L2/07-273

This is a quick document outlining some of the cases on the board for discussion tomorrow.

(I don't have mirrored glyphs, so am using ˎˎ as a proxy for mirrored „. The cases that are not accepted practice are marked with red  (and dashed outline, for black and white printing).)

The first chart is taking data as it normally would be typed in particular languages, and looks at the behavior in a RTL environment.

The second chart shows data that is modified so that it looks right in a RTL environment under different environments. Note that correct English and Swedish display in a RTL environment can be simulated by using different characters. That cannot be done for German and Swedish. The LRM character can always be used with the original data; if it is used, then 5.0 and 4.1 don't have a difference.

"Compensated"	Data	LTR Environment	RTL Environment
		Both 5.0 and 4.1	Mirrored (5.0)	Unmirrored (4.1)
English	→”ab“→	→”ab“→	←”ab“←	←“ab”←
	→“ab”→	→“ab”→	→“ab”→	→“ab”→
Swedish	→“ab“→	→“ab“→	→”ab”→	→“ab“→
	→”ab”→	→”ab”→	→”ab”→	→”ab”→
German	→„ab“→	→„ab“→	→„ab“→	→„ab“→
Hungarian	→„ab”→	→„ab”→	→„ab”→	→„ab”→