[luatex] Vital RTL issues in LuaTeX

Robin Fairbairns Robin.Fairbairns at cl.cam.ac.uk
Fri Apr 10 12:08:53 CEST 2009


Reply-to: luatex at tug.org
X-Mailer: MH-E 8.0.3; nmh 1.3; GNU Emacs 22.3.1
--------
وفا خلیقی <vafakh9 at gmail.com> wrote:

>   Even assuming the unicode bidi algorithm, someone still has to
>   define how the algorithm interacts with tex control sequences
>   and font processing. For example, how does the algorithm interact
>   with right braces ?
> 
> That is supposed to be the developer's job not the user's job.

except when the developer is designing a program for a user who has far
more experience in the subject.

who (who has no experience with persian) is supposed to guess that
section numbers in persian are l-r groups of digits arrange r-l (which
is what i _think_ you said)?

without your input, or the input of some other intelligent persian
reader, how is taco supposed to get this stuff right?

who it is that provides the "algorithm" isn't really germane here: if
development is working properly, the "user" provides enough information
that the "coder" can develop the algorithm with little trouble.

robin

who had quite enough of this sort of issue when working as a commercial
sw developer (including being told on one occasion that my program was
deficient because it failed to meet a requirement the customer had never
told me about).


More information about the luatex mailing list