[uf-new] Microformats for hidden data

[Brian Suda]

On Fri, Nov 27, 2009 at 9:19 AM, Fiann O'Hagan <fianno at jshub.org> wrote:
> I was thinking about this last night, and realised that this is
> critical. What precisely do you mean by "visible" in the context of
> microformats?

--- the way I always think about it is, out of sight, out of mind.
Files that you may link to, but are not visible in the browser on a
daily basis tend to get crufty and the data drifts from reality. I
used to have flat vCard files on my server, because i was not staring
at them every day through the browser window i forgot to update my
address, and the data was incorrect. (How many people actively are
keeping their FoaF files current via a text editor? All these files
are fine, but the originating source should be very visible to people
for editing) The same applies to <meta> elements or hidden data in
HTML. If you are not visibly looking at it every day, there is a
potential for the information to be wrong. We´re not talking about if
it is "visible" when you view source, you don't browse sites by
viewing source after following a link. Microformats have always tried
to promote that the data being encoded is "visible", as in rendered in
the browser normally so people can see it (with many eyes all bugs are
shallow) and if there is a problem, it gets fixed quicker due to this
high visibility.

> It's the final case which is most closely related to what I am
> proposing here. They have information which is part of the page but it
> is neither hidden metadata, nor is it rendered in a default view. It
> is intended to be visible to a different audience than the casual end
> user browsing the site.

--- i would say that it is best to avoid display:none on your content.
Yahoo! has chosen to do so, I personally wouldn't recommend it, but
their audience is very different than mine. Had they shown GEO
coordinates it might confuse their customers. They made a choice to
hide it and take the risks of data drift. It is their call to do so.
The microformats still work, but they are not promoted to be used in
this way. Everything SHOULD be visible by default.

> I want to do the same thing, which is to add information to the page,
> in such a way that it is accessible to humans looking at the source of
> the page, and to people with the right parser in their browser, but is
> not part of the view generally presented to end users.

--- this is where microformats are designed to work in the opposite
direction. ALWAYS visible by default. You are asking for hidden by
default, which is a recipe for crufty data. Microformats were always
designed with people in mind first, this seems to be designing for
scrapers first but accessible to people.

> If what Yahoo Local are doing is not an acceptable use of
> microformats, then I accept that it's not the right approach for us
> either.

--- I wouldn't advocate the hiding of the data, but they have their
reasons and take the risk.

> But in the world of mashups, scraping, screen readers and all
> manner of different ways of consuming HTML, it seems like a very
> artificial restriction.

--- well, we can look at it in a different way. Would you trust a
"smaller" set of data that has a higher probability of being accurate,
or loads of hidden data that has a higher probability of being
inaccurate? Search engines took the first approach. None that i know
of still utilize the meta keyword element. It was hidden data, it
drifted from reality, and wasn´t reliable (but it was everywhere).
Instead they look at rel-tag and visible data (who knows their
algorithms might even punish or ignore display:none) and use the more
trustworthy data instead.

Microformats have been shown to work in the real world already with
consuming applications and mash-ups. I don´t think this "very
artificial restriction" has been as big of a problem as you might
think.

I hope this explains abit more about visible and hidden.

-brian

-- 
brian suda
http://suda.co.uk