[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Adding Bidi to Fvwm ?

On Sun, 24 Feb 2002 02:20:09 +0330 (IRT)
 "Roozbeh Pournader" <roozbeh sharif edu> wrote:
> On Sat, 23 Feb 2002, Nadim Shaikli wrote:
> > Sounds like someone intimately involved with the Bidi libraries should
> > bring this up to the unicode folk for review so as to modify unicode's
> > underlying algorithm and requirements, no ?
> Definitely no. I will support your proposal if you bring this to Unicode, 
> but I already know what they will say: that changing the Bidi algorithm 
> is very hard to do, because every implementation will need to be modified.

Irrespective of difficulty and of Unicode's bureaucracy - all wrongs should
be righted.  With Unicode's current algorithm, Bidi will not handle
parenthesis in RTL mode and that is simply not acceptable.  I'm not taking
a harsh line, I have no problem informing Unicode of this short-coming (given
an email address and a contact person unless its the good'ole unicode
mailing-list), but I think it would carry more weight coming from a Bidi

  % echo "this MUST (WORK)" | ./fribidi --test --charset CapRTL
  this MUST (WORK)	=> this KROW) TSUM)

A reply to Unicode's potential comment,

  "changing the Bidi algorithm is very hard to do, because
   every implementation will need to be modified"

could be along the lines of -- that holds true for any oversight.  Remember
the old Pentium divide bug ?  If a company decides to ignore a bug due to
"difficulty" or cares not to take responsibility for it, it would certainly
bring that company's demise from its users.  The same should hold true in
this instance.  Unicode's algorithm should not be set in stone (unless
they've been proven to be 100% correct - which is almost impossible :-),
thus the algorithm should adapt to new issues as they come up, no ?  If this
algorithm is to be used world-wide, it needs to adopt and fix its own bugs.

 - Nadim

Do You Yahoo!?
Yahoo! Sports - Coverage of the 2002 Olympic Games