A Few Suggested Improvements

[ibrandt]

Hi,

It's safe to say that since first getting my TouchStream LP I have been a bit of an enthusiast about MultiTouch technology (and I'm not the enthusiast type). I've talked it up and demonstrated it to anyone that'd listen, from business users, to graphics professionals, to terminal junkies (and I'm definitely not one to try and sell someone something). I put your MacNTouch page on the screen of as many Macs as I could while at the Apple store in NYC. I've even skimmed through your patent application.

The one thing that strikes me is that although everyone I've introduced the technology to has appeared impressed, and fully understanding of the benefits, not a single one has gone on and bought one. As I tend to do way too often, I started wondering, why?

The first issue I see is quite simply the cost. This is self-explanatory. My response to, "How much was it?" usually results in notable shock. I presume the price will come down with time, hopefully by or close to a factor of 2 (to put it on par with current pro kbd/mouse combos). As I am required to type a lot by profession, and was suffering from significant pain in doing so, I was a cost-is-no-object consumer, but I doubt such consumers make up the most substantial market.

The second issue is the "I need tactile feedback" complaint. Everyone I talk to is afraid they won't be able to learn to type on a flat surface. Now I was never the best typist. I started with computers at an early age and just never wound up learning to type properly. I could bang on the keyboard fairly quickly without looking (for the most part), so I just never found reason to suffer through "The quick brown fox..." bit. Well, after finger and wrist pain caught up to me, and I committed to try the TouchStream, I figured why not go for Dvorak at the same time, as that'd force me to learn to type properly, and on a layout that was supposedly more ergonomic to boot. I must say I am a better typist now on MultiTouch Dvorak than I ever was on a mechanical Qwerty, but I do feel that had I been a good Qwerty typist to begin with I would have taken a step back on the MultiTouch Qwerty even after a year of experience.

What I do not miss from mechanical keyboards is the up and down clicking of the keys. The zero-force flat surface is an improvement over the standard mechanical keyboard as far as I'm concerned. Forgetting about the noise and force inherent in mechanical keyboards, the up and down of a mechanical key press wastes time compared to the tap on the MultiTouch surface, and you guys have done such an excellent job with the processing that stray "key presses" on the MultiTouch are never a problem for me. What I do miss is being able to know where my hands are on the keyboard. The four dimples on the home row of the MultiTouch are just no substitute for the textured surface of a mechanical keyboard. They just don't keep me homed. I need to pause and realign with a five finger rest fairly often, and it is a slow and inefficient maneuver.

I'd imagine that if even the slightest depression (.005mm maybe?) was buffed into the Lexan at each key position (say a 7mm diameter circle for example, or if you really wanted to get fancy key shaped depressions) you'd see a significant reduction in hand drift, as well as improve the ability to go between gestures and typing by again allowing for more efficient homing. I would further imagine that a CNC machine could be used to create precise and well contoured depressions such that they would not interfere with pointing and gestures to any real degree from a user standpoint. From the keyboards standpoint, while I'm imagining, lets say the resolution of MultiTouch technology is high enough such that the blur the depressions would create in the sensor image could be processed out.

I see that a textured surface could be the single biggest improvement that could be made to MultiTouch technology. It could reduce the fears of would-be buyers, it could reduce the learning curve for new users, and it could be a welcomed ease-of-use enhancement for experienced users.

My third suggested improvement is something I believe you've already considered: swapable printed surfaces. I'd kill to be able to swap out a Qwerty layout for Dvorak when someone else wants to use my computer, or for myself every now and then so as to stay proficient in both layouts. If each surface was fairly cheap this could also greatly help the accetpence of Qwerak and other such layouts as it would allow people to try them with almost zero risk. I know you can set the different layouts for the TouchStream internally, but that presumes that you already know the layout well enough to not have to look at the keys at all, which is rarely the case for people going from one layout to another.

So Fingerworks team, is a TouchStream with textured swapable surfaces for half the price feasible? If you can answer with an updated product, even if the price goes unchanged, you'd at least get my money again.

Regards,

Ian

[DoctorJack]

Thanks for your feedback. We really welcome suggestions on how to improve our MultiTouch products - so keep them coming.

We’re constantly working on or thinking about numerous improvements to the TouchStream. Most of these are in the firmware so there is little or no impact on production cost to implement and no risk of permanent damage to the product. Physical change to the TouchStream is another matter altogether. We have to be extremely careful about doing anything that may permanently affect product performance. In addition, an improvement change could have a negative impact on production cost. This means that physical changes will come slowly.

