On Thu, Jun 20, 2013 at 09:27:49AM +0100, Michael Everson wrote:
> On 19 Jun 2013, at 18:24, Richard Wordingham
> <richard.wording...@ntlworld.com> wrote:
> > The X11 restriction of one character per key stroke is not so easy to get
> > round.
>
> Get them to fix X11.
It looks like you think that X11 is broken. Why?
I did not yet try to fully convert my keyboard layout (see
unicode.org/mail-arch/unicode-ml/y2013-m02/0053.html) to X11, but the
limited experiments I did do not show any “brokenness” of X11. It has
its limitations, but so does any system. (Maybe MacOS is different
and gives the author full control — again, I did not try to explore it
fully.)
My working model is as follows (I'm not sure that it true to the last
dot, but it definitely is a useful even if it is a simplification):
a) You do not assign any characters at all to “strokes”. What you
assign is a keysymbol. There may be up to 8 keysymbols assigned to
a key (corresponding to state of modifier keys — and may be
something else; my understanding of the full semantic is not
complete).
A keysymbol is just an integer. Some of them are in in 1-to-1
mapping with unicode characters. Some are extras (e.g., dead_currency).
b0) There is a default mapping of keysymbols to strings. It might
be indeed that this default mapping has only strings of length 1.
b1) Keysymbols (or keysymbols with modifiers — those modifiers which
were NOT used on the step [a]; or sequences of keysymbols) can be
translated to other keysymbols or to arbitrary strings.
IMO, this scheme is flexible enough to cover most of usage scenarios.
(Well, maybe not flexible enough to cover what my keyboard is
currently doing on Windows — but my keyboard is probably the most
advanced one of those using the standard Windows keyboard
interface — the only thing missing now is the COMPOSE key.)
Am I missing some limitations?
Ilya
