> -----Original Message-----
> From: Shachar Shemesh 
> [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, February 11, 2003 10:07 AM
> Cc: [EMAIL PROTECTED]
> Subject: Re: xkb handling in upcoming 4.3.0
> 
> 
> Well, I beg to differ.

I suspect the bottom line is the same result, only it sounds like you
know this stuff.

> 
> I'm not sure what 4.3.0 did, but the correct behaviour should be that 
> shift-9 produces "open brackets" in left to right languages, 
> and "close 
> brackets" on right to left languages. shift-0 should produce 
> the other one.

I suppose that's a more technically correct way to describe the result
I want.  So it sounds like so far we actually agree.

> 
> As far as the engravings go, it is up to the display engine 
> to display 
> open brackets as left brackets in LTR context, and right 
> brackets in RTL 

Ah - that might be the key to our little "disagreement" - the keyboard
is supposed to reverse it once and then the display is supposed to reverse
things again.  As long as the end result is:
1. I press Shift-9 to get a left brace etc...
2. The generated document contains that correct character.
3. The final display is correct (e.g. in some sites I saw braces which
   kept switching on every edit).

then I'd be glad with it, and even more so happier to know we conform
to a standard.

> context. We can see, then, that we need cooperation of the keyboard 
> layout and the display engine in order to get a consistant 
> and pleasant 
> experience.
> 
> The situation thus far was not very good in X, but was tolerable. 
> Shift-0 in RTL context produced the wrong key (i.e. - close bracket 
> instead of open bracket), but this was compensated by the 
> fact that the 
> display engine did not know it was supposed to reverse them. This is 
> called, in Technical jargon "Visual Hebrew". I think I don't need to 

Again (if I understand this) - you mean that the new way is simply to teach
both the keyboad and the display that they are supposed to reverse, right?

I can't check your suggested test right now (no desktop Linux at work :-(),
it sounds like we agree on the wanted end user experience and that you know
what you are talking, so I am relaxed about this now.

Cheers,

--Amos

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]


Reply via email to