Thak you all for reading our FAQ and posting comments. A few quick notes:

1] We are working towards a demonstration project. This will not happen over night. For example, we need end user devices which send, receive and route. Currently the FCC frowns on routers, which makes deployment of a test case in the United States very hard.

2] If eletromagnetic information interfered, pin hole cameras would not work. They do. Read the FAQ carefully on interference. Also please read the supporting documents, esp. those on David Reed's pages. See also the BLAST work on reflection that were once considered a form of "interference".

3] By using underlay technologies, not yet allowed by the FCC, you make it possible to phase in Open Spectrum without requiring a massive one time simultaneous upgrade.

4] Deliberate intereference is what makes fixed spectrum solutions less secure. I want my tax dollars to buy the government the most secure, robust and resilient communications money can buy. I suggest the current fixed spectrum policy does not pass this test.

5] What is the limit of the spectrum's CAPACITY to carry information? David Reed has been looking for years for an answer. We do not know of one. The BLAST experiments suggest it is vastly larger than current approaches would suggest. I suggest you ead up on BLAST. Bob Lucky suggests it is not impossible that it could approach the infinte -- see NY Times article, last paragraph.

6] I suggest you study Motorola's Canopy Technology. It is a significant step above and beyond Wi-Fi in that Canopy offers scalable capacity. Wi-Fi does not. Today Canopy is an off the shelf solution. It works today. It is not OS, but it is another step in that direction.

7] We did not go into either the First Amendment implications of current FCC policy or the changes, pro first amendment, Open Spectrum could make possible. We hope to post comments on the First Amandendment issues shortly.

8] Historians might want to look at the FCC 1981 Spread Spectrum NOI, A very interesting precursor to someof the ideas in the FAQ.

Thank you,

Jock Gill

I was thinking about interference and rivalness last night and I realized that there were two types of interference, that which is caused by surrounding structures and other objects, and that which is created by other electronic devices. (I realize that this isn't hard to figure out, but I'm a bit slow sometimes.) The first is the type of interference WiFi and other products that use unlicensed spectrum come across, yet the first type doesn't make the spectrum rival because it is not my use that is limiting your use, it is other outside factors. Because there is unlimited carrying capacity on other frequencies not being able to use a specific frequency does not lead the spectrum as a whole to be rival.

It is the second type of interference that would lead the spectrum to be rival, yet it seems as if this second type of interference doesn't exist. So the spectrum wouldn't be rival. When we are talking about interference we need to be sure we know what sort of interference we are discussing.

-Grant M. Henninger
[ dram.teamslack.net ]

As I was writing my last comment a thought struck me, if wireless bandwidth is unlimited when we use smart radios and do all the other neat technological tricks the FAQ is talking about, why don't we try to get the FCC to release other thin slivers of the spectrum into the unlicensed domain, ones that are in different parts of the spectrum so they can be used for different purposes. If the FCC would release five or six slivers of spectrum just like the one that WiFi uses then the technology could develop and its use could be demonstrated. There is no need to try and open all of the spectrum, if the spectrum really is non-rival we would be able to do everything we wanted with very limited parts of the spectrum. This of course is not a defense for the way spectrum is licensed today, but it is an idea on how we might proceed to open up more of it so these new technologies can develop.

-Grant M. Henninger
[ dram.teamslack.net ]

Edited by author 01-23-2003 09:28 AM
Grant - a couple of comments. It's important not to confuse the terms bandwidth and "bit rate" (as so many do in hip-digerati slang). Bit rate is the capacity to carry information. Bandwidth is the width of a frequency band in the spectrum. There is a limited amount of spectrum. My point is that the capacity of that spectrum, as the number of transceivers (antennas) scales, is unlimited. To be precise, any limitation results from limitations in the architecture of the system as a whole, not in the physical spectrum.

Second, there are other measures of economic utility that we might want to optimize. For example, the total amount of information that can be sourced in a certain amount of time at a particular point in space is still limited by bandwidth, so the "open one sliver" idea doesn't seem to be very useful in the long run. It puts a cap on any one user's use.

Third, your comment about structures and other kinds of interference. Structures and so forth actually increase the information carrying capacity of a medium. They do this two ways - one, by absorbing signals they increase the spatial reusability, and by diffusing signals they increase the value of adaptive space-time coding systems like BLAST. Even thermal noise has spatial structure, which doesn't increase information capacity per se, but by adapting to it dynamically, one can squeeze even more capacity out of a fixed amount of bandwidth.

The key idea here is that "bit transport over EM waves" is not necessarily a rivalrous good. No more than audio speech in the open air.

