[XeTeX] VIQR pre-processor wrotten in (Xe)TeX ?

Zdenek Wagner zdenek.wagner at gmail.com
Sat Nov 26 11:02:34 CET 2011

2011/11/26 Vafa Khalighi <vafaklg at gmail.com>:
> Sorry, I have not followed this discussion carefully on what others said but
> if all fails, I would use an external program (a per-processor) to do that.
> You can actually can make it automatic by just using the write18 feature of
> TeX. You write your tex document in VQR, then TeX calls your per-processor
> and converts your VQR to unicode and finally compiles your unicode TeX
> document, giving you PDF.
I did not follow the discussion because someone has already suggested
TECkit maps and it seemed to tme that the problem has been solved. Do
you need more complex preprocessing than just simple mapping of input
bytes to output bytes? I can imagine that it may be complex and cannot
be done in a simple step. You can look to the xetex-devanagari
package. The Velthuis maps were created by me. Vowels have to be
treated in a special way, viramas have to be inserted in between
consonants, thus I cannot do everything in a single step. Moreovever,
there is just a small difference between Sanskrit and Hindi, therefore
the Sanskrit map is essentially the Hindi map with one more step.
Maybe these samples can help you.

> On Sat, Nov 26, 2011 at 12:26 PM, Bruno Le Floch <blflatex at gmail.com> wrote:
>> > Bruno Le Floch wrote:
>> >> Is it simply a matter of going through the string and replace various
>> >> characters by TeX accents (and take care of character order), or does
>> >> the result have to be Unicode?
>> >>
>> >> E.g., does it have to be a(a^ → ăâ, or can it be a(a^ → \u{a}\^{a} ?
>> >> If the second one is ok, then it shouldn't be too hard to write the
>> >> conversion in TeX macros. Please provide a list of the necessary
>> >> conversion rules.
>> >
>> > Vietnamese is difficult for TeX, because over a single
>> > character there can be a requirement for both an
>> > accent (to indicate a pronunciation different from
>> > that for the unaccented letter) and a tone marker
>> > (which applies to the whole word, but which has
>> > a canonical placement that may well be on an
>> > already accented letter).  Hàn Thế Thành created
>> > a way of accomplishing this using standard TeX,
>> > but XeTeX can do it natively.  Incidentally,
>> > the "ế" at the end of Thành's middle name demonstrates
>> > exactly the problem : something that TeX cannot
>> > accomplish without special fonts, since it has
>> > no primitive for positioning two diacritics over
>> > a single character.
>> Thanks for the explanation. So we need to produce a Unicode string.
>> Firstly, I should have asked: is it possible to require the Vietnamese
>> VIQR input as an argument of a macro, rather than being typeset
>> directly? e.g., \viqr{a^o' e\.}. I'll assume that it is ok.
>> Then do you have a list of all Vietnamese characters and their VIQR
>> representation? Or some precise reference to what transformation you
>> want to apply? In particular, how many different characters do you
>> expect to get on output (that should be the size of the "alphabet")?
>> Regards,
>> Bruno
>> --------------------------------------------------
>> Subscriptions, Archive, and List information, etc.:
>>  http://tug.org/mailman/listinfo/xetex
> --------------------------------------------------
> Subscriptions, Archive, and List information, etc.:
>  http://tug.org/mailman/listinfo/xetex

Zdeněk Wagner

More information about the XeTeX mailing list