Thursday, January 31, 2008

Brand and The Media Lab, 20 years after

Ah, here's the quote: Stewart Brand's The Media Lab, page 68. Media Lab's director, Nicholas Negroponte, has just described a half-gigabit channel as "effectively infinite" and asks what one would do with that much bandwidth:
The Media Lab's best and brightest stuttered. One volunteered you could have a combination of television immediacy and newspaper depth and detail available at any time through such a medium. Fine, that would occupy a teensy fraction of the bandwidth. What else? Uh, every house acquires its own broadcast capability. Uh, you could broadcast solids by having them fabricated at the receiving end. Uh, everything would be instantly available as super fax -- a daily National Geographic-quality New York Times with today's news rather than yesterday's, manufactured at the breakfast table.
That was written 20 years ago. It's tempting to take potshots at past speculation about what was then the future, but let's not. Predicting the future is hard. Myself, I'm hard put to predict the present.

On the other hand, some interesting questions do arise:

First question: What happened to the bandwidth? The Media Lab folks were stuttering 20 years ago because representatives of Bell (I mean Lucent ... I mean AT&T .... I mean Alcatel ... I mean ...) had just told them that they could run 500Mb fiber optic to everyone's home. That's around 60 megabytes/second, both ways.

There is no official definition of broadband, but current reports (such as this one from the OECD) use 256Kb as a working definition. By that measure Denmark and the Netherlands are leading the way (with Switzerland, Korea and Norway rounding out the top five). In Denmark, there is just over one broadband subscription per three inhabitants. If you factor in households with more than one inhabitant, this probably means that almost everyone who wants a broadband connection has one.

However, most of those subscribers have DSL (it doesn't say what flavor). If you're actually looking for fiber, Korea has about 9 subscriptions per 100 inhabitants, Japan about 8, Sweden about 5, Denmark about 3 and the Slovak Republic about 1. So much for fiber to every home. Most of us, it appears, are slogging along with DSL and cable. Your mileage may vary depending on a number of factors, but even the theoretical maximum listed here is nowhere close to 500Mb/s for either.

How did this (not) happen? That's an interesting story having more to do with economics than technology ...

Next question: Suppose we eventually all get a half-gigabit hookup. What could we possibly do with that? The answer seems very clear in hindsight: Blow it all on extra pixels. Late-'80s TV is one thing. HD-TV is quite another. Blu-Ray and HD-DVD play back at about 40Mb/s, easily handled by our infinite pipe, but they're not the ultimate.

Even HD-TV has visible artifacts if you look closely enough. Personally, I don't much care, but the true videophile is going to want more. How much more? Recall that IMAX is multiple gigabytes per second uncompressed, and (at a guess) around a gigabyte compressed, still over twice infinite capacity.

I doubt, though, that higher definition is why most people will want a bigger pipe. I suspect most of us who want a bigger pipe just don't like to wait. Sure, I could watch HDTV over a mere 50Mb/s connection, but who wants to wait for a program to download in real time? We want it now (because we're going to watch it later -- go figure). Now even the infinite pipe doesn't look so infinite. It's still going to take 10 minutes to download a feature film in high definition. Of course, if I can watch a movie while I'm downloading 10 more, that might take some of the sting out of it.

If you buy the "no-waiting" theory then we don't want or need infinite sustained bandwidth, but we would really like infinite peak bandwidth. Give me a gigabit pipe, and I promise I'll hardly ever use it. Or maybe I will. Usage has a way of catching up with capacity, and it seems risky, at the least, to predict otherwise.

Finally, what about the speculation in the original quote? The broadcasting solids idea has an obvious weak point: you'd need a pretty special printer to go with that (modern machine tools can reproduce solids described digitally, by any of several processes, but they're not available for home use).

As to the others: We actually do have something like the "television immediacy and newspaper depth," at least by one plausible definition, in the blogosphere and in conventional news sources. The high-def New York Times is certainly feasible. You could even print it out if you really wanted to. It's another question, and an interesting one, why no one seems to be putting out that sort of content to the mass market.


earl said...

A couple of things flop to mind. Obviously the amount of bandwidth the human brain can get any real good out of is finite, and has probably already been exceeded. This is not a problem for the human brain, which, rather than trying to drink from the firehose simply lets most of it go down the drain. Thus oversupply fails to slake our dissatisfaction.

And our capacity for dissatisfaction (arguably the source of all progress) may not no limits. It may really be that nothing is ever quite tall, fast, clear enough. Watching a baseball game on an 80's model 21" (diagonal) crt set you can see the seams on the ball on the slomo replay as it falls into the first baseman's glove. Much better than what you would see if you were at the game, but not nearly, nearly good enough.

I suppose you could always use up the capacity by insisting on speed: A full length movie realisable as an animated solid indistinguishable from real actors in your living room (they would even refuse autographs), downloadable in a nanosecond. That's pretty close to a lot, isn't it?

David Hull said...

Note to self: 3D printers