On 16 Jul 2015, at 11:30, I wrote: > the compiler admits clear characters only up to U+008F.
Up to U+007E, of course. On 16 Jul 2015, at 10:35, Hans Aberg wrote: > One still has to figure out a good map. Yes this is the primary issue for every newly encoded script, and it remains important with respect to ergonomics. I just wanted to say that I'm focussing on the Compose tree of a Latin keyboard layout. Do you mean that the US American English keymap should be thoroughly reengineered too, additionally to the solutions of ANSI, ISO, and August Dvorak? I think that on the ANSI/ISO keyboards it would be sufficient to remove the dead keys, change T29/E00 from 0x0060 to 0x02bc, and replace VK_RMENU with a Compose key. It's a bit more complicated however to get a simple *and* complete keymap for France, and surely a number of other countries using diacrited characters. Marcel
