[tex-eplain] TeX--XeT

[Laurence.Finston at gmx.net]

I don't know what you mean by two-bit or three-bit mappings.  TeX was originally designed for 7-bit fonts, which is why, for example, Computer Modern Roman has 128 characters.  Later, this was changed to 8-bit, so that a font may contain 256 characters.

While TeX's internal character encoding is, as mentioned, independent of the encoding of any given computer, in practice it is based on ASCII, so if you want to use the letter "A" for a different character from a different alphabet and the corresponding font has that character in position 65, then that will work.  Otherwise, there is something called "virtual fonts" which make possible to do a kind of indirect mapping.  However, I've never done anything with this and can't explain it.  There are some TUGboat articles about it.  A lot of work on fonts, particularly for making TeX work with PostScript fonts, seems to have been done around 2005, but it appears to have been pretty much wrapped up, as far as I know.  I don't really know that much about this subject. 

Most people seem to use TeX because of its capabilities for mathematical typesetting and the Computer Modern fonts because they are one of the only font families that provide a fairly complete set of mathematical characters.  TeX was designed for use with METAFONT fonts but apparently a lot of people wanted to be able to use PostScript fonts, for some reason which is a mystery to me, so much work was done to make this possible.  This is an issue which is at most tangential to learning how TeX and the plain and eplain formats work.

TeX provides the means to map the characters available on your keyboard to the characters in the fonts.  This may not be convenient, which is why there are various packages to make it easier.  Nor was TeX originally designed to use PostScript fonts.  Neither CJK nor getting TeX to work with UTF-8 or Unicode or PostScript has anything to do with TeX in and of itself, plain TeX or eplain.  Beyond implementing some limited compability with LaTeX, eplain also doesn't have anything to do with LaTeX.

It's possible to use the eplain macros without a deeper understanding of how TeX works.  However, if your goal is to learn how (plain) TeX and the eplain macros work, I would suggest buying or borrowing one of the books on TeX, or see what (non-bootlegged) material is available online and work through it starting at the beginning.  Things like \newlinechar aren't really central to understanding how (plain) TeX works.

> Gesendet: Freitag, 13. Oktober 2023 um 06:49 Uhr
> Von: terry.s at Safe-mail.net
> An: tex-eplain at tug.org
> Betreff: Re: [tex-eplain] TeX--XeT
>
> > Of course, 256 characters isn't enough for Asian fonts.
> ...
> > If you have the fonts, you can typeset in any RTL language.  It works for Hebrew and Arabic.
> >
> 
> That seems confusing, until thinking that LaTeX and some packages work around this. But AFAIK, they only provide 2-bit mappings.
> 
> Then I think the issue is the Asian fonts distributed with TeX Live are created for engines like pTeX and upTeX (with the caveat that pTeX uses an older "Shift-JIS" encoding). Those engines wouldn't use the old mapping system.
> 
> Again, I tried to install some using *fontinst* (spelling?) et. al a while back, but the instructions to create mappings from PostScript files were complex and didn't work. I don't have the expertise for that, and the instructions predated LaTeX's UTF-8 support. (Forget about the 3-bit issue specifically for LaTeX, I couldn't even install them + create mappings.) I have very nice PS fonts on my computer.
> 
> I did not have Internet for several years and couldn't even install TL from the web. (It would take a full day, and I don't mean business hours.) I joined TUG and was exclusively using the DVD and some VERY LIMITED time on public WiFi. I would look up as much info as I could and save it to USB for reading later. (A horrible way to learn something!)
> 
> Now I am aware there are new tools which should be appropriate to UTF-8, and should have new, appropriate instructions. I'm putting that off for the moment, though.
> 
> I could not imagine typing charcodes for each character. With MS-IME, you just type in roman characters phoenetically (Japanese has no spaces):
> arigatougozaimasu!
> And you get:
> ありがとうございます！
> 
> And that works in virtually all applications. (Source must be saved as UTF-8 or any Unicode encoding, which TeXmaker is perfectly capable of doing.) Any TeX engine ... even one with Unicode support ... will have an error when you try to compile if it doesn't support the language, i.e. have character mappings for CJK characters and perhaps specifically the "hyphenation patterns" (as was the case with OpTeX). It just doesn't understand the input.
> 
>