Re: Concerning the hyphen confusion in BiDi (was: Re: Bidirectional Bugs in Hebrew)

[Owen Taylor <otaylor redhat com>]

Tzafrir Cohen <tzafrir technion ac il> writes:

> Hi
> 
> On Fri, 10 Nov 2000, Pablo Saratxaga wrote:
> 
> > I don't think there is such a thing as "hebrew hyphen" (nor even "RTL hyphen")
> > But, on the other hand, there is "minus/dash" and "soft-hyphen" (and also
> > typographic "en dash" and "em dash"); only the "minus/dash" is a problem,
> > as it is not a neutral character like the others but is considered to have
> > the same properties as numbers (and possibly other mathematical symbols).
> 
> Hi
> 
> just one note:
> 
> in the previous thread in ivrix-discuss it was mentioned that soft-hyphen
> won't do the job:
> http://ivrix.org.il/mailing-lists/ivrix-discuss/2000/10/0052.html
> 
> IIRC unicode does have a seperate hyphen character. Any idea if most
> unicode fonts contain it?

The XFree86 fonts do. We can also make Pango display U+2010 (HYPHEN)
with the glyph at the code point for U+00AD (SOFT HYPHEN) or even 
U+002D (HYPHEN-MINUS) if there is no glyph for U+2010. Pango gives
great flexibility in this area.

So, is the right solution to simply have Hebrew keymaps have 
'minus' in LTR mode and a keysym corresponding to U+2010 
(I think the tentative standard is to put that at 0x1002010)
in RTL mode?

By making the appropriate modifications to the hebrew mapping in
Xlib's XIM support, we should be able to make sure that that keysym
still maps to 0x2d when doing direct conversion to iso-8859-8.

The main difficulty with this approach is that users may be confused
if HYPHEN and HYPHEN-MINUS look the same, but we can encourage font
designers to avoid this - even in latin typography hyphens should
generally be shorter than a MINUS sign, and may also be angled and/or
slightly serifed.

Regards,
                                        Owen