[Aleph] Mem - new release

Javier Bezos jbezos at wanadoo.es
Sat Aug 28 10:06:24 CEST 2004


Idris,

>Congratulations on your latest release!

Thank you! (for some reason I missed this line when
I first read your message).

>This is very interesting. There are some serious inefficiencies (as opposed to 
>bugs) in the otp method and I was wondering whether a perl (or python) 
>pre-processor might be in order (something that could be run by texexec for 
>example). This is especially with regards to e.g., dicaritics upon ligatures 
>in expert Arabic fonts. The regular otp method is very inefficient in my view, 
>and quite painful to configure...

Actually, what's preprocessed is the mtp file to create
an otp one, so that [POUND SIGN] is equivalent to @"00A3.
The next step would be to translate also UTF-8 sequences.
It also replaces \ by @"E000 which mem catcodes as
escape (and similarly for $, &, {, }, etc.). The problem
of preprocessors applied to documents is that macros
are skipped over--f\macrodefinedasi is not the same as
fi.

Regarding otp's are painful to configure, we are agreed
and I hope mtp files could help in the near future to
make the task somewhat simpler, so if you have any
suggestion it would be most welcome. This way we can
have prototypes to experiment with, so that in the future
otp2ocp itself could be extended with new features if
necessary. (One of the reasons I use Python is that it's
a great language for prototyping.)

Javier


More information about the Aleph mailing list