Jukka K. Korpela wrote:
(The discussion started under the heading "font-family font sizes", but I
now changed the Subject to more specific.)
On Fri, 18 May 2007, Bruno Fassino wrote:
Putting together the above, if we trust Gecko, the aspect ratio of a
font can be obtained simply observing the size of an element
dimensioned using 'ex' units.
The results are consistent with those I got with my simplistic method that
compares the letter x in large font size with a stickyard. See
http://www.cs.tut.fi/~jkorpela/x-height.html
It seems that different ways lead to the same results, implying that the
figures mentioned in CSS specifications are wrong - most notably when
they say that Verdana has an aspect ratio of 0.58, but it actually has 0.545.
This, as well as methods of finding out the x-height value for specific
fonts, is of some importance to authors who wish to use font-size-adjust
(which helps on Firefox 2 and does no harm when it doesn't).
I don't think you can rely on the size of elements dimensioned in ex
units. Eric says in his book that many user agents get their value for
ex by taking the value of em and dividing it in half. This is because
most fonts don't have the value of their x-height built-in, and it's
difficult to compute, as you discovered. Since lowercase letters are
about half as tall as uppercase letters, user agents assume that 1ex is
equal to .5em.
Also, would monitor resolutions or dot pitch have any impact on your
calculations? Since you are taking a ratio, I wouldn't think so, but I
wonder...
Lori
14 messages in org.css-discuss.lists.css-d[css-d] Estimating x-height
