Project idea: invisible chording keyboard
Stuart Langridge
whether it's rational or science-fiction, and if it's rational I'll
write it up on the wiki :-)
Put your left hand out, flat, palm up, and then grip tightly around
your left wrist with your right hand (make a circle around your left
wrist with your right thumb and middle finger). Now, move the fingers
of your left hand. You can feel tendons and muscles and whatnot moving
in your wrist, yes?
So, the idea is: make a wristband containing sensors which is
sensitive enough to differentiate these movements, so the wristband
can tell which finger you're moving and by how much. Give that
wristband a battery and a bluetooth HID chip. At that point, it's a
bluetooth keyboard, and it's just a wristband. That makes it a
chording keyboard -- for example, bending your index finger and thumb
together might mean "A", index finger and right finger might mean "B",
and so on -- but because it's just a wristband, it's invisible. No-one
knows you're wearing it. You can use it while you're walking down the
street if you want. Imagine how fast you could type SMSes with it; how
small you could make a laptop that doesn't need a keyboard.
So, is this halfway rational as a buildable thing?
Issues I see with it:
1. It'd almost certainly need to be individually calibrated for each
person. This isn't necessarily a problem, though; I'm imagining that
every morning when you put it on, it beeps, you then bend each finger
in turn in some pre-defined movement pattern to calibrate it
2. I'm thinking of it looking like one of those Make Poverty History
wristbands, i.e., not big. If it looks like you've got a doughnut
wrapped around your wrist no-one will wear it. Is it even halfway
possible to get (a) sensors (b) a power source that'll last more than
two minutes (c) bluetooth into something this small?
3. How do you recharge it? Just taking it off and dropping it on an
induction pad would be the coolest.
4. Is one hand enough for a chording keyboard, or would you need a
wristband on each wrist?
5. Can sensors differentiate finger movements enough? In particular,
there's "level 1" where you can tell the difference between "finger
bent" and "finger straight", and maybe "level 2" where you can tell
the difference between "finger straight", "finger a bit bent", and
"finger a lot bent", which greatly increases the number of gestures
you can do with one hand.
Comments on whether this idea is possible are gratefully invited.
sil
--
New Year's Day --
everything is in blossom!
I feel about average.
-- Kobayashi Issa
nikki
One thing that strikes me is the sheer number of different positions/
movements/combinations you'd have to be able to a) remember and b)
distinguish between with the sensors in order to get even an alphabet,
let alone all the punctuation etc.
If you had a system similar to texting on a mobile phone where you
have to repeat a movement say 3 times to scroll through to the letter
you want does that make it simpler, or just too klutzy to use?
Do gloved keyboards exist, or have I just been watching too much
telly? How do they work? Is it through using different combinations
that they get all the different letters covered?
Would an interesting pilot project be to control a game with just half
a dozen or so commands?
How sensitive will the sensors be? Would you end up having to make
huge finger movements for them to register? Would it be tiring to use?
Anyway, let's find out!
nikki
G Bulmer
> whether it's rational or science-fiction, and if it's rational I'll
> write it up on the wiki :-)
>
> Put your left hand out, flat, palm up, and then grip tightly around
> your left wrist with your right hand (make a circle around your left
> wrist with your right thumb and middle finger). Now, move the fingers
> of your left hand. You can feel tendons and muscles and whatnot moving
> in your wrist, yes?
>
> So, the idea is: make a wristband containing sensors which is
> sensitive enough to differentiate these movements, so the wristband
> can tell which finger you're moving and by how much. Give that
> wristband a battery and a bluetooth HID chip. At that point, it's a
> bluetooth keyboard, and it's just a wristband. That makes it a
> chording keyboard -- for example, bending your index finger and thumb
> together might mean "A", index finger and right finger might mean "B",
> and so on -- but because it's just a wristband, it's invisible. No-one
> knows you're wearing it. You can use it while you're walking down the
> street if you want. Imagine how fast you could type SMSes with it; how
> small you could make a laptop that doesn't need a keyboard.
>
> So, is this halfway rational as a buildable thing?
If the question is "could Nintendo, Sony, Intel, Microsoft, ... build
this?" the answer is yes.
If the question is "could I build one for less than £200?" the answer
may be "not yet" because the force sensors needed aren't available and
would have to be custom fabricated after a research project.
>
> Issues I see with it:
>
> 1. It'd almost certainly need to be individually calibrated for each
> person. This isn't necessarily a problem, though; I'm imagining that
> every morning when you put it on, it beeps, you then bend each finger
> in turn in some pre-defined movement pattern to calibrate it
I imagine it'd need a good amount of computer power just to cope with
the movement between wristband and wrist throughout the day. Even as
you walk down the street it will tend to move, and you don't want to
constrict blood flow.
I think it is feasible to sense movement across a bunch of sensors,
then it is a sophisticated piece of computation to recognise the
pattern and figure out what it means, then it gets worse as the
wristband moves around.