Edited by author 01-23-2003 02:28 PM
2] If eletromagnetic information interfered, pin hole cameras would not work. They do.

OK, so why do you need the pinhole, then? Why not open up the whole front of the camera, to let more light onto the film? Could it be that light impinging on the film from all directions has some undesirable effect on the film's ability to record the information you want that's coming from one particular direction? The answer is that at some point, some "sensor" has to turn the electromagnetic information into an electrical signal, or a chemical change in a dye molecule, or something else that we can use. That sensor has a certain range over which it operates; too much extraneous "electromagnetic information" impinging on it makes it impossible to recognize the information you want. Maybe you don't call that "interference", but I don't know a better term for it.

Read the FAQ carefully on interference. Also please read the supporting documents, esp. those on David Reed's pages. See also the BLAST work on reflection that were once considered a form of "interference".

I don't know anything about BLAST; do you have a link to a good reference on it? I'm going to go back and re-read the FAQ and the Times article, to see if there's anything there I missed.

I also don't understand how the bandwidth available in a finite volume (say, the Earth out to 100 miles above sea level) over a finite frequency range (say, 0 Hz to gamma rays) can possibly be infinite. Unless you assume that your transmitters and receivers are infinitesimally small and infinitesimally far from each other.

(Edited to correct formatting.)

In reply to the question on pinholes, David Reed writes:

Light is EM radiation, a form of radio. The point [of the pinhole example] was that information encoded in EM radiation doesn't take up space, because EM radiation itself doesn't take up space (it's not "rivalrous" - it doesn't exclude).

Physics says this in a simple way: Bosons have no exclusion principle. Photons are bosons, whether they are RF photons or light photons. The only way that photons interfere with each other is when the same photon encounters itself traveling on another path. If it crosses any other photon, it doesn't interfere (this is the physical definition of the term interference, not the FCC or EE definition, which includes the radios in the game).

This is different from particles like fermions, which include massy particles like electrons. Electrons obey the Pauli exclusion principle - you can't have them pass through each other. Sending lots of electrons through a pinhole does cause problems.

The pinhole thing is NOT an analogy. It's the real deal.

If all you needed was the pinhole, all would be well.

But you have to collect the light.

Both silver halide crystals and the electrons in CCD cells are rival.

Similarly, the EM radio waves will happily pass through each other, but when it comes time to build the antenna and receiver, things are not so clean and simple.

You are quite correct about interference being something that occurs in regard to receivers. For that reason, a demonstration system for Open Spectrum is not just a good idea - it is the only thing that can validate the theory. Until you build it, it will be hard to convince people that it can exist.

A separate question is whether or not you can build it affordably. I can probably figure out how to cobble together an AM radio with a few cents worth of parts. How much will your receiver cost - and is the cost worth it, both in the value of the transmitted and received info and the harder-to-quantify benefit to the public good?

Things like CDMA are a step in your direction, right? One of the things done there is power control - this avoids having to try and discriminate between two signals that are several orders of magnitude different *at my receiver*. Similarly, I'm curious to see how a single implementation would deal with the disparate problems of downtown Manhattan and northern Alaska, especially regarding transmit powers.

A final thought: is something like the Cybiko and it's messaging mechanism a good starting point for people trying to get a handle on your ideas?

Jock, you (and David Reed) are right that the photons don't interfere at the pinhole; my point is that they "interfere" at the film. Are you familiar with the response curve for photographic film? It's only (roughly) linear over some limited range of exposure; above that region it saturates. As far as I know, all detectors exhibit some sort of non-linear saturation effect if exposed to too many photons.

So as you say, photons don't interfere, but as soon as you try to convert the information into another form besides photons, you run into detector sensitivity issues. If you want to coin some other term besides "interference" for this, I suppose you could, but why not use the term that has been used to describe this phenomenon since its discovery?

Not including "the radios in the game" kind of defeats the whole point of using photons to communicate information, doesn't it?

And dpreed, isn't "audio speech in open air" a rivalrous good if you get into a shouting match? All it takes is one spoilsport screaming at the top of his lungs to ruin your nice conversation. For that matter, trying to converse with someone while another conversation goes on between people in the middle is harder than conversing in the absence of those other people.

I agree with the basic points that "Spectrum regulation should recognize physics" and "Spectrum regulation should recognize rapid change and learning, especially technical innovation". It doesn't follow that "Spectrum regulation should be eliminated".

When Open Spectrum advocates use misleading arguments like "radio signals don't interfere with each other" and "bandwidth is infinite", I get the impression that either they don't understand the physics as well as they claim, or they're trying to pull something.

