From: Donald Z. Osborn (firstname.lastname@example.org)
Date: Tue May 10 2005 - 00:09:44 CDT
The topic of keyboards has come up before so would be relevant. Keyboard
development is something that lots of folks have ideas about but few (not
including me) are specialized in.
From what I understand, the first step is a question of determining what is
needed, and this in turn responds to who the users are (IOW, what they need).
If I understand correctly, you're talking about something to facilitate input
of all unicode characters?
I still think that a technofix of a keyboard with LED keys which would display
the characters assigned to each key for each layout selected would be the
ultimate solution (I believe we discussed this on this list a while back).
There are LED keyboards available, but without the ability to change the
displays in the keys. At any rate, the physical durability of such keyboards
may not be as much of an issue as it was. The issue seems to be whether we can
make the next step to connect variable LED key displays with the information in
the software about which layout is active.
This is the dream. Let me digress a moment to an area that I have more
familiarity with. In the area of keyboards for Africa - with key & keycombo
assignments being the main appeoach - there was a tendency to think in terms
either of one keyboard that has everything (which gets complicated) or one
keyboard layout for each language (could be limiting). We're now talking more
often about layouts that meet needs for groups of languages used by a group of
users or in a geographical area (country, region). This is also complicated as
it involves some research and discussion in order to settle on standards. But
ideally it would meet user needs with standardized layouts that are predictable
and "harmonized." (The latter meaning that as you switch among layouts, most or
all characters common to the different layouts are in the same place.) Still,
there will have to be a learning & remembering of what plus what equals, say an
open-e, not to mention the problems of typing Arabic/Ajami, Tifinagh, N'ko or
whatever on a QWERTY or AZERTY keyboard. So this is still not ideal. Again the
LED keyboard seems ideal.
Since it is possible to have more than one keyboard layout installed on a
system, and for users to switch among them as necessary, perhaps one approach
to what it sounds like you are interested in (pardon if I'm off) would be a
suite of layouts, designed with some of the above concerns in mind and which
together would cover the unicode range. This might be conceived together with
regional keyboards such as for Africa, or separately. In any event, in an ideal
world, we would have LED keyboards that can tell us what the key values are in
the layout we've selected, and a suite of layouts that can cover the range of
Such a solution would obviate the need to use the Unicode character names in the
way you suggest. One problem with focusing on using the names, quite apart from
their correctness, is that users have to remember the name exactly as it is in
A-Z (and the names are not always short and easy to remember).
Quoting Hans Aberg <email@example.com>:
> Are Unicode keyboard input methods discussed on this list? The
> fastest method for a large character set (like Chinese), is
> supposedly that one types some identifier of the character, and the
> computer successively shows the possible matches. When one sees the
> right character appear, one selects it.
> Such an input method might relate to the discussion here about the
> correctness of character names. A Unicode keyboard input method, for
> standard keyboards, might be as follows:
> - First one hits the button, indicating that a new Unicode character
> is to be input.
> - One then types metacharacters A-Z plus space on the keyboard, and
> the computer displays possible character name completions, possibly
> with some other identifiers, such as a number or the Unicode
> character number.
> - When there is only one match, one hits "enter"; alternatively, one
> types the displayed number, and hits "enter".
> I think it is interesting to think about the method above, because it
> then becomes practically important to get character names and aliases
> correct. The method might also be refined, so that one can type say
> the name of a script, and get that implemented onto the keyboard.
> This would then admit quick script changes, without having to
> specially adapt the keyboard.
> Hans Aberg
This archive was generated by hypermail 2.1.5 : Tue May 10 2005 - 00:11:09 CDT