> 2. I'm thinking of it looking like one of those Make Poverty History
> wristbands, i.e., not big. If it looks like you've got a doughnut
> wrapped around your wrist no-one will wear it. Is it even halfway
> possible to get (a) sensors (b) a power source that'll last more than
> two minutes (c) bluetooth into something this small?
Low-power (class 1) bluetooth might help, or a lower power radio with
an adapter on your phone.
To be fare, I am thinking power for the sensors and computer may
dominate unless it can be powered down a lot.
How would it get turned on and off? If there is an acceptable way to
power on/off, battery life could be extended.
Maybe that is all the other wrist does; power on/off?
> 3. How do you recharge it? Just taking it off and dropping it on an
> induction pad would be the coolest.
> 4. Is one hand enough for a chording keyboard, or would you need a
> wristband on each wrist?
This product was on sale in the early '80's:
http://en.wikipedia.org/wiki/Microwriter
I tried one, and a friend had one.
He loaned it to a very good shorthand typist at his work while he went
on summer holiday, and in two weeks occasional practice, she got about
as fast as she could do shorthand or type.
It only has six buttons, so you just need two position for e.g. the
thumb, and your done.
> 5. Can sensors differentiate finger movements enough? In particular,
> there's "level 1" where you can tell the difference between "finger
> bent" and "finger straight", and maybe "level 2" where you can tell
> the difference between "finger straight", "finger a bit bent", and
> "finger a lot bent", which greatly increases the number of gestures
> you can do with one hand.
recent. Keep gestures for version 2 :-)
You might be able to get that level of discernment, but try to keep it
simple to start with; I think detecting finger movement *reliably* is
the hardest part to start with.
Also, it is NOT necessary for the microwriter chord system for
anything other than the thumb (I learned to type "The quick brown fox
jumps over the lazy dog" in less than 5 minutes, so it worked quite
well).
>
> Comments on whether this idea is possible are gratefully invited.
A simplified 'data glove' would work. You only need one movement per
finger and two for the thumb.
It wouldn't need to be anything like as sophisticated as a data glove
(http://en.wikipedia.org/wiki/Data_glove)
A small flat-ish plate shaped to the palm of your hand (maybe in a
fingerless glove) would be easier to make than a wristband, and you
could experiment with force sensors, and the whole chord typing
system.
There are affordable force sensors you could try out, e.g.:
http://www.interlinkelectronics.com/force_sensors/products/forcesensingresistors/index.html
It might be worth looking at the technology used for modern prosthetic
hands which use electrical signals to detect movement. I accept that
you don't need such a solution, but, I assume those folks have solved
even more complex pattern-matching problems, and so you could gauge
the upper-bound on what kind of computing power is needed.
Another thought is to make something that would use close to real sign-
language (the spelling form):
http://www.aspexdesign.co.uk/bsl.htm
then you might be able to use accelerometers in two wrist bands, and
conduction between fingers of the hands.
It would be delightful to have folks walking down the street texting
each other in a system of language that people who are regarded as
'disabled' could read easily :-)
Plus, you'd have a ready group of users for your creation.
Garry
Stuart Langridge
> movements/combinations you'd have to be able to a) remember and b)
> distinguish between with the sensors in order to get even an alphabet,
> let alone all the punctuation etc.
Indeed. Chording keyboards that do this kind of thing exist, certainly...
> If you had a system similar to texting on a mobile phone where you
> have to repeat a movement say 3 times to scroll through to the letter
> you want does that make it simpler, or just too klutzy to use?
Certainly makes it simpler to learn, if not to use :)
> Do gloved keyboards exist, or have I just been watching too much
> telly? How do they work? Is it through using different combinations
> that they get all the different letters covered?
Yep. (Well, gloves, don't know, but chording keyboards, yes.
http://en.wikipedia.org/wiki/Chorded_keyboard is instructive here.)
> How sensitive will the sensors be? Would you end up having to make
> huge finger movements for them to register? Would it be tiring to use?
This is one of the big things I don't know, and I don't know how to
find out either :) Some comments on where I should go to answer this
question are gratefully invited :)
Stuart Langridge
> I think the question may be imprecise.
> If the question is "could Nintendo, Sony, Intel, Microsoft, ... build
> this?" the answer is yes.
> If the question is "could I build one for less than £200?" the answer
> may be "not yet" because the force sensors needed aren't available and
> would have to be custom fabricated after a research project.
Ah that's what I don't know -- are the sensors available? I entirely
agree that if custom sensors have to fabbed from scratch then I am Out
Of Luck. But I don't know...
> How would it get turned on and off? If there is an acceptable way to
> power on/off, battery life could be extended.
The way I think that'd work is: clench your fist three times in under
a second to toggle it on/off. This means that "off" would actually
mean "turn the bluetooth off" rather than "turn the sensors off", but
I bet that'd be the main source of powerdrain anyway :)
>> 4. Is one hand enough for a chording keyboard, or would you need a
>> wristband on each wrist?
> No, one hand is enough for the full ASCII character set.