Oh, and the BLAST page is at http://www.bell-labs.com/project/blast/

With respect to detector issues:

We agree. The problem is one of receiver design and architecture which introduces problems. It is not an issue of EM interference. We need to focus on the "detectors" and using hte term interfernec badly confuses the issue.

And we of course include legacy receivers by using "spectrum underlay approaches" -- if the FCC will permit. Co-existence is a requirement and technically possible.

BTW: Reed has re-stated the basis fior this dialog in this way:

The real question is: if we look at the potential technologies that can be built to sustain wireless communications over the next 10-20 years, will we end up with more capacity and more flexibility and more innovation and more growth if: 1] we preserve the narrow-band, prescribed-in-advance frequency-based access control scheme that has been used since the first decade of the twentieth century, or: 2] can we do better with a more dynamic, networked, interoperability-based approach that is now possible with software-defined radios, digital very-wide-bandwidth signal processing techniques, ability to measure the EM propagation environment and build models of it in computer memory, etc.

The answer is obvious to me and to most who have looked closely at the problem.

The question is how quickly we can discard the legacy based technologies, replacing them with much better approaches, so that we can meet the rapidly growing demands for wireless digital communications.

It's the same question we asked in 1975 when we began the Internet project - should we continue to assume that telephony and the Bell-system circuit-switched, isochronous architecture is appropriate for all future communications needs, or should we build a system that assumes that computer-to-computer communications will dominate the demand, and that interoperability of all networks creates more value than balkanized, highly optimized networks per telecommunications service.

People thought the Internet was a joke, and that the mature Bell System was a technological wonder, never to be surpassed in its engineering quality and suitability to purpose. It was great. Then. When one had dial telephones.


Lastly, I recommend to for your consideration our new post on
Open Spectrum and the First Amendment. -- JPG

Your FAQ is improved from the first version, but it still seems to have "pie in the sky" elements. Some specific questions:

"5... minimum set of rules..." This is the most crucial element in the whole FAQ. Everything else is just a sales pitch. What kind of rules are we talking about here? Obviously you can't just let everyone shout. Do we know what the rules need to be? Do we need ten years of research before we will know? Can we open up just a bit of spectrum to start learning the rules? Until you can firm up this question the FAQ ultimately fails to be convincing.

"12... BLAST ...". I looked at some of the BLAST literature. They use multi-element antennas (essentially like having multiple antennas and looking at how the received signals differ). That's great, you can get directionality, you can look at different path effects. But it would work best for stationary installations where the differentials between the antennas have a fixed meaning. Yet obviously for wireless one of the main application is mobile communication. Will BLAST work for mobile systems? It looks like it would be a lot more difficult because the differentials between the antennas would be constantly changing.

"15... unlicensed spectrum..." This answer is very confusing. Above you said that open spectrum needs some rules. Here you say that unlicensed spectrum has some rules. How are they different? You don't say. Further you go on to contradict yourself: "The lesson [of WiFi]: opening spectrums enables innovation." So WiFi is an example of "opening spectrum"? But WiFi is unlicensed spectrum, which you just finished saying is NOT open spectrum. So is it or isn't it an example? This all ties back into your failure to elucidate the necessary "rules" for Open Spectrum. Since you haven't identified the rules, you can't differentiate it from unlicensed spectrum.

"16... keeps us in a permission economy..." Sorry, but this is just rhetoric. In practice, you need to show that Open Spectrum will work by demonstrating it in some bands before you can make a case that everything should be opened! That's the real world. This kind of revolutionary talk may sound good at your meetings, but keep in mind that ultimately this is an engineering question. Wild throw-off-your-chains language does not help your case.

"17...accepts the current metaphors and paradigms..." Here we go. The bullshit detector just pegged. I'm sorry to be rude, but this rhetoric makes you look like a bunch of wild eyed fanatics, not someone to be taken seriously.

"29...still be able to watch The West Wing..." How can this be, if we are going to change all the rules? Will the broadcasters still have a monopoly on sending strong signals in those bands? How else can we still receive their transmissions on our old TVs? And if so, then what does this mean for your rhetorical promises to throw out all the old paradigms and bring in a whole new approach?

I hope you can accept these comments in a constructive manner. I would like to see experiments that prove that the new technologies can work in practice. But you need convincing arguments, which means avoiding the extremist and unrealistic rhetoric that you slip into far too often. You need to temper your Utopian vision of a revolutionary approach to spectrum by accepting that it has to be done gradually. Ultimately that is the path which will bring your closer to your goals.

