[uf-discuss] Microformats in Form Fields

Tom Armitage tom.armitage at gmail.com
Wed Oct 4 02:12:58 PDT 2006


On 28/09/06, Drew McLellan <lists at allinthehead.com> wrote:
> That's the ticket. But you'd need a mixture of name and class to
> account for everything... e.g.
>
> <fieldset class="fn  n">
>    <input type="text" name="given-name" />
>    <input type="text" name="family-name" />
> </fieldset>
>
> (obviously incomplete example)
>
> drew.

I'm not sure that name fields are appropriate; whilst they *should* be
freely renamable, very often, at the back-end integration stage,
things like this may get changed, or (depending on the system) may be
unworkable.

I've been thinking about this a while, and I think that one potential
way of getting around the "form fields wanting data that is
microformattable" issue is by making a _minor_ change to the
microformat spec itself.

Namely: whilst microformat classes on most elements indicate what the
element *is*, microformat classes on inputs, selects, and textareas
within form fields (ie: semantically correct input elements) indicate
*what the form desires as input*.

So: an hCard on a div and some lis indicates that I've got a contact
of some form; a form with class "vcard" and  various suitably classed
inputs indicates a form that demands a vcard.

That's it. As a proposla. it obviously requires some refinement, but
it's easily enforceable in any application (because it's just based
aronud parsing html) and it keeps the microformatting to the class
attribute, where it's unlikely to ever be in conflict with any other
demands.

Thoughts?

t.


More information about the microformats-discuss mailing list