[uf-discuss] Human and machine readable data format

[David O]

On Sat, Jul 12, 2008 at 9:47 PM, Paul Wilkins <paul.m.wilkins at gmail.com> wrote:
> With the current system authors are obliged to write up a specific
> date format for computers to parse, as well as one for humans to read.
>
> They should not have to produce both types on every occasion.
> If a parser isn't able to work out tomorrow or next week from context,
> then that date could be made more explicit in the code until solutions
> are devised.

Unfortunately, natural language processing is very, very hard, even
when you're only focusing on one language, and we're attempting to
create a solution for every single language in which (X)HTML content
is publically published.  As far as I can tell, the choice here is
between an ideal solution which cannot currently be implemented, or a
kludge that can.  Also, as far as I can tell, one of the basic
principles of the microformats community from the outset was to value
things that actually worked over idealized, perfect solutions,
precisely because such solutions almost never actually get built.

> It's not right though to demand content authors to duplicate the dates
> they're entering.
> Humans first, machines second. That's how it should be.

The key word here is "should."