Re: Re: A product design project

The idea of word- or phoneme- based entry seems to escalate the issue of 
learning resistance. Consider phonetic court recorders, which have been 
around for a long time, and, as I understand it, are extremely efficient 
phonetic-based text entry systems. You would think they'd be in 
widespread use outside of courtrooms, but they aren't. As for word-based 
systems, have you ever seen a Chinese typewriter?
http://acc6.its.brooklyn.cuny.edu/~phalsall/texts/chinlng2.html
or, for a description of computer-based entry:
http://www.straightdope.com/classics/a5_177.html

A drawback to more "intelligent" word-based systems is word choice. 
Anybody who uses Microsoft Word probably has seen their words 
"corrected" as they type. This can seriously inhibit your word choice if 
you want to use a  brand name, technical term, or other word that the 
program doesn't understand. If the entry system itself is phoneme-based 
(that is, more rigidly than the alphabet is), I see no way to go back, 
when the system guesses the wrong homonym, and force it to use the word 
you want.

A rash person might suggest enforcing an artificial language that 
accommodates the vagaries of text entry. Try it. I gather the French 
can't even enforce French.

As for the ergonomics of Morse, remember that old-fashioned telegraph 
keys are indeed old-fashioned. Serious Morse practitioners use a paddle: 
pressing one side causes a stream of dahs to be emitted, pressing the 
other side issues a stream of dits. The action alternates from left to 
right and back. If you implement this using two keys on a cell phone, 
for instance, (or maybe a rocker) you could eliminate multiple 
same-finger keypresses entirely. One drawback is that you should have 
audio feedback, although you might implement some sort of tactile 
feedback (perhaps the key would "kick back" gently for each dit or dah 
emitted).

Morse isn't all that slow, either. For practiced operators (those who 
have learned not to think about it), laid-back morse-code sending speeds 
seem to run in the 25-30 WPM range (as I recall). Much faster rates are 
not uncommon within this (admittedly small, nowadays) group. Even I 
(with an old straight key) can send Morse faster than I can enter SMS text.

Further efficiency is gained through the use of "Prosigns" and "Q 
signs." These are brief two- and three-letter codes that represent 
common phrases like "end of message" (BT, run together), "send more 
slowly" (QRS), or "increase power" (QRO). There are also many accepted 
abbreviations, like "TNX" for thanks, "XYL" for wife, "73" for best 
wishes, etc. The code set is commonly known and rich enough to allow 
regular communication between operators who don't speak the same 
language. For same-language speakers, the newer IM abbreviations like 
LOL and whatnot would be just as effective in Morse. (Sorry, no smileys! 
The equivalent is HI, didididit-didit.) In practice, Morse goes farther 
than simple word-based entry by abbreviating common phrases, and one can 
always fall back to alphabetic entry when necessary.

The main advantage of Morse over a standard cell-phone keypad is that 
you don't have to look at the keys. This could also be true of your 
four-key system, perhaps, although it might not fit comfortably on a 
small cell phone. Of course, four to eight-key (more or less) chording 
systems have been tried several times (including, I believe, those court 
recorders), but they seem to require a bit more coordination than is 
common. At least, the idea always seemed more daunting to me than 
finding room for a full keyboard.

("Full keyboard," BTW, once meant a keyboard that had a full set of both 
upper- and lower-case characters. The Shift key had to be invented. In 
that sense, I guess ordinary keyboards _are_ chorded.)

Unless one learns a multi-stroke code or a chorded system, I don't see 
much difference between a low-count multi-key system and the current 
8-key cell phone text entry. In a sense, the current keypad system _is_ 
a multi-character code: A=2, B=22, C=222, D=3, E=33, etc.

Marcus

matzkelon wrote:
> I agree about morse being intriguing and fun to work with, but the
> concept of efficiency is not addressed by it. If anything - sequences
> should symbolize words or phonemes, not single letters. 3 or 5 key
> presses are hell as far as RSI is concerned, and is pretty slow even
> if you do it fast. Quikwriting is an example of what I mean. In
> quikwriting a single stroke denotes a word, not a single letter and is
> easier to perform and remember (even though it is stylus based, it
> should not be ignored).
>
> However, the concept of morse should be considered. maybe a sequence
> of lets say 9 keys could denote words and letters. Imagine the PS
> controller's 4 keys. instead of . and - performed through short and
> long presses you could have a set of 4 shapes (triangle, square,
> circle, .... X) with no dependency of keypress length. The sequences
> could be much more advanced this way.
>
> As a final word, I'd say there are good alternative methods for typing
> which should be considered first - but I do not reject the idea of
> thinking up new ones.
>
> Thank you for being an active partner. I hope to get more responses
> from other people as well soon.
>
> -Eyal (dum, dee, dum) :)
>
> --- In [email protected], Marcus Brooks <marcus@...> wrote:
>   
>> Something that you may not have considered is Morse code. This would
>>