> This product was on sale in the early '80's:
> http://en.wikipedia.org/wiki/Microwriter
> I tried one, and a friend had one.
> He loaned it to a very good shorthand typist at his work while he went
> on summer holiday, and in two weeks occasional practice, she got about
> as fast as she could do shorthand or type.
>
> It only has six buttons, so you just need two position for e.g. the
> thumb, and your done.
Sort of. I looked at the Microwriter. There are a few flaws, like, I
can't bend my little finger without also bending my ring finger, which
I think is pretty common :-)
> It might be worth looking at the technology used for modern prosthetic
> hands which use electrical signals to detect movement. I accept that
> you don't need such a solution, but, I assume those folks have solved
> even more complex pattern-matching problems, and so you could gauge
> the upper-bound on what kind of computing power is needed.
Heh. These are called "myoelectric sensors", and it's jolly difficult
to find information about them, because it's all proprietary patented
medical technology being sold at £10,000 a pop. Anyone who knows
anything about this, I'd love to hear about it!
> Another thought is to make something that would use close to real sign-
> language (the spelling form):
> http://www.aspexdesign.co.uk/bsl.htm
> then you might be able to use accelerometers in two wrist bands, and
> conduction between fingers of the hands.
huh? Conduction between fingers? What measures this?
sil
Trevor White
common practices but I wonder if you could make little sensors that
measure the change in capacitance. Maybe the tendon and muscle movements
change the local capacitance?
There are some touch sensors around that calibrate to there environment
and then detect changes in the capacitance caused by for example,
fingers coming near. The technology has been around for a few years. I
actually have some ICs at my office. I used them for touchless buttons
through perspex. They are very sensitive and work very quickly.
I have no real idea if they would work for what you are looking at but
it might be a great group experiment to have a look. It would certainly
be interesting.
Trev
G Bulmer
On Jul 9, 11:12 am, Stuart Langridge <s...@kryogenix.org> wrote:
> >> ,,,
> > ...
> Ah that's what I don't know -- are the sensors available? I entirely
> agree that if custom sensors have to fabbed from scratch then I am Out
> Of Luck. But I don't know...
effort.
>
> > How would it get turned on and off? If there is an acceptable way to
> > power on/off, battery life could be extended.
>
> The way I think that'd work is: clench your fist three times in under
> a second to toggle it on/off.
> This means that "off" would actually
> mean "turn the bluetooth off" rather than "turn the sensors off", but
> I bet that'd be the main source of powerdrain anyway :)
of the other radio's are lower power.
>
> >> 4. Is one hand enough for a chording keyboard, or would you need a
> >> wristband on each wrist?
> > No, one hand is enough for the full ASCII character set.
> > This product was on sale in the early '80's:
> >http://en.wikipedia.org/wiki/Microwriter
> > It only has six buttons, so you just need two position for e.g. the
> > thumb, and your done.
>
> Sort of. I looked at the Microwriter. There are a few flaws, like, I
> can't bend my little finger without also bending my ring finger, which
> I think is pretty common :-)
two neighbouring fingers move a bit too.
That effect is why I think it is a fairly sophisticated pattern-
recognition problem, with a need for reasonable amounts of
computation.
It will not be possible to get a 'clean' ring finger only signal for
sensors. Instead I think a bunch of sensors will change, and the
computational problem is to work out what that actually means.
>
> > It might be worth looking at the technology used for modern prosthetic
> > hands which use electrical signals to detect movement. I accept that
> > you don't need such a solution, but, I assume those folks have solved
> > even more complex pattern-matching problems, and so you could gauge
> > the upper-bound on what kind of computing power is needed.
>
> Heh. These are called "myoelectric sensors", and it's jolly difficult
> to find information about them, because it's all proprietary patented
> medical technology being sold at £10,000 a pop. Anyone who knows
> anything about this, I'd love to hear about it!
I have not looked, but I had thought some papers would be published in
medical conferences, and patents filed.
Patents are not designed to be easy to understand, but they do need to
carry enough information that their mechanism can be understood.
>
> > Another thought is to make something that would use close to real sign-
> > language (the spelling form):
> >http://www.aspexdesign.co.uk/bsl.htm
> > then you might be able to use accelerometers in two wrist bands, and
> > conduction between fingers of the hands.
>
> huh? Conduction between fingers? What measures this?
*very* low-power 'radio' signal across the body from one wrist band to
the other. When the fingers of each hand comes into contact, the
signal path would change dramatically, and that's what would be
detected. It might be simpler than it sounds.
Garry
nikki
replying before!
Can I just ask what the overall objective for the project would be?
Is it to make something small; something that's as invisible as
possible to use (ie small movements), to make an interesting interface
for a particular application; or to get to grips with a particular
type of technology.... etc
Being a bit more specific with what you want to get out of it might
naturally point the way towards particular solutions and/or highlight
which are the meaty problems to tackle.
nikki
G Bulmer
On Jul 9, 12:04 pm, nikki <genzai...@googlemail.com> wrote:
> ...
