Re: Bidi terminal emulation considered harmful
Tomohiro KUBOTA wrote on 2001-11-28 02:13 UTC:
> To study about BiDi, is the following document enough, or do you
> have some recommendation? (though I have not read the document yet).
>
> Unicode Standard Annex #9 The Bidirectional Algorithm
> http://www.unicode.org/unicode/reports/tr9/
ECMA 48 and ECMA TR/53 is probabaly more relevant. The Unicode bidi
algorithm was written for paragraph formatting systems, not for VT100
style terminal emulators. BiDi is an intrinsically conceptually messy
area anyway, you are unlikly to find a really satisfactory and elegant
solution for terminal emulators.
http://www.cl.cam.ac.uk/~mgk25/unicode.html#xterm
http://www.ecma.ch/ecma1/TECHREP/E-TR-053.HTM
http://www.ecma.ch/ecma1/STAND/ECMA-048.HTM
Read the ECMA stuff very carefully. It is somewhat difficult to digest
though, so take your time, but it is the most thorough attempt ever made
to define bidi terminal semantics. I don't think though anyone ever
implemented it.
You might end up with the same conclusion as I did: Bidi is best kept
completely out of the terminal and the vision of bidi ever working as
naturally and simple between terminals and simple Unix tools such as cat
or ls is probabaly a naive illusion. Making bidi simple will make a lot
of other things horribly complicated, and very few developers will
follow. The use of Hebrew, Arabic and Syriac in writing order should
under Unix probabaly be restricted to text editors and word processors.
In file names, urls, environment variables, etc., use either visual
order or the latin script. Really.
Markus
--
Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at acm.org, WWW: <http://www.cl.cam.ac.uk/~mgk25/>
Reply to: