Re: [tlug] OT. Linguistic ?

["Mauro Sauco" <mauro@example.com>]

HI Stephen,

On 2005/05/06@example.com:29, <Stephen J. Turnbull> wrote:
  Stephen> Wrong question.  On the technical side, you can always string together
  Stephen> multiple words, even for West Greenland Eskimo.  On the human side,

Yes , I've looked into that possibility (using phrases).

  Stephen> the problem is, how are you going to get people to remember 20,000

Humans remember a lot more than that when learning any language.

  Stephen> random numbers (base 12, yet)?  The average human being can't remember
  Stephen> one PIN for their bank card....

What I didn't do was to explain how the input device works and maybe
that is why you think you'll have to remember numbers. The Idea is
to have a glove with 12 contacts, 3 in each finger, and a common, the
thumb, then the words are permutations of 4 "contact" lenth. I made
the first prototype with alumine paper on a glove, but I am planning
(tomorrow) to get conductive polycarbonate gome to make better more
flexible glove. BTW, I found how to make 16 contact by putting contacts
on the nail side of the finger as well then the number of possible
permutations are 65,535.

  Stephen> If you've got a PDA, why not use its native input capabilities, and
  Stephen> just "spell it out"?  Allow abbreviations (a la keitai messaging) and
  Stephen> get a predicting parser (see below).

1. Inputing full lenth words using a keyboard will never be fast
enough for real time speaking (combinations of 4 finger tips
contacts are really fast) With a couple of minutes of practice , two
deaf people at a rehabilitation center were quickly getting a couple
of phrases ready, and they looked very exited BTW.

2. The device is not only intended for people with speaking
disabilities but also for people with very limited mobility,
an standard keyboard will require your both hands.

3. using less keys for writing is already proven to be faster than
an standard keyboard as you can see in "taquigrafia" (sorry only
know the word in Spanish), is the typewriters they use in court that
only has 14 keys.
http://www.esteno.net/images/estenotipia.jpg

4. Using a glove will give you the freedom to use your hands while
the PDA can be into a pocket.

  Stephen> How inexpensive?  A Zaurus SL-C7xx or SL-C8xx is going to be in the
  Stephen> neighborhood of 5-man-en real soon now, and you can probably picked up

I've seen the SL5500 (can only get it used) between 7,000 and 15,000 yen
in sofmap mobiles in Akiba, anyways it doesn't have to be a Zaurus. When I said 
inexpensive, I meant to say, reachable ... for instance there is a sign language 
interpreter, but requires cameras, motion detectors, and  lot more, making it 
unreachable. Say 50,000 JPY is still very affordable for most Japanese families.
A database of 65,000 words was 11MB filesize using sqlite. I made a test with an 
existing dictionary and a couple of randomly generated fields). Still needs to 
make more tests to find out if the SL5500 can handle festival (CMUflite), sqlite 
and python on real-time, but I am very convinced it will.

Back to " how to remember 20,000": there is a big difference between
remembering numbers and moving your fingers to convert thoughts into
words, the motor-memory is a very efficient system that does not
required the same type of memory (implicit non declarative memory) than remember 
telephonic numbers ( explicit declarative memory ), much like the way a pianist 
can play a song by hart without having to remember the exact combinations of keys 
that are required (explained by a doctor of the rehabilitation center). I found 
something here:
 http://www.thebrain.mcgill.ca/flash/i/i_07/i_07_p/i_07_p_tra/i_07_p_tra.htm.

Some studies also shown that motor-memory when used for communication
porpoise will make your brain storage and read from a different
part of of the brain also you'll skip certain processes that are necessary
the remembering other type of things. , like in sign language:
http://faculty.washington.edu/chudler/sign.html

I have seen some Pakistani kids using a very particular way of
counting with their fingers and they move their fingers very, very
fast making really quick calculations (just a comment).

I believe that learning 20,000 combinations is not a big issue as
in "taquigrafia" people normally learn it as adults and they
certainly are required to have a very extensive vocabulary for use
in court, maybe a lot bigger than 20,000. ( not words, but combinations 
of letters/keys, same a s my glove)

Then there is the practical factor that gave me a reason to start
this. I gave this toy to some kids in a rehabilitation center and
they WANT IT, they were so exited and the doctors as well.

Anyways, I can't for sure said that this will be a success until a
lot more testing is done and more people try this out. But I believe
it is worth the try. Seeing this people wanting such gives me a lot of 
enthusiasm to go ahead even if it wont be feasible after all.

Hope all this makes sense, I have problems explaining this in
English, but hope it is somehow understandable :). I will send some
pictures sometime soon.

BTW, I also got a lot off feedback of-list, thanks a lot.

//mauro//