FMi on text symbol encodings

Ulrik Vieth
Tue, 9 Mar 1999 06:54:35 -0500

>  \unfakable{capitalaccents}                % 8r,8y
>  These are ususally replaced by normal accents, so no problem.

> For fonts like Slimbach's where capital accents are different from
> lc-accents, there _is_ a problem! You're faking T1 glyphs with
> incompatible accents... The problem being Adobe that doesn't encode
> these capital accents in the first place.

Well, I was thinking of fonts with a standarad character set which
don't provide capital accents anyway.  If you have them, fine.
If not, replacing them by normal accents is a reasonable approach.

>  \unfakable{twelveudash}                 % faked 8r,8y
>  \unfakable{threequartersemdash}         % faked 8r,8y or 8x
>  Faking a twelveudash and a threequatersemdash by overlaying 
>  two hyphen characters IMHO seems like a legitimate approach.

> Not to me. i'd rather overlap two endashes. not the same weight, and
> think of fonts where the hyphen is sloped)

Well, yes.  Actually, in textcomp.mtx endashes are used for this
purpose.  I was just a little confused when I said hyphens.

>  \unfakable{hyphendbl}
>  \unfakable{hyphendblchar}
>  Could hyphendbl be replacable by a normal hyphen?

> it's easily fakable anyway, no ? (raise/lower one hyphen).

I didn't try this yet, but that a might be one way to do it.  The 
real problem, however, lies in determining the proper shift amounts.
(Same problem as for faking double square brackets.)

>  \unfakable{lira}                        % 8x (URW)

> 8x (URW) is really something different from 8x (adobe)...

Well, this just means that if you happen to have 8x (URW) fonts, you
will get a lira sign, otherwise not.  If you want to reduce TSA/TSX 
to the Adobe sets and nothing else, you probably have to throw it out.

Cheers, Ulrik.