Re: Mirroring Unicode symbols in Arabian

The Unicode report is a list of characters that *can* be mirrored.  It is
not a requirement that they be used.  As it says, they may be less than
ideal in some cases (e.g., italics, but that's not likely in math).  The
best solution is for the MathML rendering engine have a font available that
has glyphs for characters suitable in an RTL context.  Barring that,
character mirroring can be used as a fallback for those characters that can
be mirrored and actual bit flipping (i.e., rendering the character and
inverting that image) for those that can't can be done, but it is a poor
substitute for a font that has the appropriate alternative renderings.  See
(for example) http://www.microsoft.com/typography/otspec/ttochap1.htm for
more info.

Neil Soiffer
Senior Scientist
Design Science, Inc.
www.dessci.com
~ Makers of MathType, MathFlow, MathPlayer, MathDaisy, Equation Editor ~


On Thu, Feb 21, 2013 at 5:58 AM, Daniel Marques <dani@wiris.com> wrote:

> Simply put, I am more concerned about which Unicode symbols to mirror
> under the dir=RTL attribute in MathML. It seems that the decision is
> already made (http://www.unicode.org/Public/UNIDATA/BidiMirroring.txt and
> http://www.unicode.org/reports/tr9), which includes the brackets. Yes, be
> careful with the "[BEST FIT]".
>
>
>
> Another question is at what tier this should be implemented. For example,
> whether in the MathML engine or in the Font file.  I assume that this is a
> collaboration of both. For example, the render engine should be aware that
> a Font (OpenType) has or has not a mirrored version of a symbol.
>
>
>
> Dani
>
> www.wiris.com
>
>
>