[uf-discuss] re: HTML5 support

[Oli Studholme]

Hey Martin,

On Wed, Jul 14, 2010 at 10:45 AM, Martin McEvoy
<martin at weborganics.co.uk> wrote:
> On 13/07/2010 17:59, Oli Studholme wrote:
>> But how could microdata possibly disappear now that Google supports it? ;)
>
> Because Microdata is far to obtrusive to be practical in the "real world"
> for example....
>
> Microdata vcard example from
> http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#vcard
>
> <span itemscope itemtype="http://microformats.org/profile/hcard">
>        <span itemprop=fn>
>                <span itemprop="n">
>                        <span itemprop="given-name">George</span>
>                        <span itemprop="family-name">Washington</span>
>                </span>
>        </span>
> </span>

This is equivalent to

<span itemscope itemtype="http://microformats.org/profile/hcard">
       <span itemprop="fn n">
              <span itemprop="given-name">George</span>
              <span itemprop="family-name">Washington</span>
       </span>
</span>

> from a commercial and practical point of view, microdata is definitely not
> intended to be for "humans first" .

I think it would be more accurate to say RFC2426 is not intended to be
“humans first” ;-) for better or worse vCard doesn’t contain implied
“n” optimisation.

> Anyway believe what you like, microdata needs a *lot* of work before it can
> ever be considered as  "micro" as far as I can see, at the moment It just
> confuses people into using an unnecessary semantic.

Well, to use a non-English example:

<span class="vcard" lang="ja">
	<span class="fn n">
		<span class="family-name">スタッドホルム</span>・
		<span class="given-name">オリ</span>
	</span>
</span>

<span itemscope itemtype="http://microformats.org/profile/hcard" lang="ja">
	<span itemprop="fn n">
		<span itemprop="family-name">スタッドホルム</span>・
		<span itemprop="given-name">オリ</span>
	</span>
</span>

These seem pretty equivalent to me, with the main difference in length
being the itemtype URL. However there are advantages to using URLs for
specifying a vocabulary. Keep in mind the implied “n” optimisation is
arguably potentially dangerous e.g. for a social app that only
collects the user’s name, rather than two separate fields for given
and family names, and then displays this as an hCard. While some
languages that have family-name given-name order don’t use a space
separator (CJK), a quick look at http://twitter.com/boblet shows one
incorrect optimisation for my friend Channy: “윤석찬(Channy Yun)”. As you
can imagine this doesn’t optimise well. I’d look for more but it seems
Twitter’s profile page vcards are completely borked :)

I agree that for marking up a person with their name and URL — if you
can use implied “n” optimisation — microformats is superfast. However
I find I often use hCard for more data than just that, to the extent
that writing them without snippets becomes tiring. And if you’re
making snippets, there’s little difference.

HTH

peace - oli

PS just saw your reply (I can’t keep up! :) Yeah I’ve definitely
wanted an equivalent to itemref for microformats, and hadn’t come
across the include pattern before. thanks!

> in a way that can help microformats *and* be 100% compatible with the way microformats exist now

I don’t think compatibility is so important. Microformats, microdata
and RDFa all target the same basic problem space but each has it’s
strengths and weaknesses. different ideas help each technology improve
(RDFa 1.1 moving towards microformats’ simplicity for example).
Finally (I perceive) µF as an elegant hack to graft new semantics onto
HTML using the tools available; class, rel, rev, profile and coding
patterns. With the changed toolset in HTML5 (including no rev or
profile attributes) it makes sense to reassess methods, and I’m
looking at microdata and RDFa for that reason.