Path: utzoo!attcan!uunet!lll-winken!ames!mailrus!cornell!rochester!rit!ultb!jdb9608
From: jdb9608@ultb.UUCP (J.D. Beutel )
Newsgroups: comp.society.futures
Subject: Re: Keyboards - input devices
Summary: yea, how about a software-driven solution?
Message-ID: <292@ultb.UUCP>
Date: 15 Jan 89 23:33:50 GMT
References: <19626@uflorida.cis.ufl.EDU> <315@s1.sys.uea.ac.uk>
Reply-To: jdb9608@ultb.UUCP (J.D. Beutel (713ICS))
Organization: Rochester Institute of Technology (Info Systems)
Lines: 54

In article <315@s1.sys.uea.ac.uk> jrk@uea-sys.UUCP (Richard Kennaway) writes:
>
>A few years ago, in one of the computer rags, I saw an article about such a
>keyboard.  It had the keys laid out so as (it was claimed) to fit the hands
>better.  Instead of being in straight rows on a plane, they were arranged
>roughly in two clusters around the positions of the hands, in such a way
>as to minimise the amount of hand and finger movement required.  Not a
>straight line in sight.  It looked like something designed by Dali.  I cant
>remember any more details.  Does anyone recognise this?  Was it ever
>commercially produced?
>-- 
>Richard Kennaway                SYS, University of East Anglia, Norwich, U.K.


Yes, I remember something about that in some magazine some years ago
(altho I'm *some*what lacking in really useful information).
The two clusters of keys were oriented at an angle.  Simply bending
both ends of the keyboard away from the typist in a V-shape was said to
improve the typist's comfort because when you bring your hands together 
to type (as I stare at my hands now to verify this datam) they form an
angle of about 30 degrees, as opposed to the 0 degree angle imposed by
a normal keyboard with its keys that are situated in a straight line.
Such a design should be a better ergonomically whether or not the
keys are in a qwerty arrangement.

Speaking of different keyboards, my Atari ST (and perhaps other computers as
well) has a keyboard map that allows the user to assign any value to any key.
I could make mine a Dvorak keyboard, but then when I went to the computer lab
I would have to use qwerty again--it would be too confusing; I can barely
handle keeping track of the ctrl key being different as it is now.

A solution would be to have a tty driver that will map keys.  The user could
define what map he or she wants to use, like a termcap, and then no matter what
terminal he or she uses, the keyboard is whatever he or she wants.
No hardware would be affected, and the solution would allow a uniform and
personal choice of a Dvorak or other such change-the-names-of-the-keys-but-
don't-actually-move-any-around type keyboard.  

I think such a tty driver would need only two changes from ones used now:
	1)	enable it to read a key map
	2)	any char it recieves is used as an offset into the key map to
		determin the char that is passed along--i.e., translate

A problem would be that reading a key map from disk every key stroke would be
very inefficient, so the key map would need to be kept in memory with the tty
device driver.  However, since each user might have a different key map,
the device driver would need an array of key maps (less than or equal to 
128*2 bytes per user record), or perhaps a limited number of pre-defined
keyboard types corresponding to a user-modifiable minor device number?

Are there any Un*x gurus out there that would care to comment on the 
feasability of such an improvement to the tty driver?
-- 
11011011_____jdb9608@ritcv.UUCP_____prefered==>__jdb9608@ritvax.BITNET