[comp.text.tex] No accented characters

A4422DAE@AWIUNI11.BITNET (Konrad Neuwirth) (09/29/90)

About the subject of the new fonts.

  Reading the last mail about the fonts having no accents at all,
I should say that I am rather shocked. Coming from a language that
relies on accented characters, it is hard for me to understand that
you want to throw those right characters out. TeX educates us to
go for quality, so why should we use second class umlauts? Umlauts
are not, in fact, accented characters but characters on their own
(if you care to look into those old lead things, they had their
umlauts as distinct characters (which could even have a rather quite
different shape than those characters unaccented). If you understand
written german, look it up in `Das Buch des Setzers' by Genzmer).

   Also talking to people from sweden, they have the same problem, too.
Especially with things like the ^"A or so (the A changes, believe me).
So it *IS* necessary to have those things. We are not just PostScript
corrupted.

   I am also somewhat dissapointed that the necesary ligatures like
ch, ck and ft didn't make it into the font. Some other characters are
missing, too, to make some reasonable typesetting. But then, I think we
are talking about different characters!

   I am sorry to say the follwing that geralized, but the Americans always
have a tendency to think that one (at best theirs) solution is good for
every problem.

   The creators of the new scheme never claimed that it solves all the
problems of all fonts of the latin family. It is a better tool for some
languages like German, Swiss German, Polish, Hungarian, Swedish and so
on, but it does not claim to be the best thing for every language. It also
is not meant to replace CMR10 (the math characters are missing then), but
to be used for text typesetting. And if you want to have a different layout,
so then go forth and use it. Knuth gave us a tool to simplify all those things,
so that maybe german typesetting, one day, will be correct again, using the
right ligatures and the right uppercase letters and have real umlauts.  Maybe
we will create a font that does that, but we won't recreate it all in metafont.
Virtual Fonts is the answer!

   Come on guys! You got to be more liberal than that (even if it is hard
for americans.
(set flame off).

have a nice day,
//konrad

+----------------------------------------------------------------------+
 As I'm not working for anybody, I can't be representing their opinion!
  Konrad Neuwirth                   Earn/BitNet: a4422dae@awiuni11
  Postfach 646                      VoiceNet: KONRAD!!!!!!(louder).
  1100 Wien, Austria (Europe)       Bloody BIX is too expensive.

dhosek@linus.claremont.edu (Hosek, Donald A.) (09/30/90)

In article <9009291501.AA02718@lilac.berkeley.edu>, A4422DAE@AWIUNI11.BITNET (Konrad Neuwirth) writes...
>About the subject of the new fonts.

>  Reading the last mail about the fonts having no accents at all,
>I should say that I am rather shocked. Coming from a language that
>relies on accented characters, it is hard for me to understand that
>you want to throw those right characters out. TeX educates us to
>go for quality, so why should we use second class umlauts? Umlauts
>are not, in fact, accented characters but characters on their own
>(if you care to look into those old lead things, they had their
>umlauts as distinct characters (which could even have a rather quite
>different shape than those characters unaccented). If you understand
>written german, look it up in `Das Buch des Setzers' by Genzmer).
[etc.]
>//konrad

Well I personally have no problem with providing pre-accented
characters. I've been saying for years that it's the only way to
get quality typesetting (although there are cases, e.g., Hebrew
and Arabic where floating accents are a necessity) for latin
typesetting. However, speaking from the perspective of a font
designer, somebody tell me what character code to use for the ct
ligature. How about long-s and all its associated ligatures. What
about the extra ligatures I need for my Italic typeface? It's not
that the TeX world has become PostScript-corrupted, they've
become TeX-corrupted. I'd like to see two standard latin
encodings that will be far more inclusive and leave (at least) 16
vacant glyphs for special typographic features. Otherwise, I'm
not going to bother with that coding scheme for my typefaces at
all.

-dh

---
Don Hosek                       TeX, LaTeX, and Metafont support, consulting 
dhosek@ymir.claremont.edu       installation and production work. 
dhosek@ymir.bitnet              Free Estimates.
uunet!jarthur!ymir              Phone: 714-625-0147
                                finger dhosek@ymir.claremont.edu for more info

phil@ux1.cso.uiuc.edu (10/01/90)

I read and hear these things:

1.  Scalable fonts are bad, particularly for small fonts, because the
    proportions do need to change somewhat with actual size for proper
    readability.

2.  Umlauts (and quite likely others) are more than just mere letters
    with "things" stuck on them.

Solving these problems is not easy either.  Treating umlauts like accented
letters is not the best way to go, but it does solve the immediate problem
which is a total lack of such characters in the American character sets.

However, because this approach has solved some other problems as well, I
do not want to see it go away until ALL of the problems it solves are
solved by other means.  Umlauts for Americans is just one of them.

One of the reasons I like TeX is that it allows me to typeset text in the
Esperanto language as well as many others.  The approach of treating these
letters as accented letters made it possible (albeit not perfect) to have
these letters without someone having to create the fonts for them AND TO
MAKE SURE IT IS DONE FOR ALL LANGUAGES.

Of course it is not yet even close to perfect.  Cyrillic is missing.  And
languages like Arabic, Japanese, Hindu, Georgian, etc., cannot yet be
typeset.  A lot of work is due.

When the work is done, however, just be sure not to removed the facility
that exists now until you can prove that the work is absolutely complete.
This is not likely for quite some time.

Now if you have some scalable PostScript fonts for ALL the possible letter
shapes in all the languages whose alphabets is based on the Roman alphabet,
then I am interested.  If it omits one or two languages, the value of it
is diminished very greatly.

--Phil Howard, KA9WGN-- | Individual CHOICE is fundamental to a free society
<phil@ux1.cso.uiuc.edu> | no matter what the particular issue is all about.

Damian.Cugley@prg.ox.ac.uk (Damian Cugley) (10/01/90)

From:		 Konrad Neuwirth <A4422DAE@AWIUNI11.BITNET>
Message-Id:	<9009291501.AA02718@lilac.berkeley.edu>

> X-Unparsable-Date: Sat, 29 Sep 90 16:54:58 MEZ

>   Reading the last [article] about the fonts having no accents at all

(a) I'm not against accents
(b) I'm not against having accented characters as separate glyphs
(c) I do think trying to fit every accented letter used by every
    language ever existing into a font is just unworkable

I would suggest having separate fonts per language/language group with a
smallish number of accented letters & the facilities to make good
accents of other letters for short quotes in other languages.  Remember,
I use an accented language, too (there are English words with accents.)

The proposed Latin8 encoding isn't *completely* terrible, it just has a
few problems, which just happen to be mainly where it attempts to be
more PostScript-like - e.g., ASCII circumflex, straight quotes etc.
Thus my assertion that PostScript was rotting people's brains.

I am more annoyed with the inclusion of straight quotes, ASCII
circumflex, visible space etc. than with the accented letters.  


>    I am also somewhat dissapointed that the necesary ligatures like
> ch, ck and ft didn't make it into the font. Some other characters are
> missing, too, [needed] to make some reasonable typesetting. But then,
> I think we are talking about different characters!

One of the things I am griping about is the apparant obsession with
cramming 256 characters into a "standard" font encoding which leaves no
room for features like more ligatures - e.g., German ch, ck, or more
eccentric ones in display fonts.  Which ligatures are wanted
**probably** depends on what language is being used; different countries
*do* have different traditions.


>    I am sorry to say the following that generalized, but the
> [anglophones] always have a tendency to think that one (at best
> theirs) solution is good for every problem.

That's interesting because it was me that was ranting about the proposed
"one solution" to the problems of different languages having different
accented characters.  (Maybe because it doesn't include the Esperanto
accents!)

teexdwu@ioe.lon.ac.uk (DOMINIK WUJASTYK) (10/11/90)

In article <1048100005@ux1.cso.uiuc.edu> phil@ux1.cso.uiuc.edu writes:

>One of the reasons I like TeX is that it allows me to typeset text in the
>Esperanto language as well as many others.  The approach of treating these
>letters as accented letters made it possible (albeit not perfect) to have
>these letters without someone having to create the fonts for them AND TO
>MAKE SURE IT IS DONE FOR ALL LANGUAGES.

This is a very important point indeed, and must not be overlooked by
those people struggling with the character set problem.  Any character
set, even one with pre-composed accents for Swedish or whatever, really
should have the separate accents too, for use with \accent, so that
unusual words can be dealt with.  And unusual words are very usual.
I too came to TeX partly because it could typeset all the Indic 
words I use that have macrons over the vowels and underdots under
many consonants, s-acute, and so on.
 
>Of course it is not yet even close to perfect.  Cyrillic is missing.  And
>languages like Arabic, Japanese, Hindu, Georgian, etc., cannot yet be
>typeset. 

TeX is more perfect than you think.  I would guess that you are not a
member of the TeX Users Group, or you would know from the excellent
journal TUGboat that all the above languages and scripts can be typeset
by TeX already.  (If by "Hindu" you mean "Devanagari", and by
"Georgian" you mean "extended Cyrillic".)

Phil makes an extremely important point.  Text fonts that omit floating 
accents will not be widely adopted, mark my words.

Dominik