[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Faking ff ligatures



  The solution seems to me to be this: in those
 cases where a fount ends up with horrible-looking ffl and ffi `ligatures'
 made up of a real fi or fl ligature preceded by an `f', the best fix is to
 replace the `fi' and `fl' ligatures (in ffl and ffi only) with `f' `i' and
 `f' `l'.

that sounds reasonnable to me. For instance, in Lino Didot italic, the
f in the fi lig has a shape (even slope) completely different from
that of the standalone f, hence the string <f><fi> is an
abomination. In that case, it is better to use <f><f><i> with proper
kerning, yes.

 I was assuming this; I was considering playing around with the kerning in a
 TeX file until I'd found something that looked okay, and then using *that*
 kerning for *that* fount only.

ah, that was unclear, as you were talking of changing latin.mtx...

 > fontinst's complex way of doing nothing
 >is the only sane line here.

[the sanest way being maybe to suppress reference to inexisting ligs
at all, but that's another matter, discussed at length already...]

 I reckon the appropriate modification is writing a file ffl+ffi_hack.mtx
 and putting this in it:
 
 \setglyph{ffi}
    \glyph{f}{1000}
    \movert{\kerning{f}{f}}
    \glyph{f}{1000}
    \movert{\kerning{f}{i}}
    \glyph{i}{1000}
 \endsetglyph
 
 \setglyph{ffl}
    \glyph{f}{1000}
    \movert{\kerning{f}{f}}
    \glyph{f}{1000}
    \movert{\kerning{f}{l}}
    \glyph{l}{1000}
 \endsetglyph
 
 and then saying something like:
 
 \installfont    {pgslhr8t} {pgslh8r,latin,ffi+ffl_hack}
                                                {T1} {T1}{pgslh}{m}{n}{}
 
 Is this less mad?

YMMV (your madness may vary...)

Now, how will look your final efforts: efficients?
Thierry Bouche, Grenoble.