Message: 24
Date: 6.1.95
From: <nicholas@media.mit.edu>
To: <lr@wired.com>

Digital Videodiscs: Either Format Is Wrong

Here we go again. Big guns and big stakes have been pulled into the fracas over two competing digital videodisc formats. Both of them will store studio-quality, full-length movies on a compact disc, but each format will do it in a slightly different way. And people are taking sides. Even big players from Hollywood are jumping in, as if they knew the difference between an angstrom and a micron.

These mature and savvy businessfolk, as well as the press, refuse to understand that the issue is not video, but bits. Bits are bits. The moving image is only one data type. Surely, we cannot expect consumers to buy one machine or technology for video, another for audio, another for data, and yet another for multimedia. The screwball idea of owning a digital videodisc, which is nothing more than a movie player, is tantamount to digital obscenity.

Certainly, we must increase the number of bits per square millimeter on CDs, but we also need to treat those bits as nothing more or less than what they are. We do not need to agree in advance on precise digital standards and formats, and we cannot speculate in advance on all conceivable uses. Instead, we need to agree on "meta-standards," a way of talking about talking about those bits.

Sound like double talk? It isn't. Listen. Today, you can store roughly 5 billion bits on one side of a CD. If someone provided the means of increasing that by a factor of 10, it would be absolutely terrific. But I hope that whoever comes up with that scheme makes those bits as flexible as possible. They may be used for video and they may not be. Don't use the typical "standards committee" mind-set to remove the potentially rich new forms of information and entertainment that haven't even been thought of yet.

Unstandard standards
The physical world is unforgiving, so standards are desperately needed. Nonetheless, we cannot agree which side of the road to drive on. Europe has 20 power plugs. And once standards are set in the world of atoms, they're nearly impossible to undo.

But the world of bits is different, more forgiving. Why can't the entertainment industry understand this? A string of bits can contain information about itself: what it is, how to decode it, where to get related data. Surely there are multiple applications and options for future digital formats. The world is not just about movies, movies, and more movies. We must not lock the format of the bits into a single standard and call it video.

By contrast, what we must get from the outset is the atoms: the diameter of the disc (the only variable that's not in dispute), the physical property of the small pits in the disc off which the laser bounces (as much as anything, an issue of choosing a wavelength of light that everyone can agree upon), and the thickness of the disc. If we don't agree on these, we are in very deep trouble. Although nobody is saying it, this is what the debate is really about, not video.

Beyond paperback movies
We must not forget the obvious: discs are round. Unlike a long tape, the geometry of CDs is intrinsically random-access, and thus interactive. We need to pay serious attention to what the physical standard will do for interactive applications versus just looking at a movie from beginning to end. For example: How does the disc rotate? Is it like an audio CD, changing speed and slowing its angular velocity as the head moves to the outer tracks? Does it rotate at a constant angular velocity like your magnetic hard disc? Or should it be variable and pump out more bits the faster it spins, up to some practical limit?

The biggest issue concerns single- or double-sided discs. Curiously, the debate is largely rhetorical, as proponents of two-sidedness do not propose to read both sides simultaneously with a two-headed player. One very appealing feature of reading the top and bottom of the disc at the same time is that doing so allows the sort of seamless multimedia applications in which one head "plays" while the other "seeks." This feature is not even part of the current debate.

CD's last stand
It's crucial to get this new CD right and not lock it into old-fashioned thinking. Why? Because it will probably be the last CD format we will ever see. Package media of all kinds are slowly dying out, for two important reasons.

First, we are approaching costless bandwidth. Shipping all those CD atoms will be too difficult and expensive in comparison with delivering the bits electronically. Think of the Net as a giant CD with limitless capacity. Economic models may justify CDs for music and children's stories, making them last longer because they are played over and over again. But for the most part, CDs and all sorts of other package media (like books and magazines) will wither during the next millennium.

Second, solid-state memory will catch up to the capacity of CDs. Today, it seems outrageous to store a feature-length film in computer memory, but it won't be outrageous tomorrow. Solid-state memory offers the important feature of no moving parts.

In fact, 100 years from now, people will find it odd that their ancestors used any moving parts to store bits!

So please, Sony, Philips, Toshiba, Matsushita, and all your partners in Hollywood, don't give us a digital videodisc. Give us a new medium to store as many bits as possible. Learn from CD-ROM and let the market invent the new applications and new entertainment customers want. We'll be much better off.

Next Issue: Affordable Computing

[Back to the Index of WIRED Articles | Back to Nicholas Negroponte's Home Page | Back to Media Lab Home Page]
[Previous | Next]

[Copyright 1995, WIRED Ventures Ltd. All Rights Reserved. Issue 3.06 June 1995.]