[tei-council] for Council to consider: new datatype

[Syd Bauman]

SR> data.webMeasure?

I like that a lot better. Of course, XSLFO may well be used for
printed pages. But 'data.outputMeasurement' seems kinda clunky.


LB> (a) why restrict it to web measurements? there might well be
LB>     other attributes for which we'd want to specify number+units 
LB>     in a single value.

Darn good question. The difficulty is that if we use it for other
things, we need to include all sorts of units, some that won't make
any sense for a given attribute. So either we make up a separate
datatype for each set of units a set of attributes wants, or we live
with a system that permits ridiculous values like
  <graphic height="1.1m" width="76kg"/>
  <person height="256px" weight="3em"/>
  <when interval="14.7mg"/>
Personally, I would much prefer a few extra datatypes. (There can't
be that many needed, can there?


> (b) some of these units look a bit exotic to me : isn't a "rem" a
>     unit of exposure to radiation? couldnt find it in css3 anyway

I think they're outright weird. Got 'em straight from 
http://www.w3.org/TR/2005/WD-css3-values-20050726/#numbers0, where
"rem" is listed as "the font size of the root element".


> "dimension" suggests the axis along which the measurement is made
> to me, not the actual measurement.

I think you're right, "measurement" is better. If we go with one
datatype fits all, then "data.measurement" makes sense. If we go with
different datatypes for different kinds of measurement, then things
like "data.lengthMeasurement" and "data.massMeasurement" or
"data.measurement.typesetting" and "data.measurement.length" and
"data.measurement.mass" might make sense. (We don't really have any
mass ones, I don't think; just using that as an example.)