Proposing new arrow characters with Bidi_Mirrored=Yes

[Eli Zaretskii]

> Date: Sun, 13 Apr 2025 10:45:15 +0000
> From: Nitai Sasson via Unicode <unicode at corp.unicode.org>
> 
> 
> I just stumbled upon this:
> 
> https://www.unicode.org/versions/Unicode16.0.0/core-spec/chapter-23/#G13001
> 
> > Three pairs of deprecated format characters are encoded in this block:
> >
> >     Symmetric swapping format characters used to control the glyphs that depict characters such as “(” (The
> default state is activated.)
> >
> > [...]
> >
> > Symmetric Swapping. The symmetric swapping format characters are used in conjunction with the class of
> left- and right-handed pairs of characters (symmetric characters), such as parentheses. The characters thus
> affected are listed in Section 4.7, Bidi Mirrored. They indicate whether the interpretation of the term LEFT or
> RIGHT in the character names should be interpreted as meaning opening or closing, respectively. They do
> not nest. The default state of symmetric swapping may be set by a higher-level protocol or standard, such as
> ISO 6429. In the absence of such a protocol, the default state is activated.
> >
> > From the point of encountering U+206A INHIBIT SYMMETRIC SWAPPING format character up to a
> subsequent U+206B ACTIVATE SYMMETRIC SWAPPING (if any), the symmetric characters will be interpreted
> and rendered as left and right.
> >
> > From the point of encountering U+206B ACTIVATE SYMMETRIC SWAPPING format character up to a
> subsequent U+206A INHIBIT SYMMETRIC SWAPPING (if any), the symmetric characters will be interpreted
> and rendered as opening and closing. This state (activated) is the default state in the absence of any
> symmetric swapping code or a higher-level protocol.
> 
> These deprecated format characters differ the ones proposed now in the following ways: 
> - They only apply to characters that have the Bidi_Mirrored property, meaning they don't solve the use cases
> raised. 
> - They apply to an entire span of text, whereas the proposed characters bind to a single preceding character
> each time. 
> - It is ambiguous, at least to me, whether symmetric swapping should be inhibited or activated after U+206B
> ACTIVATE SYMMETRIC SWAPPING .. U+206A INHIBIT SYMMETRIC SWAPPING. The proposed characters
> would leave no ambiguity.

Doesn't the BPA algorithm, nowadays part of UAX#9, supersede what the
above intended to do?