> - more importantly, we lost the control of the final result, because
> a faked accented letter may be not exactly the same as an actual composite
I am not quite sure what you are saying is the problem here.
Do you think that these two should be `the same' in some ways in which
they are not in current fonts?
In particular, are concerned about:
-- only differences in the final glyphs
-- differences in the metrics (these can cause major differences in
In fact differences equally great in the actual glyphs can happen
simply because two versions of what claims to be the same font are
finally used in the actual rendering on some actual physical device.
> It so happens that no TeX installations are the same and perhaps
> a different font in selected in another system just because a file has not
> been installed.
This is true and a fact of the universe (and always will be).
The interesting question is: what differences in the font resources are important?
How can a typesetting system usefully interact with these differences?
> Despite that, I think that is the right way, and I'm studying how to solve
> these issues. Any ideas?
Probably: but please be specific about what issues you think need to
be dealt with --- then maybe we can deal with some of them by turning
them into non-issues:-).