What a nice distraction this thread is.
Localizing hex input to different alphabets misses a point - I have not
seen it mentioned in my somewhat random sampling of this thread, so I'll
make it here.
Observing that:
* Hexadecimal input is a mapping of a KEY sequence to a Unicode Scalar Value.
* The Key sequence needs to be correlated to the code tables so that the
users can use the code tables to know what sequence will give which character.
* The only officially sanctioned code tables are published using 0-9A-F.
* Practically anyone, anywhere in the world, has (to have) the ability to
type letters A-F anyway, and all physical keyboards that I have seen, allow
this with an appropriate mode.
Therefore:
* Since the hexadecimal input method can work directly with the KEYs for
A-F, even if the keyboard is not switched to that mode (the Ctrl-Shift is
enough of an announcer to override whatever shift mode the keyboard is in)
it can be programmed to work with whatever keys can be used for A-F for a
given native keyboard.
As an aside it's fun to observe that the method is actually harder to
program for Latin keyboards. Non-Latin keyboards usually provide the
English or qwerty layout for Latin letters, so they would less of a problem
than some of the non-English layouts. After all the French do want to use
A-F and not Q-F ;-). So for the latter case the input could not work with
the raw keys but would need to be cognizant of the (Latin) layout.
I have used jury-rigged software that used similar input schemes and they
are very handy for anybody who needs to produce a particular code, without
knowing the language or scripts and its customary input mechanisms.
Programmers and Software testers are first in line for this, of course, but
possibly in the future editors and some data entry (if we ever get to be
multilingual to such a degree). And users of ML URLs - that had been
mentioned. For some of the latter cases it might make more sense to provide
a more powerful input method that allows selection by sound, or shape
similarities.
A./
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:46 EDT