[uf-discuss] hCard: url and tel

[Christopher St John]

On Jan 7, 2008 8:14 PM, Tantek Çelik <tantek at cs.stanford.edu> wrote:
>
> The distinction of properties, values, types, schema etc. are well
> documented computer science terms.
>

Actually, in knowledge representation terms they're
usually not. To get around the "what's meta" problem
people generally just pick a level that seems reasonable
to the problem at hand and go ahead knowing that other
choices might have been equally valid. (Computer geeks
can think Java Reflection or the Lisp MOP. When is a
type actually data? Just don't go there :-) )

In HTML for example, the "sematic level" of the various
tags varies quite a bit: <p> is very generic, <cite>
very specific, so denying the question isn't helpful to
those trying to write a new format (or understand the
logic behind existing formats)

I generally agree that the discussion of meta-levels can
be unproductive, but there are choices to be made. A
better answer to the question about data in class
attributes might be:

"Yes, it's data, and there are some fairly deep
questions about what is appropriate and what is not. We
tried to cut through the Gordian knot by simply avoiding
the deep questions. When possible, names are just stolen
from existing standards (hCard). Otherwise, authors have
just used intuition to make some reasonable choices.
There is no hard and fast rule. Different microformats
have very different sorts of "stuff" in the class
attribute (just compare xoxo to hReview), the key is to
make the "stuff" appropriate to the task at hand. If you
want to author a new microformat, you're going to need
to make some choices and experience has shown the
community (and lots of research) will help you with the
appropriateness of your vocabulary and its 'semantic
level'. There are also guidelines on the wiki that have
proven useful in other efforts. Long discussions of the
what counts as meta often end badly, so don't worry
about it too much. Instead, concentrate on existing
practice and trust the community to help with judgement
calls."


> One way to
> learn more about such distinctions is to pick up a book or two on computer
> science and data structure and learn about them.
>

I don't personally mind a little heat in my technical
discussions, but this is exactly the sort of remark Andy
was banned for, and it's unfair to hit a person who
can't hit back.


-- 
Christopher St. John
http://artofsystems.blogspot.com