Rainbow's EndThe first bit of dumb luck came disguised as a public embarassment for the European Center for Defence Against Disease. On July 23, schoolchildren in Algeria claimed that a respiratory epidemic was spreading across the Mediterranean. The claim was based on clever analysis of antibody data from the mass-transit systems of Algeria and Naples…

So begins Rainbow’s End, by VernorVinge, “A novel with one foot in the future.”

The first chapter of the book, which is about the central problem that frames the remainder of the book, seamlessly intertwines:

This is only the beginning. The book is loaded with ideas.

The entire text of the book is now available online, for free: http://vrinimi.org/

The book assumes the SoftSingularity? scenario. That is: Technology is developing at accelerated rates, but not so accelerated that human beings can’t co-develop in time.

Some of the technologies and issues of the 2020’s:

This is just a small sampling of the ideas; there are far more.

The book delves deeply into the interactions between the technologies; it is NOT guilty of the single advancement problem. “Oh, look, we have spaceships,” but people do intelligence ye olde fashioned way. C3PO and R2D2 can’t communicate by wireless. [7]

The book hinges on both the undeniable, powerful, and crucial strengths, and the weaknesses, and pettiness, of ThePublicWeb.

It also has a lot to say about age; The interactions between younger & older. Talent, and mentorship. As we develop into the future, the intellectual advantage of being older declines.

Everybody is on the same page, regardless of age; Something I personally have learned and confirmed, as a programmer. Even deep wisdom is communicable and communicated, a la the DesignPatterns movement, and wealth of publications on various general systems theories.

And yet, there are things that can only be known, and only be had, through age. History and path dependency are forces of their own.

The book has a lot to say about intelligence, as well. The AutomaticTranslationFiction shows up in the story as JITT, or just-in-time-training, which is illegal, but also, …

The book is just full of ideas. It’s idea-dense. They’re just spilling out, all over the place.

This is, undoubtedly, a reference book, until at least 2010. Science fiction authors who need to write about the 2020’s will, for a time, be writing straight into VernorVinge’s concept of the 2020s. It’s simply too well put together, and too well articulated, to be ignored. It will take a lot of effort for someone to deviate, and make a convincing story. It can plausibly be done, but it’ll require a lot of intelligence to do, and VernorVinge has a very elite pool of analysts, industry sensors.

See Also


http://del.icio.us/rss/tag/rainbowsend returned no data, or LWP::UserAgent is not available.

http://feeds.technorati.com/feed/posts/tag/rainbowsend returned no data, or LWP::UserAgent is not available.

RSS parsing failed for http://unalog.com/tag/rainbowsend/rss: mismatched tag at line 42, column 2, byte 1390 at /home/alex/perl5/perlbrew/perls/perl-5.26.1/lib/site_perl/5.26.1/x86_64-linux/XML/Parser.pm line 187.


Incidentally, I have no idea how you’d deliver (alone!) a full 1024x768 displays worth of data at 60 Hz to a contact lense, much less the far greater resolution that would be demanded of a full augmented reality system. If anybody has a clue how that could conceivably work, please leave a note here.

The size, weight, and power of computers and their displays continues to plummet. The smallest and lowest-power display I am aware of (as of 2006) is the “virtual retinal display” (VRD), which scans a small, low-power beam of light across the retina. Microvision ( http://mvis.com/ ) is apparently commercially selling VRDs, based on research at the virtual retinal display group ( http://www.hitl.washington.edu/projects/vrd/ ). I’ve been told that someone (?) is trying to build something similar into cell phones, so that cell phones can get even smaller, but have a screen that looks larger and is easier to read.

For a SF story, it would be plausable enough to say that the contact lens contains (VCSEL) surface-emitting lasers.

DLPs and LCD projectors are competing technologies that are also pretty small and shrinking.

I’ve worn the NOMAD visor display, which I believe makes use of the VRD.

The path of bandwidth for the NOMAD is clear: You have a big fat cable from the device to the visor, and the visor is the thing emitting the laser that bounces off a reflective semi-transparent plate.

What I’m baffled by, is how we’ll get display data to the contact lens of the future.

Options, it seems to me:

  • Generate the complete image on the contact lens itself, taking cues from the environment.
  • Wirelessly send the complete image to the contact lens.
  • Somehow route data to the contact lens through the body.

I doubt that the first and second are plausible, even with sophisticated nanotechnology. I don’t know what the bandwidth carrying capabilites are of the human body, and the eye: I suspect it is much less than radio.

I think it’s important to remember that computer chips are already nanotechnology. The smallest feature size (I still don’t know exactly what that means: Are they saying the complete width of a transistor? What exactly is a “feature?”) is 65 nanometers. It’s plausible we’ll make it to 15-20. It’s not plausible that transistors as we know them will get down smaller. The ITRS believes we’ll be petering out in miniaturization around 2018, unless we figure out something dramatically different. We probably will, but it’ll probably only get us another 1 or 2 more halvings (Carbon Nanotube, or whatever.) It’s pretty bleak, from then on.

