[XeTeX] No underdot in Adobe Garamond Pro?

[Benct Philip Jonsson]

On 2009-05-07 Ulrike Fischer wrote:
> I would certainly never use \"u instead of ü. But I at 
> least do find it much easier to use commands for symbols
> I use rather seldom than to set up a lot of shortcuts in
> my editor which I will probably not remember when I need
> them.

That's true, but there again I use XIM, so,
I'll rather type <Compose><!><d>.

I guess the real difference between us is that I feel
those \<CHAR>{<ARG>} commands make for utter unreadability,
and my main reason for using XeLaTeX is exactly that I
won't have to use them.

 > And even with pdflatex I don't use all chars I
> could input directly. E.g. I always use -- and --- and
> not the corresponding unicode glyphs because at my
> opinion it is much more readable.

I'm all with you on that one!

> Also using the direct glyph will not solve the problem 
> discussed here. If a glyph is not in a font it doesn't
> reappear if you use another input method.
> 
> And at last: commands can be defined to work in various 
> circumstances. \d{<ARG>} will work with new open type 
> fonts *and* with old OT1/T1 encoded fonts and with a lot
> of <ARG>'s -- and you can easily redefine it. It is quite
> possible that a future xunicode defines \d in such a way
> that it tests if the glyph is in the font and use a
> fallback if not.

I would rather have the software be smart enough to check
for each Unicode character in my input file if it exists
in the font, failing that to try replacing it with the
corresponding Normalization Form Canonical Decomposition and 
failing that
to try some other fallback.  That kindh of checking and
rote memorization is what computers are supposed to be
good at and we shouldn't have to use a markup command to
triger that checking; it should be done automatically
for every multi-byte character.  Is it too much in this day 
and age to expect the software to be aware of Unicode
equivalence and do someting smart with it?  If xunicode
already is smart enough to have "\d{d}" trigger a check
if \char"1E0D exists in the output font why can't "ḍ" on its 
own trigger such a check?