Re: Bidi terminal semantics

From: Jonathan Rosenne (
Date: Wed Sep 15 1999 - 14:03:41 EDT

ECMA-48, which preceded Unicode, envisaged an implicit alternative but
concentrated on the explicit model. We probably thought that the implicit
algorithm will guess the explicit directives. Both implicit and explicit
are logical models, and the results are similar.

TR/53 is theoretical. At the time, Arabic terminals were implicit and
Hebrew terminals were visual.


At 06:57 15/09/99 -0700, Markus Kuhn wrote:
>I just discovered
> ECMA Technical Report TR/53
> Handling of bi-directional Texts
>which provides a nice short tutorial with examples on how to implement
>the ECMA-48 bidi terminal emulation. The curious thing is that is also
>describes a mode in which received characters have an implicit
>directionality, very much like the Unicode bidi algorithm. But unlike
>the Unicode bidi algorithm, ECMA TR/53 does not just discuss a flat
>character stream, but also interactive editing, tabulation, and many
>other things of concern to the developer of a VT100-style terminal
>It will now be interesting to see, how compatible the
>ECMA TR/53 implicit directionality mode is with the Unicode
>bidi algorithm, and whether it is possible to implement
>a terminal emulator that follows both specifications.
>Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
>Email: mkuhn at, WWW: <>

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:52 EDT