Wow Guenter,
congratulations for the excellent, clear, and concise explanation of the
(to my eyes) incredibly messy TeX/multi-language/multi-script situtaion. I
have been using TeX for a few years now, and try to read up on the system,
inclusing reading Tug and so forth. I have never seen such a clear and
compact narrative. I wish I had read it when I started using Lyx and Latex.
It is very helpful. Perhaps you could reformat it into a wiki page? I am
sure it would prevent many users from falling into the semi-confusional
state they are bound to fall into as soon as they start mixing languages in
their files.
Thanks again,
Stefano
On Fri, Mar 15, 2013 at 3:15 AM, Guenter Milde <mi...@users.sf.net> wrote:
> On 2013-03-14, leonid baranov wrote:
>
> > But a multilingual multiscript is somewhat different. It may
> > look like a rarity in academic community. Even in the larger
> > world of written and artistic communications.
>
> > Yet, if you consider a combinatorial space of bilingual pairs,
> > it is no longer such a niche. But then it seems meaningful
> > to seek for a single uniform multilingual solution.
>
> Yes, of course. This is why the Unicode standard was invented (years after
> the invention of TeX).
>
> This changed everything: when I started using LaTeX, it was in order to
> be able to type my name and get it printed properly (with Umlaut) while on
> study in Scotland where Keyboards and Computers did not support any
> non-ASCII characters. It was easier then to write Russian or Greek or
> extended Latin with TeX than using all the different code pages.
> Nowadays, Unicode makes multi-script documents "dead easy", but TeX lags
> behind. Fortunately, it catches up --- with Xe- and LuaTeX.
>
> > Multiscript is just a degenerate case of a "new" language
> > with only the script being different, all the rest is same.
>
> I used the term multiscript for one aspect of a multi-language document.
> Most of my documents are single-script multi-language documents
> (German-English). This is well supported by 8-bit LaTeX also for custom
> fonts.
>
> > Thus, a multilingual multiscript. Which also seems like
> > a least and most natural increment of the original TeX.
>
> Your original question was about the combination of multilingual
> documents and custom fonts. Extensions of TeX/LaTeX in these
> different dimensions exist (far longer than Unicode):
>
> * NFSS ("inputenc" and "fontenc" packages) set up a "TeX standard" the
> font selection,
> * "babel" cares for multi-lingual documents.
>
> Combining these orthogonal extensions of course compicates matters by an
> order of magnitude.
>
> The "babel" package that is part of every basic LaTeX installation since
> LaTeX2e. This means that
>
> * the script,
> * automatically generated text,
> * hyphenation, and
> * typograpic traditions
>
> all are adapted to the specified language.
>
> As the development of the very complex babel package stalled during the
> indroduction of the Unicode-aware Xe-/LuaTeX engines, the replacement
> "Polyglossia" was created. This solved the problem that babel changes the
> font encoding for different languages (required with 8-bit 256 character
> fonts) but Unicode-aware TeX engines will then use the TeX fonts -
> loosing the advance of Unicode-encoded multi-script fonts.
>
> Polyglossia works well, but unfortunately it has a different API, so that
> on the LaTeX level you need \if... constructs to care for use of either
> package. Fortunately
>
> * this is no issue with LyX, that will do "the right thing" when generating
> the TeX file.
>
> * the new "babel" package will be compatible with Unicode TeX engines and
> fonts. http://www.ctan.org/pkg/babel-beta/
>
> > Considering that, do you know if Prof Knuth ever offered
> > any insight as to how best to move beyond the limitation
> > and towards the multilingual TeX?
>
> It is clear that Prof. Knuth himself will not be evolving TeX any
> further. He prefers the stability of a program that does what it was
> originally intended to do. Others have taken up the development, though.
> The future with the TeX-based new engines. Part for part the LaTeX
> packages will adapt or replacements be made. Even now, it is possible to
> write documents that work with both engine types and more and more
> packages hide the decision (e.g. the new "libertine" package uses either
> traditional font setup or the fontspec package to set up the Linux
> Libertine fonts and automatically chooses either the TeX-encoded or OTF
> version.
>
> On the other hand, work continues on solving the issues of multi-script
> custom font documents for traditional TeX engines - see, e.g., the "lgrx"
> and "substitutefont" packages.
>
> Günter
>
>
--
__________________________________________________
Stefano Franchi
Associate Research Professor
Department of Hispanic Studies Ph: +1 (979) 845-2125
Texas A&M University Fax: +1 (979) 845-6421
College Station, Texas, USA
stef...@tamu.edu
http://stefano.cleinias.org