> Can I just ask what the overall objective for the project would be?
> Is it to make something small; something that's as invisible as
> possible to use (ie small movements), to make an interesting interface
> for a particular application; or to get to grips with a particular
> type of technology.... etc
>
> Being a bit more specific with what you want to get out of it might
> naturally point the way towards particular solutions and/or highlight
> which are the meaty problems to tackle.
>
> nikki
Supplementary, what are the priorities, and what is 'non-
negotiable' (i.e. if it doesn't do this, it isn't worth doing).
For example, is it enough to do no more than choose between 2
telephone numbers and dial?
GB-)
Stuart Langridge
> Is it to make something small; something that's as invisible as
> possible to use (ie small movements), to make an interesting interface
> for a particular application; or to get to grips with a particular
> type of technology.... etc
>
> Being a bit more specific with what you want to get out of it might
> naturally point the way towards particular solutions and/or highlight
> which are the meaty problems to tackle.
Heh.
Short term goal: typing SMSes on my phone is a pain and I want it to be easier.
Medium term goal: I want a laptop the size of a cigarette packet, not
the size of an A3 piece of paper. (This is the second project that I
have on my mind, about which I shall write an email at some point :))
Long term goal: I want to be online wherever I am, everywhere, without
having to take my phone out of my pocket or my laptop out of my bag. I
want a head-up display on my glasses that overlays augmented reality
on what I'm looking at. i want the line between the real world and the
internet blurred into meaninglessness. I want to be able to do things
online without having to switch away from what I'm doing into "be on
the computer" mode.
Being able to get any repeatable output out of the wristband at all is
a good first goal, though. If you can type letters with it, everything
else is merely software after that. I can do software ;-)
G Bulmer
On Jul 9, 11:12 am, Stuart Langridge <s...@kryogenix.org> wrote:
> >> ...
> >> wristband on each wrist?
> > No, one hand is enough for the full ASCII character set.
> > This product was on sale in the early '80's:
> >http://en.wikipedia.org/wiki/Microwriter
> > I tried one, and a friend had one.
> > He loaned it to a very good shorthand typist at his work while he went
> > on summer holiday, and in two weeks occasional practice, she got about
> > as fast as she could do shorthand or type.
>
> > It only has six buttons, so you just need two position for e.g. the
> > thumb, and your done.
>
> Sort of. I looked at the Microwriter. There are a few flaws, like, I
> can't bend my little finger without also bending my ring finger, which
> I think is pretty common :-)
I got to use his Microwriter for several hours, and my chum used it
for years.
The pressure needed to press the buttons is just enough to avoid the
problem. I never had a problem with that effect, and AFAIK neither did
he, and he could type as fast with one hand as many trained typist
could type on a QWERTY keyboard with two.
If you could crack the problem, there is no reason to restrict it to
walking down the street, and you may find pressing against, e.g. a
table top, may make it easier.
Stuart Langridge
> for years.
> The pressure needed to press the buttons is just enough to avoid the
> problem. I never had a problem with that effect, and AFAIK neither did
> he, and he could type as fast with one hand as many trained typist
> could type on a QWERTY keyboard with two.
>
> If you could crack the problem, there is no reason to restrict it to
> walking down the street, and you may find pressing against, e.g. a
> table top, may make it easier.
Yep. If it's considerably easier to sense the differences if fingers
are pressed against a surface rather than just bent, then that's fine;
you can just "type" against the side of your leg, or the inside of
your coat pocket, or the back of your other hand :)
My real "non-negotiable" criterion is that the "keyboard" should be
able to be worn all day without it impeding anything else you're doing
and without anyone noticing. This is why data gloves or a thing
strapped to your palm (both of which already exist) aren't really what
I'm personally looking for. Obviously this invisibility criterion
doesn't apply to prototypes :-)
Ade
Personally I write my SMS with a stylus and handwriting recognition -
feels much more natural than typing
^ My 10 cents
G Bulmer
> > Can I just ask what the overall objective for the project would be?
> > Is it to make something small; something that's as invisible as
> > possible to use (ie small movements), to make an interesting interface
> > for a particular application; or to get to grips with a particular
> > type of technology.... etc
>
> > Being a bit more specific with what you want to get out of it might
> > naturally point the way towards particular solutions and/or highlight
> > which are the meaty problems to tackle.
>
> Heh.
>
> Short term goal: typing SMSes on my phone is a pain and I want it to be easier.
Skeletal (i.e. not big and obtrusive) chord keyboards exist, and would
likely be straightforward to make in metal or plastic.
Or just use the phone body to be a chord keyboard.
> Medium term goal: I want a laptop the size of a cigarette packet, not
> the size of an A3 piece of paper.
I don't understand how that constrains the a solution to be a
wristband keyboard;
a simple, small, input device would be fine. Like the laser projector
keyboard:
http://www.virtual-laser-keyboard.com/
Or a small one handed keyboard (size makes two handed seem pointless)
I assume you'd like quite a big display (my eyes are going :-)
> (This is the second project that I
> have on my mind, about which I shall write an email at some point :))
> Long term goal: I want to be online wherever I am, everywhere, without
> having to take my phone out of my pocket or my laptop out of my bag.
