From: Kent Karlsson (kent.karlsson14@comhem.se)
Date: Tue Nov 04 2008 - 11:32:24 CST
Michael Everson wrote:
> On 4 Nov 2008, at 09:35, Asmus Freytag wrote:
>
> >> Given the mandatory nature of the bidi algorithm, such a loophole
> >> should either be created, or historical scripts should never be
> >> encoded with default RTL directionality.
>
> Old Italic has default LTR directionality, and directional overrides
> have been specified to be used when RTL directionality is desired.
> This mirrors the glyphs.
The bidi alghorithm *allows* the individual glyphs to be mirrored in
this case. But only if an (unspecified) higher-level-protocol says so.
Only characters that have bidi_mirrored=true are required (except for
technical or version limitations) to be mirrored, and no higher-level-
protocol saying not to mirror is to be heeded.
> Egyptian Hieroglyphs have default LTR directionality, and directional
> overrides have been specified to be used when RTL directionality is
> desired. This mirrors the glyphs.
The bidi alghorithm *allows* the individual glyphs to be mirrored in
this (same) case. But only if an (unspecified) higher-level-protocol
says so. Only characters that have bidi_mirrored=true are required
(except for technical or version limitations) to be mirrored, and no
higher-level-protocol saying not to mirror is to be heeded.
> Old Hungarian should have default RTL directionality, and directional
> overrides have been specified to be used when LTR directionality is
> desired. This mirrors the glyphs.
Currently the bidi alghorithm *does not* allow the individual glyphs
to be mirrored in this case. *Not even* if an (unspecified) higher-
level-protocol says so. In particular, characters that have bidi_
mirrored=true are *not* to have their individual glyphs mirrored (and
this latter part should not be changed, even if one adds permission
for a higher-level-protocol to mirror individual glyphs for characters
that have bidi_mirrored=false in this case).
I agree with you that your assumptions about bidi may be reasonable,
but they do not agree at all with how the bidi algorithm currently
stands. See also my (just) previous email on the topic.
/kent k
> Michael Everson * http://www.evertype.com
>
>
>
This archive was generated by hypermail 2.1.5 : Tue Nov 04 2008 - 11:34:54 CST