Processing Sanskrit texts in utf-8 notation with Omega TeX
Some of the tools in the package are not platform specific, most
should be portable, but they are tested in and designed for LinuX,
GNU-Emacs 21 and Omega (preferably version 23.2).
A more recent LinuX distribution should contain all the components
required, but make sure to install emacs support packages like auctex,
mule-ucs, teTeX including Omega, which are probably not selected by
the "automatic" installation procedure. My own system is SuSE LinuX,
and here one has to switch to manual package
selection. Furthermore the tipa fonts and macros are required, which
are not usually part of TeX distributions.
Unfortunately there are problems specific to particular LinuX distributions,
for instance, in the current SuSE 9.3 omega does not work out of the box. It
took some time and the kind help of Thomas Esser to trace the error. If you use
that distribution, please change the value of extra_mem_bot in the file
/etc/texmf/web2c/texmf.cnf to 0 and recompile the format files with
fmtutil --all. Otherwise omega will produce awkward errors.
I also recommend reading the detailed information on Christian
Coseru's page on Sanskrit
Unicode Text Processing.
- Download utf-skt05.tgz
- Expand the archive (tar -xzf utf-skt05.tgz), or start "mc"
(Midnight Commander) and open the archive by pressing "Enter". You may
then copy the files in the subdirectory /texmf to the appropriate
locations as indicated through the subdirectories in the archive.
Since a certain knowledge of TeX and the LinuX system is necessary in
order to make use of these tools, I would recommend an introduction to
TeX, the teTeX manual and a brief description of Omega for all those
who do not know where to copy the files. Do not forget to run texhash,
so that the files can be found by TeX. The file
/etc/texmf/web2c/updmap.cfg has to be changed so that it can find dvng.map. Run updmap.
- The files in /user have to be copied to the directory of the
user(s) and may not work without adaptations and reconfiguration of
preferences. One could start emacs from the .xinitrc with the line:
emacs -fn *misc-fixed-bold*15*iso10646* &
Other Unicode fonts can be used, but note that not all of Unicode
fonts display the accented characters as expected.
Known Problems with older versions (pre SuSE 9.3)
In the process of testing the package one bug in the Omega program
surfaced which was reported and corrected by Norbert Preining. It
caused a segmentation fault after omega had read a certain length of
Nagari text. Norbert has also made the compiled omega binary available
for download (here). For those who have
to or want to install it a few notes may be helpful.
Firstly it should be sufficient to copy the files to the proper locations:
omega -> /usr/texmf/teTeX/bin
omega.pool -> /var/lib/texmf/web2c
Please check your systems, this applies only to LinuX Suse 8.0.
On my system the paths were not found by the omega binary, that means
iniomega complained of not finding files in the TeX tree. There are
certainly more elegant ways to solve this, but copying some of these
files to the directory where the omega binary resides has helped,
specifically these are:
texmf.cnf (from /etc/texmf/)
With these two in the same directory as iniomega/omega the creation of
the format file lambda.oft succeeded. The latter had to be copied to
the directory, in which omega or TeX can find them. Perhaps the latest
distributions of TeX are already free from that error.
J Hanneder (firstname.lastname@example.org )