Doesn't mean to be a wrist band to me yet.
> I
> want a head-up display on my glasses that overlays augmented reality
> on what I'm looking at. i want the line between the real world and the
> internet blurred into meaninglessness. I want to be able to do things
> online without having to switch away from what I'm doing into "be on
> the computer" mode.
Use "thought-detection" and avoid the physical completely.
>
> Being able to get any repeatable output out of the wristband at all is
> a good first goal, though.
They are unobtrusive as a physical object, but I don't understand what
is unique.
You could use combinations of toe movements, sub-vocalise or hum and
have an in-mouth mike, track eye movement, shake your hands, think a
thought, wiggle your ears, tense muscles groups, ...
Can you explain what is key about the wristband, and what you'd accept
as a second-best?
GB-)
Stuart Langridge
>> a good first goal, though.
> I don't understand why wristbands are key.
> They are unobtrusive as a physical object, but I don't understand what
> is unique.
> You could use combinations of toe movements, sub-vocalise or hum and
> have an in-mouth mike, track eye movement, shake your hands, think a
> thought, wiggle your ears, tense muscles groups, ...
>
> Can you explain what is key about the wristband, and what you'd accept
> as a second-best?
I'm not totally wedded to the idea of a wristband, not at all. If
someone can come up with an alternative which is all of
(a) essentially invisible
(b) totally unimpeding to everything I do
(c) trivial to take on and off
(d) works when you're walking
(e) allows me to roughly emulate a keyboard, since that's an input
device that can do everything
then I'm more than happy to talk about that instead :)
The best thing I can come up with that meets those criteria is a
wristband. Tracking eye movement invisibly is currently not rational,
since the thing which tracks eye movement needs to be somewhere in
front of your face. Shaking your hands requires detection somehow,
which seems wristbandish to me. Toe movements is doable, but it's less
trivial to take on and off. I don't know how you'd do muscle-group
tensing -- myoelectric stuff might work for that, but it's
non-trivially expensive as far as I can tell. Detecting thought is
science-fiction for us, as far as I can tell -- it's doable, if you're
a university research lab, but it's massively in its infancy.
Subvocalisation...I keep reading about that in books, but I can't find
a lot about how itwould actually work in practice. Thoughts on that
invited. :)
G Bulmer
> >> ...
>
> > Can you explain what is key about the wristband, and what you'd accept
> > as a second-best?
>
> I'm not totally wedded to the idea of a wristband, not at all. If
> someone can come up with an alternative which is all of
> (a) essentially invisible
> (b) totally unimpeding to everything I do
> (c) trivial to take on and off
> (d) works when you're walking
> (e) allows me to roughly emulate a keyboard, since that's an input
> device that can do everything
> then I'm more than happy to talk about that instead :)
Is it approximately in priority order?
>
> The best thing I can come up with that meets those criteria is a
> wristband.
> Tracking eye movement invisibly is currently not rational,
> since the thing which tracks eye movement needs to be somewhere in
> front of your face.
if it portable, would be in front of your eyes too.
> Shaking your hands requires detection somehow, which seems wristbandish to me.
do.
If you don't mind looking like a complete wierdo ('compromising' one
of your requirements:), you could use this technique with navy
semaphore :-)
> Toe movements is doable, but it's less trivial to take on and off.
Sorry, I don't understand what the requirement to be 'trivial to take
on and off' is about.
> I don't know how you'd do muscle-group
> tensing -- myoelectric stuff might work for that,
Or very simple strain-gauge-style resistance. My forearms change is
circumference (a bit) when I tense them.
I imagine there are other effects too like changing blood flow (which
may be measurable, e.g. with a reflective sensor).
> but it's
> non-trivially expensive as far as I can tell. Detecting thought is
> science-fiction for us, as far as I can tell -- it's doable, if you're
> a university research lab, but it's massively in its infancy.
But more seriously, I am just trying to understand, so might it be
*inferior* to you, in some tangible way, to wristbands?
> Subvocalisation...I keep reading about that in books, but I can't find
> a lot about how itwould actually work in practice. Thoughts on that
> invited. :)
I think he said there is a musician who uses it, but I may be wrong.
The "real thing" was quite expensive (over $1000), but we couldn't
understand why. We were thinking of a very simple experiment with a
small mike in a little plastic bag on a wire. We'd look like hooked
fish, but we'd get some idea of what information is available.
GB-)
Stuart Langridge
>> someone can come up with an alternative which is all of
>> (a) essentially invisible
> hidden within every day wear, and requires small, normal, movement?
Yep, that's a better way of putting it.
> That's a great list!
> Is it approximately in priority order?
Speaking purely for myself, any solution which doesn't meet all those
is not something I will actually use, I think...
>>
>> The best thing I can come up with that meets those criteria is a
>> wristband.
