Readers critique the science in David Weinberger's "The Myth of Interference."

Published March 18, 2003 8:30PM (EST)

Read the story.]

The author of "The Myth of Interference" article has it wrong.

Radio frequencies are considerably different than light. The entire radio spectrum occupies the range from a few kilohertz to a few hundred GHz. Of that entire range, only a very small portion allows reasonable size components and antennae, and still smaller areas allow waves to propagate beyond line-of-sight with any reliability.

The author proposes a system that must communicate both ways, so the receiver can exchange information with the transmitter. We are not talking about one receiver; we are talking about millions of receivers that would suddenly become transmitters with similar range to the original broadcast transmission.

In the very narrow spectrum that allows propagation over large distances, efficient antennae must be very large. While a very small, very inefficient receiving antenna will hear strong signals, the same antenna could not possibly broadcast any useful signal back to the transmitter.

In order to reach both ways, power levels must be similar. We would have millions of receivers with internal high-power transmitters, all of which would broadcast back to millions of receivers at the transmitter so the transmitter knows "how it sounds" to all of those receivers.

The author, like many people today, thinks in terms of perfect worlds with no physical limitations. Unfortunately, the radio has to exist in the real world.

In the real world, very simple devices like power supplies radiate considerable radio interference. Hard-wired computer networks cause interference to communications systems, often for many miles. Unintentional electromagnetic pollution of radio frequencies is considerably worse now than it was even 10 years ago, and the author suddenly wants to add millions of intentional transmitters to the mix.

Even if the interference problems could be handled, the power and physical size of portable receivers and antennae able to communicate back similar distances to 50,000-watt transmitters would make hardware requirements unworkable.

It would be much more productive if Reed and other "architects of the Internet" spend time finding solutions to EM pollution caused by switching power supplies and digital systems, rather than proposing ways to make problems worse in areas they clearly don't understand. We must always understand how a system works and what the problems actually are in order to improve the system.

-- Tom Rauch, RF Design Engineer

His opinions are certainly enlightening and newsworthy, but what David Reed neglects to mention is that the "polite," agile radios that would allow the free use of open spectrum can't be built yet. The signal processors needed to make radios polite to one another are years off in the future. We may need to wait through generations of advances in semiconductor manufacturing to develop enough processing horsepower for this purpose.

Even when these processors do become available, older, impolite radios will make the polite ones less efficient, lowering the effective throughput Reed predicts.

Reed's observations are excellent, and he made a convert of me several months ago, but he should find out more about what the current obstacles are to real-world applications of his theories. The obstacles don't just consist of spectrum license holders and regulatory bodies.

-- Alan Morrison

While one of Reed's suggestions (frequency hopping) is certainly useful for making use of unused spectrum on an immediate basis, it doesn't really address the problem or the natural solutions to what is presently perceived as a limited availability of spectrum.

First, as anyone with a wide band receiver (even to as low as 1GHz in frequency) will quickly discover, there are broad swaths of spectrum that can be monitored for days without a single signal popping up. These portions of the "rainbow" have been reserved by the FCC for one use or another, but remain largely unused by those for whom they were reserved. Many of these owe their existence to a legacy of thinking of bandwidth as something to be assigned rather than managed efficiently.

While shortsighted, this mode of operation was easier to comprehend at a time when supply and demand didn't require the FCC to make a better effort on the management side. As demand has increased for bandwidth, more and more users are being packed into the same slots of spectrum -- all to avoid having to deal with industries and agencies whose underutilization of their assignments should be questioned and the contentious meetings and rationalizations that would result.

Unused but "unavailable" spectrum aside, the second problem focuses more on the technology being employed by those many new users who are being packed into the remaining spectrum slots. Perhaps the most contested and expensive slots today are those whose assignments are for "common carriers." Bidding for (and sometimes reneging on payment for) limited spectrum made available by the FCC to the cellular industry has turned the federal government into an overpriced auction house.

In reality, there is no shortage of spectrum for this industry. For the same reason that many low-power regional broadcast stations can all share the same frequency without ill effect, there is no technical reason why a much larger number of cellular users cannot share a very limited amount of spectrum. The problem is that a combination of cellular handset and site radio equipment, and antenna placement and height conspire to spread a signal on one frequency so much farther than is really necessary as to make it unreusable by another cell site at a reasonable distance away.

Imagine, if you will, an increase in cell sites to perhaps 10X what we have today, each with a much more limited range, and as a side benefit, resolving the problem of the many "black holes" in cell coverage that many current subscribers experience daily. There isn't any technical reason why every third or fourth cell site in a line can't reuse the same piece of spectrum so long as it is sufficiently distant from any other site using the same frequency -- far enough apart to avoid interference that can't be readily sorted out by the handset or cell site receiver. Additional benefits include the fact that towers could be lowered or better disguised, handset transmitter power could be lowered substantially (a concern to some already, however much the science may disagree), and the spectrum used over and over again within much smaller geographic areas -- "micro cells."

Why are cellular providers using the "clear channel" approach instead of the "local broadcaster" approach? Money. Cell sites, even equipped for handling what could be a lower capacity each, have certain fixed costs. That said, we hear a great deal of weeping over the cost of spectrum. It is evidently not yet so precious as to convince providers that a capital investment in equipment to solve the problem doesn't outweigh the cost of spectrum bidding wars conducted by the government.

-- Chris Anderson

I was very disappointed in the science presented in this article. While it is true that a pinhole camera does produce an image of the outside world, that does not void the issue of interference.

To carry the analogy to the next level, where interference does occur, consider poking another hole or two in the side of a pinhole camera. Now, each pinhole will be presenting an image of the outside world, but the images will interfere with each other.

To carry it even further, as the number of holes increases, the image will gradually fade to a blur. In the extreme case (remove the side -- an infinite number of pinholes), the inside of the camera is illuminated by reflected light. The information from the myriad pinholes has been lost in the sea of noise.

-- Bryan Ewbank

David Reed is wrong in a very narrow technical sense. Electromagnetic waves do exhibit interference: think of the famous 2-slit experiment in quantum mechanics. Furthermore, as I'm sure he is well aware, Claude Shannon proved that there is a maximal amount of information that can be carried in a given bandwidth. Having made those technical points I'll concede that the information content of radio or even TV programming would allow us to use much, much narrower bandwidths if we replaced our current radio technologies, and here is where he makes a more significant error.

He glibly dismisses the huge effort that would be required to shift this basic consumer technology. Look at the pathetically slow adoption of HDTV by consumers. My household hasn't been in any rush to pick up HDTV or satellite radio. We certainly aren't in any hurry to replace our eight or nine radios.

-- Charles E. Grant

By Salon Staff

MORE FROM Salon Staff

Related Topics ------------------------------------------