Why my PDA is an iBook, and what the next one will be

[bnewman]

I had a Palm III for a while, but I didn't really use it. Today, I have an iBook, and I use it as a PDA more than I ever did the Palm, because it's a real computer and I don't have to worry about synching it (which never really worked right for me). The thing is, in most situations where I can afford to devote both hands to computing (one for the PDA, one for the stylus), I can afford to sit (or squat) and take out my iBook (I'm typing this on the bus right now). The Palm wasn't more portable in the sense of being more convenient to use, just in the sense of being smaller and lighter, and I have a nice, padded computer backpack and routinely pack my iBook when I'm going anywhere except shopping.

So, what would qualify as more portable than this iBook? A real wearable computer would. I'd like a dataglove (which, now that we have the Wiimote, should be well within reach) and a transparent, binocular heads-up display which places virtual objects at in an arms-length-radius sphere around my head, organizing data using a BumpTop-like spacial persistence model. An Exposé-like show-me-my-desktop gesture moves everything to the periphery so I can see the real world.

The thing is, this doesn't require a new computer — I'm already carrying a computer everywhere that's more than capable of all this. All I need is my bluetooth dataglove and HUD, and an appropriate desktop manager, and I'm good to go. So where are they? The dataglove should be easy after the Wiimote. The HUD will be harder.

Anyway, after we have these the possibilities open up a lot. To the sensor which locates the dataglove and HUD relative to my body, add another that locates them relative to the room. Now I can place objects at fixed locations in space, which would be great for someone like me who likes to pace while developing an idea. I can even save the arrangement of virtual objects and reopen it in another room of the same dimensions. And if any spam gets past my sentient AI secretary, I can crumple it up and throw it into the trash from across the room.

Another killer app would be location-based services — using GPS or, for greater precision, a neighborhood council's localizer grid, I could see Mapquest, Google maps, or Google Earth overlaid on reality. I could click on a restaurant and open a PDF of their menu. This should be enough of a gain for local businesses that it shouldn't be too advertizing-polluted. (I would expect the virtual world to have a bit more advertising density than the real one, but not by that much. This may be hopelessly naive.)

I'd pay $50-$100 for the dataglove and $200-$400 for the HUD, assuming both lived up to my expectations. It should be possible for these technologies to hit those price points within the next 5-10 years... here's hoping.

Comments

[orawnzva.livejournal.com]

I think I'd prefer voice, maybe with a sensor attached to my throat to detect subvocalization so I don't have to speak out loud. A good fully-featured wearable system would include a variety of text input methods, including voice, air-writing, air-typing, and ASL fingerspelling (or even full ASL, if we're getting ambitious).

[foxfour.livejournal.com]

from an NLP point of view, "awww, do i have to?"

seriously, all those would be great. handwriting seems the most easily feasible, though.

(i like the idea of a virtual tablet, on which one can write — air writing to all around you, but to you — a floating scroll! or whatever!

and think about the possibilities with a zeroconf network and sharing objects.)

[foxfour.livejournal.com]

addendum: problem with airwriting and the bumptop-style interface: in both cases, pressure is important. what will the glove measure pressure against? distance from the reference point? it would need to correct for slight movements of a reference point that you carry, and need to allow for wobble in hand-placement against the air. but could be done.

[orawnzva.livejournal.com]

BumpTop just happened to be on my mind, but really an interface involving objects floating in space is going to be very different from one involving objects sliding across the surface.

Anyway, you'd use gestures — make a grasping gesture to pick up an object, open your hand to let it go, etc.

[foxfour.livejournal.com]

that sounds good.

[sildra.livejournal.com]

I know many people who type more quickly than they handwrite, and several others who can't even read their own handwriting. I think I'd prefer air-typing to air-writing. I actually have very legible handwriting, but I write at a pretty average speed, and type almost fast enough to take dictation. Part of the reason I've never been attracted to PDAs in general is I don't like it that you have to handwrite--what's the point of having a computer if you can't type?--and my fingers (way on the small end for an adult) are too big to ten-finger type on those things.

[foxfour.livejournal.com]

points. i've not tried air-typing, but my intuition is that i would have lower accuracy than with real typing or air writing, though.

[orawnzva.livejournal.com]

In the case of air typing or air writing, of course you would have lower accuracy if it were tracking the movement of your fingers relative to a virtual keyboard or page located in space, but it won't be doing that, it'll be tracking the movement of your fingers relative to your wrists. Try making writing motions in the air and see where the information is really being generated.

Of course, only real touch-typists will be able to air type — since there's no virtual keyboard per se, there's no way to hunt and peck. That's why I, who can't type very fast, want voice, possibly augmented with some derivation of Dasher for fast correction/disambiguation.

[tibicina.livejournal.com]

Well, why not have a virtual keyboard? I mean, it's a virtual world of virtual objects - how hard could adding a keyboard be?

[ccommack]

I actually have a mild objection to devoting too much power/hardware to input. Not only do I think it's inelegant, there's a practical limit to how much attention you want people to be able to devote to a computer environment when they're walking around in the world. I think restricting input to a few basic functions (clickwheel on a wiimote) would be fine, and saves the input for keyboard-and-pointer. (Yeah, I'm a CLI partisan. Shoot me.)

[orawnzva.livejournal.com]

I'm a CLI partisan, too, that's why I want voice! You need hands-free text input to be able to take notes in a hands-free situation. I don't want to have to be fumbling for a thing when I'm trying to pay attention to what smeone is saying so I can write it down. Yes, someday someone will get hit by a car because they tried to cross the street while playing solitaire on their HUD. That future person is an idiot. I am not going to do that.

[ccommack]

I'm just unconvinced that voice is a good way to interact with a CLI.

[orawnzva.livejournal.com]

It's not a good way to interact with a current CLI because they aren't designed for it. Current voice-recognition interfaces are either makeshift skins on top of other kinds of interfaces or voicemail mazes where you only have a few choices. A voice interface that is really a voice interface has yet to be developed.

[orawnzva.livejournal.com]

In fact, my GPS-enabled, map-aware computer of the future will beep at me if I try to cross the street while interacting with a graphical virtual object that is not located on the street.