> I can understand why you like them, but what might second best be?
Don't know. I'm open to suggestions there :)
>> Tracking eye movement invisibly is currently not rational,
>> since the thing which tracks eye movement needs to be somewhere in
>> front of your face.
> Agreed, but I thought you wanted some sort of head-up display, which,
> if it portable, would be in front of your eyes too.
Yep, but that's in the "long-term" section for a reason, and it's that
an invisible HUD is science-fiction at the moment :-)
>> Shaking your hands requires detection somehow, which seems wristbandish to me.
> Agreed. I thought accelerometers/gyros in wrist bands would be easy to
> do.
> If you don't mind looking like a complete wierdo ('compromising' one
> of your requirements:), you could use this technique with navy
> semaphore :-)
ha! You are correct, but any suggestion that involves me doing
semaphore while walking down the road, frankly, is going to lead to me
continuing to use my laptop :)
> Sorry, I don't understand what the requirement to be 'trivial to take
> on and off' is about.
Think of it like a watch. I wear a watch every day. When I have a
shower, I take my watch off; when I get out of the shower, I put my
watch back on. It's a trivial action to do that. If, to be able to
know the time, I needed to strap a four-part Heath Robinson
complicated item to three separate parts of my body, I'd do without
it. That's what I mean by trivial to take on and off -- I carry a
phone because all I need to do is drop it in my pocket. If it was hard
to have a phone (for example, I needed to take a phone and a backpack
everywhere, like you did in the 80s), I'd do without the phone.
>> I don't know how you'd do muscle-group
>> tensing -- myoelectric stuff might work for that,
> Yes that would be ideal.
> Or very simple strain-gauge-style resistance. My forearms change is
> circumference (a bit) when I tense them.
> I imagine there are other effects too like changing blood flow (which
> may be measurable, e.g. with a reflective sensor).
Worth thinking about this, although I can't for the life of me imagine
that it's capable of more discrete motions than my fingers ;)
>> but it's
>> non-trivially expensive as far as I can tell. Detecting thought is
>> science-fiction for us, as far as I can tell -- it's doable, if you're
>> a university research lab, but it's massively in its infancy.
> Yea, but it might be cute :-)
> But more seriously, I am just trying to understand, so might it be
> *inferior* to you, in some tangible way, to wristbands?
No, hell no. If you can make an input device where I can think stuff
and that's enough, I'll have that instead. It admirably fulfils all my
criteria :)
>> Subvocalisation...I keep reading about that in books, but I can't find
>> a lot about how itwould actually work in practice. Thoughts on that
>> invited. :)
>
> Let me get back to you on that. A friend mentioned it a while ago.
> I think he said there is a musician who uses it, but I may be wrong.
> The "real thing" was quite expensive (over $1000), but we couldn't
> understand why. We were thinking of a very simple experiment with a
> small mike in a little plastic bag on a wire. We'd look like hooked
> fish, but we'd get some idea of what information is available.
I'd be really interested in hearing your further thoughts...
G Bulmer
an alternative, but ...
It should be practical to make a 'cover' for a mobile phone, either
using a stretchable cover, or with 3D printing.
Then, we could put a simple, 6 sensor/button, chord keyboard into
that.
To use, pull your mobile phone out of your pocket, and squeeze it in
chord-patterns to type. Not better than ow, but not worse either.
Not as nice a solution as the wrist band, but, it could be powered
from the phone, and, with a bit of hacking (or judicious choice of
phone) might be relatively easy to get to work.
It would use a small processor, a connection into the phone, and a few
'squeezy' sensors
(these might be good enough for experiments:
http://www.mutr.co.uk/product_info.php?products_id=1144).
While it is inferior to your original idea, I think it may also be
significantly better than the current (IMHO naff) 3 or 4-characters/
key unpredictive-text approach.
Just a thought,
GB-)
Richard Rothwell
>
> I realise this suggestion is so far short of your idea that it isn't
> an alternative, but ...
>
> It should be practical to make a 'cover' for a mobile phone, either
> using a stretchable cover, or with 3D printing.
> Then, we could put a simple, 6 sensor/button, chord keyboard into
> that.
> To use, pull your mobile phone out of your pocket, and squeeze it in
> chord-patterns to type. Not better than ow, but not worse either.
>
If you are into chordal keyboards, you should look at the old
Microwriter system.
http://en.wikipedia.org/wiki/Microwriter
I hacve both an original MW4 and an AgendA in my old computer collection.
The chording system is extremely easy to learn. You can still buy the
CyKey http://www.bellaire.demon.co.uk/bellaire_cykey.html
(I think)
Sensing the movement of the fingers remotely might be a bit more of a challenge.
rgds,
Richard
PS I have some Python code that I've been developing that allows you
to decode 'predictive' type input, like mobile phones, but with a
bespoke coding scheme - right down to one button for vowel and another
for consonant.
Very rough code, but it works.
