations of how to display dates
#
aaronpk one of the things I realized is that for rapidly changing feeds of data, the relative date can be ok to use (ideally along with the absolute date), and displaying events is often clearer if you also show the day of the week
#
tantek I've even proposed to WHATWG to expand <time> to permit compositing nested time elements, e.g. the above example could be represented even briefer as <time class="dtpublished"><time>20:40</time> on <time>2012-08-24</time></time>
#
tantek http://wiki.whatwg.org/wiki/Time_element#reducing_DRY_violations_overview and following sections
#
tantek but it was rejected for now, with the point that not enough adoption of the <time> element itself has been seen yet to justify extending it further, which is a reasonable reason to reject (for now)
#
tantek til then, we can keep using vcp+<time> (as I have it running live on my site)
#
tantek now let me see if I can fix those other out-of-date wiki references you found
#
tantek was it primarily just the hAtom spec itself? or were there other pages that may have been misleading?
#
aaronpk it may have been just taht one
#
tantek well that's an important one so I should fix it ASAP
#
tantek especially as affects Indieweb deployments
#
tantek and potentially protocols we come up with (e.g. the comment / pingback thing)
#
tantek edited /value-class-pattern (+189) "/* Examples in the wild */ forgot to add tantek.com a while ago" (view diff)
#
aaronpk should proabbly go update his site to use the <time> attribute
#
tantek element ;)
#
aaronpk er, element
#
tantek :)
#
tantek oh cool - it looks like the VCP docs already handle nested <time> elements: http://microformats.org/wiki/value-class-pattern#Date_and_time_parsing
#
tantek "if the element is a del, ins, or time element, then use the element's datetime attribute value if present, otherwise its inner-text. "
#
tantek now to provide a couple of examples
#
tantek edited /value-class-pattern (+1572) "/* Date and time values */ add preferred HTML5 examples to all date and time parsing examples" (view diff)
#
tantek ok that should help a bit
#
aaronpk great examples!
#
tantek edited /hatom (+208) "replace datetime-design with value class pattern (reflect modern usage as of 2+ years ago)" (view diff)
#
tantek alright, hopefully that helps drive better usage a bit
#
tantek edited /hatom (+47) "/* Semantic XHTML Design Principles */ semantic HTML update" (view diff)
#
tantek thanks aaronpk!
#
tantek ok I think I've at least patched the misleading references in hAtom. will add better inline examples next. at least I think I've stopped the bleeding as it were.
#
aaronpk yea that looks a lot better. will take another look when I go back to redo my page tags
#
tantek awesome - thanks again for the constructive feedback - helped a lot
#
aaronpk do you know of a good tag or class to use to indicate the enclosed content is the main content of the page? I'd like to hint to crawlers where the body text lives vs navigation and footer elements for example
#
tantek well I think the way that's handled is by marking up the nav and footer elements :)
#
tantek then crawlers can focus on the other stuff in between
#
aaronpk hm that's a thought
#
tantek besides that, hentry does the trick
#
tantek or entry-content in particular
#
aaronpk actually i'm writing a crawler so I get to choose what to look at
#
tantek oh fun
#
tantek hopefully you can also try parsing for microformats2 as well
#
tantek http://microformats.org/wiki/microformats-2
#
aaronpk i'm writing a classifier for posts on my site so I can show "related posts"
#
tantek and try consuming both the uf2 on my site and on barnabywalters :)
#
aaronpk entry-content is good, I will start there
#
tantek note the use of e-entry-content in uf2: http://microformats.org/wiki/microformats-2#h-entry
#
tantek edited /microformats2 (+20) "/* h-entry */ h-entry backward compat - for 'entry-content', parse as e-" (view diff)
#
tantek in uf2, the 'e-' prefix means parse and use the element's entire innerHTML as its property value (including markup, not just text node)
#
tantek benward convinced me we needed "e-", in particular for hAtom / h-entry, which already had such "entire embedded element tree" functionality for its 'entry-content' property in order to solve the "syndicate an entry *with* markup" use-case
xtof_fr, Flam9, chiui, nonge_, romainneutron, grosroro and csarven joined the channel
#
nonge aaronpk, regarding "do you know of a good tag or class to use to indicate the enclosed content is the main content of the page?": it can be 'article' (HTML5)
#
nonge especially if there is only one 'article' not nested in other sectioning content
lgedeon joined the channel
#
cemk04 edited /existing-rel-values (+57) "/* proposals */" (view diff)
MacTed and polvo joined the channel
#
tantek rel rokbox?
#
tantek anyone heard of that?
#
tantek certainly not a proposal level rel, not even a brainstorm (no documentation provided). going to undo that.
#
tantek edited /existing-rel-values (+66) "summary for rel-payment, Undo revision 47156 by [[Special:Contributions/Cemk04|Cemk04]], no spec or summary provided for rel rokbox, sounds library-specific" (view diff)
#
tantek nonge - indeed, article can be useful for that, however article typical may include a header/footer of its own as well
#
tantek s/typical/typically
#
Loqi tantek meant to say: nonge - indeed, article can be useful for that, however article typically may include a header/footer of its own as well
#
tantek e.g. I use a <footer> inside each <article> for the date published information in my posts: tantek.com
#
tantek edited /existing-rel-values (-112) "/* formats */ rel-js-modal is not part of any existing format, and certainly not part of XFN thought that was probably copypasta. No spec provided, wrong table." (view diff)
#
nonge tantek, yeah that's right. one could argue if the header/footer inside of a 'article' belong to the content or not (I'd vote yes, the publishing date in the footer and the heading/title in the header are part of the content/article/blogpost/whatever)
chiui, romainneutron, barnabywalters, tantek, chiui_ and grosroro joined the channel
#
@phixed @davidmihm I feel like with schema/hcard/microformats/etc this just adds unnecessary confusion.
#
tantek following that tweet, here is the latest unnecessary reinvention of hCard, in a side-file too: https://github.com/fesja/businesstxt
#
tantek what's with the recent fashion of "just use a .txt at a well known local path" trend?
#
tantek e.g. humans.txt
#
tantek etc.
billyoc joined the channel
#
aaronpk I agree it's kind of a silly thing to do, but restaurant websites haven't had a good track record of using html anyway
barnabywalters joined the channel
#
tantek yeah
Cat4D joined the channel
#
@theferf @adamauden love the idea - but didn’t it already fail with hcards and microformats? http://t.co/WiEPtJ1B
romainneutron joined the channel
#
@DominicNanni It disgusts me that @MittRomney would pull the birther hcard while in my state of Michigan. Grow up, Mitt.
#
@mycueb Nice little semantic address creator using the hCard microformat standard. For humans & machines http://t.co/oAakgFss #html #hCard
Cat4D and romainneutron joined the channel