reinhard.kotucha:
What the new engines offer is that one can write a text in a readable
form (H. N.i instaed of H\`a N\d\ocircumflex i). This is what people
expect nowdays.
I agree, the intput methods for various languages are needed. For example
I started with TeX 22 years ago only after TeX converted from seven bit to
eight bit and I was able use usual input methods for my language. The
seven bit TeX seemed to me as totally unusable (it was unsupposable to
write Ol\v s\'ak instead Olšák for example in common Czech text). On the
other hand, when I sent my TeX source to the world I convert these Czech
characters to 7bit TeX notation because this is robust. But we are
discussing about typesetting specific characters (--, ---), no language
specific characters.
If you want to preserve the old behavior, you can
provide macros like \def\alpha{.} after all (dot is real UTF8 alpha here)
Do you plan to remove \mathchardef\alpha=... from basic TeX macros and
replace them only by \mathcode"03B1=... ? Why? The only reason is that
there exists one input method in one text editor? Is it sufficient reason
for such major decision? Why is inpossible to have both declaration
(\mathchardef\alpha=..., \mathcode"03B1=\alpha) in paralell?
This decision is similar to removing -- and --- ligatures as default
ligatures from TeX fonts.
Petr Olsak
