Computer systems will soon mediate our interactions with the outdoor, physical world. Some of these systems will be carried with us, others will be built into our environment. The portable and fixed systems will interact in interesting ways. New applications and new risks will arise. Here are some ambient computing technologies I would like to help create.
It is becoming possible to repair a growing range of injuries to people's nervous
systems, whether by stimulating re-growth of damaged nerves or re-wiring the nervous
system with artificial connections. Although much remains to be done before
paraplegia can be cured, the question is raised: what if we go further and
give people better-than-human control over their bodies?
As a rock climber, I can think of several situations in which it would be desirable to have conscious control over my autonomic nervous system. I would like to be able to suppress the sweat glands on my hands and feet to maintain grip.
It would also be advantageous to be able to completely rest muscles that aren't in use, even when other parts of the body are under load. When one is under stress, the whole body tenses up, wasting energy.
When wearing heavy winter clothing one could remain warm by suppressing the sweat response over one's entire body, thus preventing one's clothing from becoming waterlogged. Damp clothing conducts heat much faster than dry.
Another application for Smart Nerves would be
to control one's facial muscles to project any desired
emotion, or to hide one's real emotions. Obvious applications are business
negotiations, stage acting and poker. One's entire physical affect
(emotion as evidenced by posture, expression, breathing, sweating, pupil dilation,
heart rate) could be brought under explicit control to good effect.
A system that continuously analyzed which of a user's muscles were in tension would have a wide range of applications. Such a system would be easier to build than the Smart Nerves described above, and could be noninvasive, based either on video analysis or on a network of electrodes. Some examples:- For an actor, a facial analyzer could give a continuous readout of the emotion being portrayed.
Interested capabilities become possible when you combine a GPS unit with a network-connected smart 'phone and bluetooth-enable the device for easy connectivity with a home computer. You can then use GUI applications on the PC to manage location-dependent services delivered through the smart phone, for instance an automatically updated map of where your friends are at any given moment. These devices are getting cheaper every year, so it's possible to imagine a family going to the CNE and attaching one to each of their kids: the parents can find the kids just by looking at a continuously redrawn map of the grounds that has all the buildings labeled, with the kids visible as dots moving around! Even if a kid goes into a building, a history function could show his trail for the last half hour...
For a more commercial instance, let's say the user has a wish-list application running on their computer. Since it's a couple of years from now and all product offerings have XML descriptions stored in an on-line database, the wish-list application can be powerful: all the offerings from various manufacturers can be viewed side-by-side, viewed from all angles, their attributes compared in automatically generated charts, with consumer ratings and reliability reports. Because of the richness of a product's XML description, it's also possible to view products for similar purposes, from the same manufacturer, in the same product line, made with the same material, etc. This includes used products, perhaps through an E-bay link.
As the consumer modifies their wish list on their computer, their smart 'phone keeps in sync, and as they walk along, stores having products on the list are highlighted on the 'phone's screen.
Dating is another obvious application: describe your ideal match in a user-friendly application. Your 'phone syncs up and if you happen to be close to someone both compatible and on the prowl, your 'phones will notify each of you of the other's presence.
Oddly, what 'phones really need to make their services useful is a voice interface... Not that you can't speak into them already, but the ability to respond to voice commands will make a lot of functionality more useful. At a desktop machine the user can usually point, click and type more efficiently than a speech interface, but cell phones need a way to get around their tiny keyboards.
Although the peasants and the livestock may accept the marketing-driven reality constructed for them by advertisers and politicians, the rest of us may someday filter our reality through glasses that filter out the advertisements on billboards and in newspapers. Instead, the content of billboards and ads will be replaced with content of our own choosing, perhaps art from our collection, or text from books we've chosen to read, or news updates, or whatever we choose to see.
One objection to heads-up displays has been that as the computer generating the images can't generate frames fast enough to keep up with the user's head motions, the user gets vertigo or nausea because what he sees doesn't agree with what his inner ear is telling him.
This effect might be ameliorated by having the computer generate a view larger than that actually seen by the wearer. As the glasses wearer makes head motions, an inertially-guided lens pans over the computer-generated image.
This would tie in with having actual lenses form part of the vision system.
I recall reading about next-generations lenses for glasses. By mapping people's actual corneas and creating lenses to fix the vision at each point, better-than- human vision could be achieved. Something like these enhanced lenses could be inserted into the fiber optic pathway through which the computer-generated travels to get to the user's eye. The lens researchers had also found a way to produce these lenses in the form of a "correcting mirror", which would produce a clearer reflection (for a certain viewer) than the original image. Perhaps this technology could be used to create the viewing plates on the interior of the rose-coloured glasses.
When shopping, the rose-coloured glasses from the previous section might also put haloes over "good" products, pitchforks over bad products, arrows on the floor pointing to desired products on your wish list, ghostly barbed wire across paths to undesirable objects.