Adrian Godwin
The concept is based on an eight-point compass as defined by the switch combinations of the joypad. A keystroke is a motion that starts at one point and ends at another, hence there are 56 (7*8) basic keystrokes and the possibility of shifts and single-point operations.
I played with this a little and prototyped it with a PC gamepad but eventually started chasing something else. It would, however, be interesting to resurrect it as an alternative to onscreen keyboards on touch phones. It could also be implemented as phone cover as suggested by GB.
For a device that isn't handheld, it might be possible to use your tongue to trace patterns on a plate held in the roof of your mouth, though i think that would be slower than the original wrist-sensor suggestion.
-adrian
Peter Oliver
because it's just a wristband, it's invisible. No-one
knows you're wearing it. You can use it while you're walking down the
street if you want. Imagine how fast you could type SMSes with it;
If people forget to switch it off, I look forward to a splurge of random twitter messages everytime someone performs a mundane everyday task :-)
Have you seen the guy that does interactive whiteboads and finger tracking with the camera on the Wii Remote (http://johnnylee.net/projects/wii/)? Could you strap something like this to your wrist to look for your fingers?
People have pointed out that some solutions would need significant processing, but remember that some of this could be offloaded to whatever you're imputting data into.
More realistically, until you can get hold of a magic head-up display, you're going to need some kind of screen to provide UI, feedback for typing errors, etc. If you're thinking of trying to actually make something that works today, you could just glue a chording keyboard to the back of an iPhone :-)
--
Peter Oliver
G Bulmer
wrote:
> ...
> Have you seen the guy that does interactive whiteboads and finger tracking
> with the camera on the Wii Remote (http://johnnylee.net/projects/wii/)?
My favourite is the perspective display. Knockout:
http://www.youtube.com/watch?v=Jd3-eiid-Uw
> Could you strap something like this to your wrist to look for your fingers?
One could !-)
Why not disguise the camera in a hat?-)
http://www.spytechs.com/spy_cameras/hat-camera.htm
I think this is what Stuart really needs:
http://graphics.cs.columbia.edu/projects/mars/images/AR/ISWC99/LowWithHandheldAndUserBroad.jpg
:-)
>
> More realistically, ... If you're thinking of trying to actually make something that
> iPhone :-)
Make the chord keyboard slim, tidy, no-switches, and an integral part
of the phone so that it is almost invisible, and quite natural.
You could sit on a train, and people wouldn't realise your texting,
they'd think your just reading the screen of your phone.
If you got very good with it (not a prototype, I suspect :-) you could
use it while still in your pocket, and few people would leap to the
conclusion you are secretly texting someone!-)
GB-)
Phil Dawes
(2 handed though)
Gradually drifting offtopic: Does anybody have any experience with using
chording keyboard layouts here? I switched to dvorak a while back for
RSI reasons. I'm wondering if I should experiment with a layout that
doesn't involve moving fingers about.
Cheers,
Phil
P.S. Hi Garry, I used to work at Parallax while you were there
Richard Rothwell
>
> This might be interesting: http://www.youtube.com/watch?v=0vhKc94JmlY
> (2 handed though)
>
> Gradually drifting offtopic: Does anybody have any experience with using
> chording keyboard layouts here? I switched to dvorak a while back for
> RSI reasons. I'm wondering if I should experiment with a layout that
> doesn't involve moving fingers about.
Yes, i used a Microwriter AgendA for about 2 years and may have a
CyKey somewhere.
rgds,
Richard
----
Richard Rothwell
G Bulmer
Lovely to hear from you. How are you? Are you coming to the next
meeting?
Glerk, my message got lost, I'll try again.
On Jul 9, 9:31 pm, Phil Dawes <googlep...@phildawes.net> wrote:
> This might be interesting:http://www.youtube.com/watch?v=0vhKc94JmlY
> (2 handed though)
>
> Gradually drifting offtopic: Does anybody have any experience with using
> chording keyboard layouts here? I switched to dvorak a while back for
> RSI reasons. I'm wondering if I should experiment with a layout that
> doesn't involve moving fingers about.
excellent Microwriter.
single handed: http://chorder.cs.vassar.edu/forside
http://www.chordite.com/
A list of products: http://www.tifaq.org/keyboards/chording-keyboards.html
commecial: http://www.tekgear.ca/index.cfm?pageID=90&prodid=22§ion=99
http://www.computer-takeaway.co.uk/index.php?option=com_content&view=article&id=112:chording-keyboard-ekatetra-ekapad-ekapackage&catid=36:input-peripherals&Itemid=9
One handed keyboards: http://www.frogpad.com/
http://www.billbuxton.com/matias93.html
http://www.techready.co.uk/Assistive-Technology/One-Handed-Keyboards?gclid=CJCX9r3gyZsCFYUA4wodXBGBJg
http://www.freepatentsonline.com/6952173.html describes an
interesting, almost gestural, one handed keyboard patent.
Two handed chord keyboard: http://chordkeyboard.com/
(not helpful for portability, but might be handy for sufferers of RSI)
GB-)
Stuart Langridge
> with the camera on the Wii Remote (http://johnnylee.net/projects/wii/)?
