Re: Element for Numbers

David Woolley <david@djwhome.demon.co.uk>:
>
>>   Either a signed decimal number with optional signed two digit exponent
>>   and mantissa separated by a dot (e.g. -1.2345E+03),
>
> Some non-transcendental number can be represented exactly in finite
> strings but not in this format.

I probably should have looked up an existing rule for representing numbers
and refer that.

>> ISO [ISO 8601]
      ^insert "date"

> I'm also concerned that this would be an excuse for presentational
> hacks in the element contents, when a semantic language requires that
> the content be meaningful in its own right.

Can you provide an example, please? I cannot see a reason for your fears.

>> Dimension
>>   A case dependent string describing a dimension according to table 1:
>
> Much too complex for HTML,

It isn't really, it just looks like, because there are so many tables. If
there was a normative resource I'd been aware of, I would have cited that
instead. In practice you wouldn't care much, just hack in the unit you
desire -- it should be there, if common.
Probably just keywords were better than '&' or '*', though.
When dividing datetimes and other numbers into two elements, the dim
attribute would become even simpler. I included a preceding dollar sign
('$') for currencies, which might as well be hyperfluous.
I forgot units like px, pc and pt; where the last collides with the unusual
'pico ton' (aka µg).

> and inevitably going to become out of date.

Huh? I'm pretty sure SI units and ISO date are going to outlive HTML.

Christoph Päper