Re: Hebrew in LaTeX

Hi Nadav, 

Welcome on board! I really hope you can help get Hebrew LaTeX out 
of the mess it's in. I think the main problem at the moment is that 
there's nobody who takes it upon themselves to be responsible for 
the system as a whole, so various contributions go to different 
places and we end up with little bits and pieces that are not quite 
compatible. Here's a bit of history from the time I was involved in 
the project, I hope it helps.

When I started contributing around mid-2003, the person maintaining 
the IvriTeX distribution was Tzafrir. I would send him occasional 
fixes, and Tzafrir would put them up on the IvriTeX distribution. 
Not very methodological :-)  I think I pretty much fixed the 
problem of Hebrew numbering (the tet, yod, yod-alef, yod-bet 
order), and I also added some code that was supposed to standardize 
the treatment of Culmus fonts. All my patches were also sent to 
the IvriTeX list, so you can find them in the archives.

Somehow Tzafrir and I ended up with somewhat different versions. 
The setup on my computer worked quite well for me ;-)  Around the 
end of 2003 I started working on integrating vowel points, but ran 
into an implementation problem which I didn't find an easy solution 
to, and I gave up on it because of time constraints. As a result the 
system on my computer became unstable (only with regards to vowel 
points), and two years later it's still in that state...

Around the same time (late 2003) Rama Porrat had a problem with 
using the SourceForge version to integrate Culmus fonts into her 
Hebrew distribution for Windows, so I sent her my personal system
(which was still stable). This seemed to work for her, and I think 
she incorporated that into her distribution. I still have a copy 
of the texmf tree I sent her, so I can give it to you if you're 
interested.

Also around that time some parts of the IvriTeX distribution got 
picked up by the main babel distribution, but I'm not sure which. 
I'm pretty sure they did get the addition of "heb" to the command 
names for Hebrew letters (e.g. \alef from Boris Lavva's package got 
changed to \hebalef), and I'm also quite certain that they did not 
incorporate my 8-bit font encoding which included vowel points and 
cantillation marks.

Then in mid-2004 Maxim Iorsh came out with version 0.100 of Culmus, 
and I modified my system to work with it; however, I didn't have 
time to iron out the problems with vowel positioning so I wasn't 
able to distribute these changes, and this modification resides on 
my laptop alone. Other people posted different solutions to deal 
with Culmus 0.100, and I lost track of what got incorporated where.

Since moving to England in August 2004 I haven't added any code to 
my Hebrew system (due to lack of time, not interest). My impression 
from following the list is that there wasn't much activity by 
others, or at least that they didn't discuss this on the list.

As I said, I think the most important thing the project needs is a 
leader -- someone who can bring the various systems in harmony, and 
who can integrate the various bits of code that others produce. I 
can't take this role upon myself, but if somebody did this I 
probably would contribute more than I do now (which is zero).

As for the technical way to go, here are my views; everybody please 
feel free to disagree.

Two factors (at least) contribute to make the IvriTeX distribution 
a mess. One is the fact that the main Hebrew support system which 
is integrated with babel is developed and distributed together with 
a few Technion-centric classes. Ideally, all Hebrew support should 
be part of the babel system, which should work with the standard 
classes. In all my patches I consistently ignored the specialized 
classes, and to this day I don't know what's in them.

The other factor is that the system is distributed as a Linux-style 
development version, with a very unwieldy Makefile; this requires 
you to build the package from scratch using all sorts of tools that 
you might not have or not know how to use. A much better way would 
be to distribute the system as bits of a texmf tree, which you can 
just plug into your existing TeX system no matter how it's set up.

I think the fact that the current system is integrated with babel 
(due to Boris Lavva's heroic efforts) is a great plus which should 
be maintained. Also, the best way to distribute Hebrew LaTeX is as 
part of babel, so it would be available to everyone "out of the 
box". In order for babel to accept an upgraded system it had better 
be as stable as what they have now.

As for fonts, I feel that Culmus is the way to go, since they are 
freely available and initial support already exists. I disagree 
with the snippet Baruch forwarded from Tzafrir's year-old letter: 
in my opinion, there's no way to develop LaTeX support independent 
of the fonts, because any system must have *some* fonts. If we make 
Culmus support separate from babel, then babel will have to resort 
to some other fonts as a default -- probably the ugly fonts which 
have been around since LaTeX 2.09.

The only problem I see with making Culmus the default under babel 
is one of distribution -- babel should ship with the actual fonts 
themselves, as it does, for instance, with Greek (and Hebrew, for 
that matter).

In order to support all the characters from Culmus, the Hebrew TeX 
community has to agree on a standard 8-bit font encoding (which of 
course would be idiosyncratic to TeX, but there's no way around 
this). I tried to push such a proposal two years ago but I don't 
think it was widely accepted. Anyway, I have distributed in the 
past a working installation for an older version of Culmus, and I 
have a private sort-of-working installation for Culmus 0.100. I use 
it for all my Hebrew writing needs (not much at the moment) and it 
works fine. And I'm happy to share.

Another font matter: LaTeX in general is not very good with mixing 
fonts, and if you want to use more than three families in the same 
document (serif, sans and monospaced) you have to resort to some 
non-standard methods. So supporting the full range available from 
Culmus will introduce some idiosyncrasies. 

Anyway, I'll be happy to share with you what I have and to go into 
technical discussions, in private or on the list. 

-Ron.
----
Ivrix-discuss list. See http://ivrix.org.il.
To unsubscribe, please send mail to [email protected] with
only the following line in the message body (NOT SUBJECT!): unsubscribe