I can say that we have long thought about adding more tactile features to the surface. However, we still have much to do before we will be able to offer something like a removable cover with just the right amount of surface features.

[John Meacham]

each key should have it's character in braille raised bumps on it. that would rock, you could type and learn braille at the same time. this is a rehash of an idea of mine for regular keyboards...

http://www.halfbakery.com/idea/Braille_20Keyboard

[JerryKnight]

I share your views on the Touchstream, as well as your wondering about "other people." Most of the people I talk to either can't shell out the money (when they buy other computer items far more expensive), or they are pessimistic about learning ZF typing. I guess the general mindset of people today is that "keyboards should be cheap." I mean, the most expensive conventional keyboard is, like you said, roughly 1/2 the cost of this.

But my argument is that you spend good money on your CPU since it is always in use. You spend good money on your monitor for the same reason. Why not then spend good money on the next most commonly used piece of computer equipment, your keyboard and mouse. Except now the two are combined into one in such a way that you get even more benefit than the two things alone (gestures). I was thrilled with mouse gestures in Opera. Keyboard shortcuts are great, but not if the other things you do are with the mouse. Now with gestures, it is like using Jedi mind trick on the computer to get it to cut, paste, go back, refresh, etc.

A few issues to address: about the swappable face idea, I made the effort to learn dvorak before getting the LP, simply because I didn't want to learn 2 things at once. So I was already a decent dvorak touch-typist (same story -- never a phenominal typist before, just adequate), so I just got the QWERTY layout, in the off chance that someone would want to type on it. I only look down sometimes when I need to realign my hands, but usually that is done by the Braille dots.

About the indentations or Braille dots on every key: The developers have mentioned, and I have noticed, that even the small raised dots interfere with the sensor interpolation that is done by the keyboard. The result is that while mousing past a dot, the path of the cursor will "bend" around it ever so slightly. If you put even more dots on the surface, the mouse cursor will always travel in a wiggly fashion. I would rather not have this. The ultra-micro-shallow indentation would be the way to go, but it would undoubtedly raise production costs of the LP significantly, which would delay the price reduction or worse yet, increase the price even more. I even heard a good idea of putting very slightly rougher texturing over the center of the keys. I am not sure of the "feelability" of this, but it would not affect mousing much.

Continue the enthusiasm, and eventually the mass mindset about keyboards will change. For a while, I never took my LP anywhere, since it is "my precious" but taking it to the lab gets more than a few interested looks. There is a market for these things, if people can get away from the "cheap keyboard" idea and realize that it is worth alot of money to improve their computer input experience.

-Jerry

[-jeffB]

This month's issue of Communications of the ACM has an article from the Canesta folks talking about their experiences building their virtual keyboard. It projects an image of a keyboard on the table, and uses optical (IR) sensors to detect finger positions.

While this has some aspects in common with the TouchStream experience, it also has significant differences. On the negative side, they don't seem to have thought much about ergonomics, and they aren't doing anything with gestures. On the positive side, they're heavily focused on "cheap".

They did mention one enhancement that might be helpful for the TouchStream: modulated auditory feedback. They say:

We found that a faint click generated electronically upon recognition of a keystroke is an enormous help and markedly increases typing speed. We also experimented with modulating the quality of the sound as a function of where a key is hit. Although this trick has hardly any effect on typing speed, users learn from it and tend to type more reliably (their fingers drift less), even if they do not look at the keyboard.

A simple key-click is easy enough to do with standard third-party software, but the enhanced/modulated clicks would require a fair amount of work in the firmware. Is this something you've looked at?

I'd love to see a cooperative venture between Canesta and FingerWorks. Canesta is poised to break into the mainstream for PDAs, cell phones, and other very small devices. If they could incorporate any of the intelligence that's built into TouchStream products, the possibilities are staggering. (A "TouchStream Virtual" keyboard that's the size of a cigarette carton, runs for a day on a small battery, and costs $99.95?)

Of course, it's never that simple... :)

[JerryKnight]

I saw this "projected" keyboard a while back, but I have never seen one in use. It seems interesting, but I wonder how sophisticated the finger detection is or if it could handle multi-fingered gestures.

Keyboards for a palm-like device should definitely be cheap and small. But, for desktop computers where hours are spent at a time typing, I reiterate my previous statements. I personally would like "quality" before "cheap." Fingerworks is heavily focused on quality and usability, which should not change.