What is good? If you subscribe to an ethics server, you can allow organizations whose morals you agree with to express their moral judgment of your environment automatically. By adding a moral dimension to your landscape, walking around can become educational. Staring at an object brings up an overlay of the back story behind your ethics servers' opinions: where it was produced and how, using what materials, by whom under what working conditions, best and worst alternatives, and the messages you would be sending by buying or not buying this product. If adopted on a mass scale, this would enable the masses to make their message heard, not only in a mass "Wal-mart is evil" sort of way, but with nuances.
Another dimension of good is how healthy food items are. Those containing trans fats would appear with little skulls and crossbones, as would anything possibly containing nuts if you the user had a nut allergy.
This is reminiscent of 3rd Voice, a technology that in its early days enabled people to attach comments to websites. Anyone with a 3rd Voice plugin could see the messages everyone else had posted when viewing the same website. (Note desirability of moderation services; perhaps a Slashdot moderation model with multiple axes of moderation). Within the roseglasses paradigm, people could leave notes attached to various real-world objects, or painted on the objects with virtual paint.
The same information that pops up when wandering store aisles could be repackaged in a more informationally dense format for use when a user is building his or her wish list.
The user begins by selecting a recipe or objective: "I want to make brownies", or "I want to fix the garage door." Various online and local recipe and how-to databases can be searched to find existing recipes (note consider business models for recipe/how-to book publishers to append their recipes to universal databases. Perhaps once you buy a cook book you get a code unlocking the corresponding recipes in the online database) or the user can add their own recipes. Subscription services and micropayment schemes for per-recipe purchases are possible, as are organizations who want to promulgate their way of life by donating their recipes en masse. Note the desirability of a universal recipe/how-to format that would enable anyone to start a recipe/how-to server. Then one's wishlist client could query multiple data sources seamlessly.
A selection of recipes can be nominated as "frequently-made." One's glasses will guide the way to bulk bins at the supermarket, or to stores that sell that constellation of ingredients in bulk. In the Map View of the Wish List application, ethically sound places to pick up desired equipment and ingredients/parts are displayed.
Recipe filters can modify recipes in various ways:
- localize recipes for high altitudes;
- bread machine makers could supply filters that modify bread recipes for their particular bread machine.
- Diabetics could apply a filter that replaces sugar with acceptable alternatives,
- high blood pressure users apply a salt (including hidden salt, as in soya sauce) filter. Note this last implies that some ingredients need a measure of how important they are to the recipe. Just as you can't make Jolt Cola without 3x normal caffeine, a certain amount of salt is essential to most bread recipes. Applying a high-blood pressure filter to a Foccaccio (salty Italian bread) recipe should produce a warning message that you shouldn't be eating Foccaccio if you have high blood pressure.
- Filters could theoretically be used to screen out or substitute ingredients from given manufacturers, but this may already be implicit in the later "provisioning" stage of the wishlist fulfillment process.
Recipes contain equipment and ingredients. How-to manuals reference equipment and parts. The user's wish-list editor accesses lists of the equipment and ingredients the user owns in various locations. The editor has lists of alternatives for equipment and ingredients, as well as assessments of how good the alternatives are. For instance, if you don't have a pipe wrench, you can 86% pretend with a strap wrench when wrenching 2" PVC pipe. In The Kitchen at The Cottage I have almost everything I need to make Brownies. Printing the recipe for production in The Kitchen at The Cottage produces a list of the ingredients or alternatives already in stock at the cottage, plus a list of stuff to get, and instructions modified for use with the stuff you (will) actually have to produce the recipe with.
Like everyone else and their dog, I'd like to become involved in the design of navigation systems for cars. First you give cars an accurate bunch of sensors so they can tell where they are, what the landscape is like, what vehicles are nearby, and what those vehicles are doing. Then you network the cars together so they can share information about the landscape and what they're all planning to do. Then you distil all this information and present it on a heads-up display for the driver. Because the car's "sight" involves GPS access to a fine- grained terrain map, detailed radar and infrared cameras as well as visual-spectrum cameras, plus telemetry and terrain database updates from other nearby vehicles, a car can have a pretty detailed idea of its surroundings. Not only should the driver be presented with an evolving precis of this information, it can be used to allow the car to refuse to do anything dangerous.
That's the first stage. The next step is for all the cars to be chatting with each other continuously and cooperating on how to space themselves out on the road, speeding up or slowing down slightly so as to allow crossing traffic to traverse intersections without stopping. (Stop signs are necessary only if drivers have human limitations.)
Note: Intersections themselves and busy sections of road would also have wide-spectrum cameras and other sensors which would drop information into the car navigation information network for access by approaching vehicles.
The final stage is to be able to say to the car, "let's go to the cottage," and then have the driver snooze the whole way as the car percolates through the road system on its own, stopping at automated filling stations as necessary. One could also say, "let's take the scenic route," and the car would choose roads near points the driver's personal profile indicated might be of interest.