> Could you strap something like this to your wrist to look for your fingers?
Not easily. I can't think of a way of having a wrist camera that can
see your fingertips. Also, Lee sticks little reflective bits on the
ends of his fingers :)
> People have pointed out that some solutions would need significant
> processing, but remember that some of this could be offloaded to whatever
> you're imputting data into.
Agreed.
> More realistically, until you can get hold of a magic head-up display,
> you're going to need some kind of screen to provide UI, feedback for typing
> errors, etc. If you're thinking of trying to actually make something that
> works today, you could just glue a chording keyboard to the back of an
> iPhone :-)
I am currently totally handwaving the notion of feedback. I've got
various ideas about an audio UI, but ignore that for the moment.
Pretend that you've got a screen. :)
Phil Dawes
> Phil
>
> Lovely to hear from you. How are you? Are you coming to the next
> meeting?
>
I think so. It's on the 15th right?
nikki
http://www.fizzpop.org.uk/hacksessions/hack-session-15th-of-july/
Please spread the word and also make use of the comments to keep track
of who's coming and working on what.
Cheers,
nikki
Martin Meredith
>
> OK, one idea I had for a thing. I'll describe it; you all tell me
> whether it's rational or science-fiction, and if it's rational I'll
> write it up on the wiki :-)
>
> Put your left hand out, flat, palm up, and then grip tightly around
> your left wrist with your right hand (make a circle around your left
> wrist with your right thumb and middle finger). Now, move the fingers
> of your left hand. You can feel tendons and muscles and whatnot moving
> in your wrist, yes?
>
> So, the idea is: make a wristband containing sensors which is
> sensitive enough to differentiate these movements, so the wristband
> can tell which finger you're moving and by how much. Give that
> wristband a battery and a bluetooth HID chip. At that point, it's a
> bluetooth keyboard, and it's just a wristband. That makes it a
> chording keyboard -- for example, bending your index finger and thumb
> together might mean "A", index finger and right finger might mean "B",
> knows you're wearing it. You can use it while you're walking down the
> small you could make a laptop that doesn't need a keyboard.
>
> So, is this halfway rational as a buildable thing?
>
> Issues I see with it:
>
> 1. It'd almost certainly need to be individually calibrated for each
> person. This isn't necessarily a problem, though; I'm imagining that
> every morning when you put it on, it beeps, you then bend each finger
> in turn in some pre-defined movement pattern to calibrate it
> 2. I'm thinking of it looking like one of those Make Poverty History
> wristbands, i.e., not big. If it looks like you've got a doughnut
> wrapped around your wrist no-one will wear it. Is it even halfway
> possible to get (a) sensors (b) a power source that'll last more than
> two minutes (c) bluetooth into something this small?
> 3. How do you recharge it? Just taking it off and dropping it on an
> induction pad would be the coolest.
> wristband on each wrist?
> there's "level 1" where you can tell the difference between "finger
> bent" and "finger straight", and maybe "level 2" where you can tell
> the difference between "finger straight", "finger a bit bent", and
> "finger a lot bent", which greatly increases the number of gestures
> you can do with one hand.
>
> Comments on whether this idea is possible are gratefully invited.
Skipping out most of the conversation and replying to this one, as it's easier
this way.
Couldn't a wristband be made out of something like Pressure sensitive fabric,
and then use something along the lines of something created at [1] to do the
whole different fingers bit?
Might be worth experimenting with this, and seeing if we can get some decent
results out of it to differentiate the fingers. If we can, then we can move on
from there.
By the way sil, you do know that if we make something like this, the commercial
properties for selling it could make us all millionairres? :P
G Bulmer
> ...
> Couldn't a wristband be made out of something like Pressure sensitive fabric,
> and then use something along the lines of something created at [1] to do the
> whole different fingers bit?
>
What is the "something created at [1]"?
GB-)
Martin Meredith
Forgot to link it...
http://www.instructables.com/id/Pressure_Sensor_Matrix/
--
Regards,
Martin "Mez" Meredith
Jon Spriggs
Have you seen the GKOS [1] project? It's two handed, using three buttons on each hand, so you could probably use the wristband idea to put three buttons on each palm? I know there were people on the mailing list talking about making it bluetoothable, although they seem to be focused on making box-of-matches sized devices.
I'd be interested in seeing what any of you can do with this, and I'd love to see what we can put together with something like the fonera 2.0 [2] hardware, USB audio and Festival :) I might even make it down from Manchester to have a look at it :D
[1] http://gkos.com/
[2] http://wiki.fon.com/wiki/La_Fonera_2.0
--
Jon "The Nice Guy" Spriggs LPIC-1 Certified
Adrian McEwen
You don't happen to work for Nokia do you? They're looking at building
some gesture-sensing wearable sleeve kind of thing, or at least looking
to patent such an idea -
<http://www.experientia.com/blog/nokia-patents-gestural-phone-controls/>
:-)
Adrian.