Some things I would want to see in a typesetting system:
- Having a typesetting quality and control like TeX does, including mathematical typesetting, although improvements could be made to that, too. (I think I would also like the syntax more like TeX has)
- You can include PostScript codes within the document which can run during the typesetting process (rather than only during output), therefore allowing it to affect decisions of page breaks, etc, as well as allowing PostScript to draw diagrams and control text rendering. (I had written a PostScript program to load PK fonts, so that would make it possible to use PK fonts from TeX as well.)
- Full support for non-Unicode text (without converting it internally to Unicode). (It is OK if it also supports Unicode as long as the code to support Unicode is avoided when not using it (in the entire document or in a part of it).)
I think Unicode is messy and is not the best way to do i18n and m17n and l10n and a11y and many other things (although there are also problems with existing implementations that do not have to do with the character set, the character set is one of the problems), and I also think that it is not good to insist on using one character set for everything (especially if that character set is Unicode).
UTF-8, UTF-16, UTF-32, UTF-EBCDIC, etc are encodings of Unicode. EUC-JP is not an encoding of a subset of Unicode; it is an encoding of JIS, which is a different character set. The PC character set is not a subset of Unicode; it is the PC character set. Being able to be mapped to Unicode in some cases does not make it Unicode (nor does it mean that these mappings are necessarily "clean", but even if they are, that still wouldn't make non-Unicode character sets to be (subsets of, or even supersets of) Unicode).
- Having a typesetting quality and control like TeX does, including mathematical typesetting, although improvements could be made to that, too. (I think I would also like the syntax more like TeX has)
- You can include PostScript codes within the document which can run during the typesetting process (rather than only during output), therefore allowing it to affect decisions of page breaks, etc, as well as allowing PostScript to draw diagrams and control text rendering. (I had written a PostScript program to load PK fonts, so that would make it possible to use PK fonts from TeX as well.)
- Full support for non-Unicode text (without converting it internally to Unicode). (It is OK if it also supports Unicode as long as the code to support Unicode is avoided when not using it (in the entire document or in a part of it).)