The click thing is interesting, even though I have sworn a lifelong oath never to allow keyclicking again (remember ICQ???). Perhaps if the keyboard itself had some sort of piezo clicker thing (extremely small and faint), but this gets into hardware design change...

-Jerry

[-jeffB]

I'm with you on "dumb" keyclicking -- the tactile feedback of my finger hitting the surface seems more than adequate -- but I'm intrigued by the possibility of richer feedback as a training mechanism.

I also agree with your comments about putting quality before price for a desktop keyboard. However, it seems like there are a lot of circumstances where small size and portability are important. If one could incorporate FingerWorks' recognition intelligence into a different physical sensor arrangement, it would (1) open new revenue sources for FingerWorks, and (2) make me a lot more likely to put up with a handheld/wearable platform... :)

[Bob deWitt]

I am not currently a user of any MultiTouch products, but as someone who is contemplating purchase, one of the issues that has caused me to reconsider my initial enthusiasm was the lack of any kind of feedback with 'pressing' keys. I think it would be difficult to get used to no feedback. Since obviously tactile feedback would not be practical, then some kind of audio feedback would be desireable, it seems to me. I know that when browsing the internet if I click on a button with the mouse, for instance, that does not provide some time of feedback (a click, a 'depressed' state, a glow....something), it leaves me with the uncomfortable feeling of wondering whether my input has been accepted. I would think that the same thing would apply to the 'keys' on a TouchStream keyboard.

Just my two cents worth.

Bob

[JerryKnight]

Really, the lack of tactile feedback is a down-side to the LP for me. But it wouldn't be possible to have feedback and the other wonderful features, such as mousing, gesturing, etc. If there were feedback, then there would be nothing different about TS.

You're obviously intrigued already.. The feedback issue is big, but for me all the enthusiasm came from the ability to do all input on the same surfaces. Typing, mousing, gesturing - all is possible at any time. If you try it and you're still dying for feedback, then it is possible to experiment with the audio feedback, like key-clicks. I think most flavors of linux provide that feature, and I bet the shiny new windows XP or OS X (they caught on to the whole 'x' thing) can do key clicks as well. When mousing, if you click on a link, the page loads (or at least, the busy icon-thing moves). If you click on a window or button, something reacts to the click. The same goes for typing and gesturing. Usually you see what you typed immediately (unless your computer is seriously bogged down or it's a slow terminal connection). There is rarely no visual response to something you do with the keyboard or mouse. And when that happens, just trust that the keyboard recognized your key/gesture. The keyboard and its firmware are really quite robust.

But you should take a leap of faith and try it. The typing will be a down-side, but hopefully you can get used to it and let the benefits outweigh it. I actually like typing on my LP now that I am somewhat used to the idea. When I go to the lab and type on a mechanical keyboard, I miss alot of keys since I have to push so hard! I get spoiled by being able to type on the LP ever so lightly instead of having to pound away on the keys.

[Bob deWitt]

Thanks for your valuable feedback, Jerry. I have decided to get my feet wet with an iGesture Pad first. If I like the experience with it, perhaps I'll consider upgrading to the keyboard before the 30-day evaluation period is over.

Bob

[nomaded]

Bob, that is exactly what I did. Back when FingerWorks first announced the TouchStream keyboards, I still wasn't sure about how the technology would work, so I picked up one of their iGesture Mini keyboards (nee TouchStream Mini) to try it out. After having it for a couple weeks, I preordered a TouchStream Stealth. After a month of two, the Stealth arrived and I've been using it ever since, while at work.

And I just got my MacNTouch for my TiBook, so I'll be using FingerWorks MultiTouch technology almost always now. Whee! :D

[aegis]

The improvements I'd like (to an already impressive product):

LED's for CAPS and NUM lock, I end up using key clicks to avoid missing the caps word gesture.

Detachable USB cable, to make the ST more portable and make me worry less about damaging it!

Daniel.

[aegis]

Next on my wish list is swapping hands. Normally, being right handed I have mouse on right surface, cursor on left. When I'm using the editor I'd like to reverse this. A palm slide gesture to switch hands would be very useful. I'm aware that this can be done from the feature selector but there doesn't seem to be a macro for it.

Daniel.

[losvedir]

As it's been brought up here and a lot of places, it appears that the lack of depression is a major turn off to people. I'm kind of wondering why. When I first learned to type, I loved it so much I would do it everywhere; I would imaginary type on my lap, on desks and tables, and all over the place (much to the chagrin of my mom..) so I don't see this being much of a problem. So anyway, what about no tactile feedback don't you like?