To get an idea of what Open Spectrum can expect from the broadcast industry check this out:

http://www.nab.org/newsroom/issues/lpfm/responsetofcc/

And for a real hoot listen to these audio files of "actual" interference:
http://www.nab.org/newsroom/issues/lpfm/re...cc/noiseandhiss.asp
http://www.nab.org/newsroom/issues/lpfm/re...technicalrecord.asp

The NAB blanketed Congress with these bogus CDs and the scary part is that Congress bought it and passed "The Radio Broadcasting Preservation Act of 2000", effectively killing low power radio.

The interference tests were based on the following 28 radios:
http://www.nab.org/newsroom/issues/lpfm/studycomparison.asp

Prediction - A new bill called "Broadcasters' Public Service Preservation Act of 2003" (or something like that) will be introduced soon. NAB chairman Eddie Fritts (Trent Lott's former college roommate and best buddy) will plead with Congress to save free over the air TV from those Open Spectrum commies with their Marxist style manifestos.

(Hint: tone down the "one world", "air is free", "End of Broadcast Nation" fanaticism).

Jock, you (and/or Reed) compare your Open Spectrum movement to the beginnings of the Interent. But the Internet wasn't started by people saying in 1975 "We're going replace Ma Bell with packet switching. Throw away your out-dated dial phones, voice over IP is coming!"

Think about the principle of "rough consensus and running code". Save your "overthrowing the current metaphors and paradigms" until after you've demonstrated something working. That sort of rhetoric is suitable for the VCs when the next bubble starts, not the geeks who are going to make things work in the first place.

jleader [?][Would you care to identify yourself?

A good point, but please consider this contradiction. In the early days of the internet there were no regulatory blocks to building demos that could create "rough consensus and running code". You could "just do it". Perfectly legally.

Current FCC regulations do not permit either 1] spectrum underlay or 2] repeaters in end user devices. Thus building demos which create "rough consensus and running code" is not legal in the US today. Thus we have to start with an attempt to make the demos legally possible - a political argument.

It is areal shame that the FCC 1981 Spread Spectrum NOI was blocked by parties with other agendas.

We are also looking at creating off shore "radio havens" in which to establish demos that would lead to "rough consensus and running code".

I would also refer you to the excellent work being done by the GNU radio people. Combine this with some of the new radio chips Intel is talking about and .....

Regards,
JPG

Cypherpunk,

I refer you to my earlier post re the requirement for co-existance via spectrum underlay. We are not proposing an either or dichotomy.

I also refer you the fact of the current political environment that makes testing illegal in the US. So we face first a poltical barrier which must be removed before we can address demonstration projects in the US. It would be a shame if these new ideas were developed overseas by competitors.

Regards,

JPG

> Current FCC regulations do not permit
...
> repeaters in end user devices.
> Thus building demos which create "rough consensus
> and running code" is not legal in the US today.
> Thus we have to start with an attempt to make the
> demos legally possible - a political argument.

Doesn't a WiFi access point operating in infrastructure mode act "like a repeater"? I don't claim anything more than a bare minimum of undertstanding the of the relevant regs, but wouldn't what you are doing be permissible as pow-power work within Part 15?

Part of my thinking here is that you are not going to build a "repeater" in the classic, analog radio, sense of that word. Bouncing everything you receive back out will eventually cause something to cause problems - it never stops propagating around.

If I understand enough here, you need to make clear when you *stop* sending stuff you have just received. But the very act of making that decision places an intermediary between the 'input' and the 'output', making your device more like a transponder, and less like a receiver.

As a separate issue, you have a very strange attitude here. You claim that you can revolutionize the world, but first you need your revolution vetted by law???? We might have a little more confidence in your creativity if you could find a way to do some simple tests instead of throwing up your hands and saying "oh, we're not ALLOWED to do that". Heck, put the entire experiment inside a Faraday cage. I was also under the impression that Part 15 depended on a complaints process, rather than a certification process - so proceed carefully, make sure you don't interfere (for maybe more than a second or so) until you get the basics worked out. More to the point, if what you are doing is always going to cause interference in existing receivers, then that is a serious issue, which I hope you would even recognize, given that you propose co-existence with current systems. If it won't cause interference, then it seems you should be able to operate inside Part 15.

Jock, my name is Jeremy Leader. Nice to meet you!

If I haven't mentioned it before, I strongly agree with aligning FCC regulations (and other government policy) with modern technological capabilities, I just object to disinformation and overblown rhetoric in support of that goal. I don't trust groups like the NAB because they twist facts to support their position; doing the same thing from the other side of the question is a poor way to try to gain my support.