Our 3-D processors have already gone past the CPU, and moved to specific hardware encodings of specific algorithms. So I don’t see that as “a way out” either.

So, in short, I don’t see how we’re going to put a 3-D rendering & visualizing system-on-a-chip onto a contact lens.

But glasses? That’s practically a given! We can be near-certain that people will be wearing AugmentedReality glasses by 2025. I personally believe that 2015 is very likely. There are basically zero obstacles, and several people are already doing limited forms of it. Commoditization, new desire, all the factors behind it are on the way.

I would give more likelihood to direct neural interface, or cybernetic manipulation of the eye, than contact lenses. “Messy?” Yes, but, clearly possible.

Or perhaps we’ll have the contact lenses, but you’ll have to have a cap as well, and the lens will be more of a display, receiving and rerouting data directed to it from the cap.

I recently saw that Vernor Vinge confessed that the contact lenses are probably impossible. See roughly half-way through the interview with Vinge; He thought the same thing.

So, there are only 3 problems: Getting the information to the contact lenses, processing the information, and getting the information from the contact lenses into the optic nerve.

Silly me, I thought you were stumbling over the contact-lens-to-optic-nerve path.

Standard IrDA? devices already communicate at 4 Mbit/s. You may have already had one pointed right at your eyeballs and never even noticed. Just a couple of orders of magnitude faster (and/or some video data compression), and it could transmit full-motion video.

… OK, let’s go with glasses. That gives us nearly all the benefits of the contact lenses, and makes them far easier to build.

“The smallest feature size”

A person doing VLSI layout looks a lot like a person doing pixel art. Often she’s zoomed way in on a tiny piece of the whole image, making everything look jagged because it’s made of visible squares. (The size of every square is called “L”, or in the really prestigious schools, “lambda”).

The “smallest feature” on a chip is the width of the poly “wires”, which are 2 squares (“2 lambda”) wide.

“a MOS transistor can be viewed as the overlap of a diffusion primitive and a polysilicon primitive (see Fig. 2.1a).” – http://www.rulabinsky.com/cavd/text/chap02-1.html

Whenever a chip fab is built with better resolution equipment, designs can be shrunk down. Then you can squeeze more chips onto the same-sized disk of silicon, making each chip cost less. This also makes the newer chips run faster.

Some more pictures of VLSI layout: http://rer.vuse.vanderbilt.edu/RER/cadence/manual/man_advance2.html

Some VLSI layout easter eggs: http://www.microscope.fsu.edu/creatures/

Does that answer your question?

It is getting hard for science fiction writers to stay ahead of what people have already built. The “display on a contact lens” is no longer fictional.

Babak A. Parviz and his students have built a crude version of a device mentioned in that book – a device that we all thought was completely unrealistic when the book was published.

Photos: http://www.spectrum.ieee.org/image/829135 Article: "Augmented Reality in a Contact Lens" by Babak A. Parviz, 2009-September


My concern is that, a lot of time has passed since RichardFeynman wrote TheresPlentyOfRoomAtTheBottom?, and that – today, we’re pretty close to the bottom. Femto tech looks very unlikely; I don’t see how we would ever have machines made out of subatomic particles.

Whether we use LEDs or lasers for our display, the area available for optoelectronics on the surface of the contact is really small: roughly 1.2 millimeters in diameter. The display must also be semitransparent, so that wearers can still see their surroundings. Those are tough but not impossible requirements. The LED chips we’ve built so far are 300 µm in diameter, and the light-emitting zone on each chip is a 60-µm-wide ring with a radius of 112 µm. We’re trying to reduce that by an order of magnitude. Our goal is an array of 3600 10-µm-wide pixels spaced 10 µm apart.

That’s basically a 60x60 display, and I think they’re circularly packing, so we’re not talking about a whole lot of resolution. If they use a laser, that’s probably the better route, but I wonder, what can they address?

The other thing that gives me doubt is that VernorVinge himself doubted. He said that if he were to write it for more realism, he would have bio-electrical interfaces that splice the optic nerve. He reported that he consulted with his engineering friends to see “Are the contact lenses possible?”, but his friends said, “Probably not,” and outlined their cases to him. Vernor Vinge is pretty well connected.

There is another problem – of analyzing the eye movements and reacting to them very fast. The human eye is in constant motion, we trace the edges of objects and generally move our eyes to help interpolate their poor resolution. Such a lens would need very accurate information on the position of the eye, and possibly also on its focal point. At this point glasses are so much easier.


Define external redirect: IrDA SoftSingularity TheresPlentyOfRoomAtTheBottom

EditNearLinks: UserInterface VernorVinge DesignPatterns AugmentedReality RichardFeynman