>Not so bad when the testing can be done automaticly. Consider the
>following font description:
>
> Line height (maximal) is 16pts == 16/72in. == 2/9in.
> Line height (upercase maximal) is 12pts == 12/72in. == 1/6in.
> Width of string is 232pt == 3.5in
> Font name is MadeUpFont => Class is Serif
>
>An automated search can be performed with minimal built-in knowledge
>(font class) to find an appropriate font. It may not be the "correct"
>font, but it'll work. A user preference would determine if screen
>resolution is taken into account -- if not, assume 72ppi, so 1pt == 1
>pixel.
Anthony,
you're talking about font substitution, right? This would basically work,
but we'd also need to remember average character width, else we might get
Garamond Narrow substituted with New York and then all our button titles
would be cut off to the left and right.
>How about letting people do it as they please: You may enter sizes in
>the unit of your choice. Even lightyears, if you happen to like
>decimals a lot <g>. Internally, I guess we'd probably store it as the
>OS's expect it -- points.
ClarisWorks uses a similar technique: In the German version you enter
measurements in millimeters, but internally they're kept as inches.
Conversion loss is awful. I enter 1 cm and get 0.9 cm next time I open the
dialog window. Thus, this may be a feature of the editing environment, but
it isn't that useful in other cases as the users might get different values
than they are expecting...
Cheers,
-- M. Uli Kusterer
------------------------------------------------------------
http://www.weblayout.com/witness
'The Witnesses of TeachText are everywhere...'
