building a font picker

[Graham]

On Sat, 19 Jul 2003 01:06:38 -0600, Alex Rice <alrice at ARCplanning.com> wrote:

>re: comparing I vs. Z or the whole alphabet. Picking two letters I
>would take I and M. I think M is supposed to be the widest letter in
>general, hence an "em-dash"... I could be wrong about that.
>
>But my reason for using entire alphabet instead of I and Z is that I
>think it could be possible for some font to have an I and Z have a
>difference in width that is less than the resolution of "the
>formattedWidth". I can't visualize what the font would be, but it's
>certainly possible. That's why I would make it more reliable by
>comparing multiple letters. Thoughts?


For normal fonts, I think you are safe with i and M (that's lower case i 
and upper case M). These are traditionally the narrowest and widest letters 
in a font. Of course it would be **possible** to design a proportionally 
spaced typeface which breaks this rule to the extent of making these two 
letters equal, but it would be a very strange one. However I have just 
thought of two other potential showstoppers - (a) the incomplete font sets, 
where for example there are only capital letters with all the unused ones 
just the same letter (usually printed as a small rectangle); and (b) 
display fonts where the 'letters' are in fact symbols (as in Zapf Dingbats 
etc). In these fonts all bets are off for individual letters, so if you are 
simply running through all the fonts on your system to divide monospaced 
from proportional fonts, you might have to go back to comparing the 
complete alphabet as before.

Graham

---------------------------------------------------
Graham Samuel / The Living Fossil Co. / UK & France