Well, his message is getting through. Salon has an introductory piece on the "there's enough bandwidth for everyone" utopian arguments of the oft mentioned David P. Reed. Still, the actual math is so hard - assuming you dig in deeper than the Salon article - that I doubt many people have much beyond a metaphorical understanding of what he's talking about.

Slashdot posted the story. Here's the +5 rated comments (the best comments.) They almost universally savage Reed (in the best/worst tradition of /. commenting.) The message here seems to be: "this guy doesn't know what he's talking about" which is sort of what I was scared of.

Still, while a close reading of all the objections is instructive, I do think most people misunderstand his point. No doubt Salon's intro isn't the best text to base a technical refutation on. Sure, we can't do what Reed says with today's radios. But Reed isn't claiming we can.

Part of what he's saying is that software defined radios are going to allow this sort of thing to happen. Assuming we can build (program) them. And assuming we can change the (soon to be) out of date broadcasting regulations that would disallow such devices. And the first step in changing such regulation is to have people believe that something better is possible.

So perhaps he's overly optimistic (as most /. comments complained) but I don't think he's wrong. Or at least not yet. Let's see how the software comes along over the next year. But a little optimism might well be warranted. I think getting the idea that something wildly better is possible into non technical people's heads will be for the better. And it sure won't hurt anything, since all these claims will have to be demonstrated eventually anyway. It's not like we're going to dismantle the FCC until there's some running code and working gear.
- jim 3-13-2003 12:15 am

Many of the comments on slashdot focus on dismantling his metaphors. He does oversimplify, but to a non-technical audience that seems pruduent. Perhaps he should sharpen his metaphors so that the non-techies get the drift and the techies don't spin off on diatribes about photons and Schrödinger.

Fundamentally, he's saying that the RF spectrum is very inefficiently utilized given that the allocation of much of the spectrum is based on the limitations of early 20th century technology. And that new communication paradigms based on new technology can and should make much better use of this resource.

- mark 3-15-2003 1:24 am


I know nothing about RF engineering, but he sometimes appears to say that the spectrum is not just underutilized, but actually infinite. It's this claim that gets people kind of worked up (some hopeful, some angry.)

David Reed responds on his (group) site.

Pretty interesting, but I'm still not clear if he's saying the spectrum is infinite (in a useful way.) But then again, I'm not sure it matters. As long as it keeps getting bigger as a factor of advancing radio technology.



- jim 3-19-2003 1:43 am


Reed's comments at that site are a much more clear statement of his ideas than the much longer article at Salon, assuming one knows about basis functions. But once I was able to confuse some grad student TA's at Stanford when I used the term "basis functions" in a discussion of the FFT vs. the DCT. And these were TA's for a course on signal processing. I would have even lower expectations for physicists and RF engineers to understand the mathematics emboddied in that term. Perhaps this is why Reed uses the metaphors that get him into so much trouble.

I wish I could write a succinct definition of "a set of orthogonal basis functions" but I'm not that good.

Here's a metaphor. The human cochlea is able to analyze a squiggly waveform of pressure variations and break this down into a set of "pitches" and corresponding "loudnesses".

The basis function concept is sort of like that. Replace "set of pitches" with "set of orthogonal basis functions" and "corresponding loudnesses" with "corresponding coefficients". It's a mapping from one domain to another. No information is lost in the process, but certain aspects of a signal may come into sharp relief.

Two ears working together can break down their inputs into "direction", "loudness" and "pitch". Your ears might use this analytic ability (plus some "symbol" oriented processing) to tell you "there's an oboe on the front porch playing rag time and a flute doing a jig in the kitchen". We need radios that can do that.


- mark 3-19-2003 2:39 am





add a comment to this page:

Your post will be captioned "posted by anonymous,"
or you may enter a guest username below:


Line breaks work. HTML tags will be stripped.