Peter T. Daniels wrote:
> > The second inaccuracy concerns the use of the term
> > "right-to-left". Although
> > the majority of the text will be written right-to-left,
> > numbers are still written left-to-right (LTR).

This only applies to computer character sets, such as Unicode, where digits
in numbers are always encoded with most-significant digit first.

But, in handwriting, numbers can be written either LTR or RTL, provided that
the resulting sequence is visually the same.

E.g., if I want to write one hundred twenty three:

1
12
123

or

3
23
123

> > In addition, right-to-left text will often
> > include borrowed or foreign words written in their native
> > left-to-right script [...]
>
> What a lame excuse. If a word in another script is dropped into the
> first script, the first script doesn't become "bidirectional"! If it
> did, why wouldn't the same apply to a passage in English with a Hebrew
> word dropped in, as is often found in discussions of biblical text?

I agree: this use of "bidirectional" doesn't make sense out of the realm of
computer technology.

Moreover, even in a technology context, it is misguiding to say that a
*script* (or even less so a language) is "bidirectional". It is a character
*encoding* which can be bidirectional or not.

E.g., Unicode and ISO 8859-6 (8-bit Arabic) are bidirectional character
encodings, because they allow to mix in a text file both LTR and RTL
characters, and both kinds of character are in "logical order" in the
computer memory (i.e., in the same order as they are written with a pen).

But *any* text encoded in a bidirectional encoding is virtually
"bidirectional". Even a Unicode text file completely written in English is
bidirectional, because no one forbids me to edit it and add an Urdu word in
it...

The only real "bidirectional scripts" are ancient writing systems such as
Egyptian hieroglyphics or the Etruscan alphabet, which could be written
either RTL or LTR, by mirroring all the non-symmetrical signs.

_ Marco