RE: BIDI: possible fix

From: Marco Cimarosti (marco.cimarosti@essetre.it)
Date: Tue May 22 2001 - 09:40:32 EDT


Roozbeh Pournader wrote:
> Well, I have got convinced now that the algorithm should not
> be change for
> this. But please note that if you want to underline the whole
> number, the
> separator with the digits around it, you still need the change in the
> algorithm to make it right.

Why?

        
<LRE>1<CIRCLE>2<CIRCLE>3<CIRCLE>,<CIRCLE>4<CIRCLE>5<CIRCLE>6<CIRCLE><PDF>

Is not less left-to-right than:

        <LRE>123,<CIRCLE>456<PDF>

> > [...] re-shipping all Windows NT, Linux, or Mac OS [...]
>
> Do you really believe that all these are compliant?

Touché. But you think that they will invest time to implement your circled
separators just for being compliant?

> > But I think that Roozbeh is very interested in
> interchanging text in a
> > standard way so, apparently, a higher-level protocol is not
> the way to go.
>
> Yes. I want to avoid higher-level protocols as much as
> possible. If you
> were living in a country that has about 5 or 6 common
> document formats,
> each having a different codepage and its own considerations about the
> script, you would understand that why I want nothing other than
> standards...

But you missed the second part of what I said: the suggestion is to use a
higher-level protocols only *inside*, and convert it to *standard* BIDI
controls in external files.

I think that an application can safely use whatever private encoding or
protocol -- as far as it is able to convert it to standard data when
speaking to the outer world.

In the specific case, you can define your own version of the bidi algorithm
that behaves as you described.

The only important thing is that the abridged algorithm is able to convert
to the standard conventions before saving/transmitting the text (i.e.,
detects the non-standard usage and standarifies it by adding extra BIDI
controls).

(BTW, I live in such a country in a sense... If you peep in Italian
newsgroups, you will notice that most people avoid the local characters à è
é ì ò ù « » in favor of a' e' e' i' o' u' " ", to avoid problems with the
different character sets in use. Luckily enough, the rest of the Italian
alphabet is in ASCII.)

_ Marco



This archive was generated by hypermail 2.1.2 : Fri Jul 06 2001 - 00:18:17 EDT