Eric Muller wrote:
> Munzir Taha wrote:
>
> >Second, why then Unicode choose some characters like
> parantheses to have two
> >glyphs whereas others like sqrt haven't. What's the point?
There is a misunderstanding here: the square root character *does* have the
mirrored property, just like parentheses.
If the RTL glyph is not available for this character, it is a pure font
issue.
> >Third, I am still searching for the right font but can't
> find it yet. Can you help me?
> >
> The case of square roots is beyond what the mirrored data can handle.
> You really need a combination of layout engine and fonts that
> can work together to achieve this. I know that OpenType does
> not accommodate that today (that's why we need a new feature).
> May be AAT or Graphite already handle this?
The fact that square root is declared to be "mirrored" by Unicode means
that:
1) In an ideal world, all fonts should have *both* the LTR and RTL glyphs
and select the proper one according to the bidi context;
2) In a cruel world, any font should have *either* the LTR or the RTL glyph,
whichever is considered more useful by the font designer.
So an OpenType (ot even TrueType) font designed for Arabic could and should
have a RTL square root right *now*. Of course, it cannot have *also* the LTR
glyph; not until "smart" fonts will be smart enough to handle this.
Unfortunately, Munzir Taha's question remains: I have looked in all the
Hebrew and Arabic *TrueType* fonts I have, but the square root is either
missing or it is LTR.
_ Marco
This archive was generated by hypermail 2.1.2 : Fri Mar 29 2002 - 14:16:45 EST