[uf-discuss] Human and machine readable data format

[Breton Slivka]

>
> I think approaching ISO dates as metadata rather than content will remove
> the need to compromise on core principles.
>

I think you'll find that metadata of any kind is a comprimise of the
"microformats core principles". It's information hiding, and the
example that tantek uses is the "meta" tag, which is the prototypical
failure of the metadata approach.

Let's rewind a bit. The problem isn't necessarily that ISO dates
aren't human readable. We've demonstrated that they are, as long as
someone is familiar with the format and what it means. It's not fun,
it's not friendly, so it violates the principle of "Humans first,
machines second". That's an issue, but that's not the most important
issue. The real problem that sparked this whole debate, is that they
aren't machine readable. More specifically, they are read incorrectly
by screen readers. Any solution that involves a quick snap judgement
as to whether a peice of text is "legible" to a sighted human is
irrelevant to that problem. We need to focus on solutions that target
screen readers specifically, because that's what's wrong with the
current solution.

One way to approach this problem is to fix the screen readers. But we
can't do that, so in the meantime, how about just an alternative date
format that a screen reader converts to speech correctly?