Quoting Trent W. Buck (trentbuck@gmail.com):
The only reason I can think of that a cellphone's transceiver couldn't do the same thing, is 1) the FCC are scared of consumers using (pre-built) programmable transceivers to "shoot down planes and stuff"; and 2) it'd cost the hardware vendor an extra twenty cents per unit.
Here in the USA, there was for many years a low-level flamewar between Linux activists saying there's no reason completely open-source radio stacks (for open source mobiles, software radio, etc.) would violate the law and a coalition of (1) proprietary-software apologies (notably Atheros people) and (2) a small but vocal group of control-freak amateur radio ('ham') people claiming it's against FCC regulations to permit open source. As you probably suspected, I was in the former group. The other folks spun out an endless stream of rhetoric, dire predictions of doom, and what seemed tantamount to threats, for years never _quite_ specifying what they were talking about. (We of the first camp remained polite but detected the whiff of metaphorical rodent.) I finally chased down ex-Atheros engineer Jim Thompson on this topic on July 27, 2006. The conversation remained twisty and noisy, but made some progress. Part of it was public: http://lists.svlug.org/archives/svlug/2006-July/050194.html Brief except from the even more noisy _private_-mail discussion (my side of an exchange with Thompson) follows. A 'tl;dr' summary would be: It's bullshit, but an informal conspiracy of nervous radio people and paranoid corporate interests _grossly_ misinterpret an FCC regulation designed to prohibit "user accessible controls" likely to be misused to violate spectrum assignments and similar provisions, as (supposedly) prohibiting any implementation of frequency-related code in open source. So, in short, it's less FCC being scared than the industry self-censoring and bullshitting the open source community. ---<begin long dialogue>---
Here are the relevant sections of CFR 47 Part 15 (http:// www.access.gpo.gov/nara/cfr/waisidx_01/47cfr15_01.html)
[snip]
(b) An intentional or unintentional radiator must be constructed such that the adjustments of any control that is readily accessible by or intended to be accessible to the user will not cause operation of the device in violation of the regulations.
So, Atheros's interpretation of CFR 47 15.15(b) is that "controls readily accessible or intended to be accessible to the user" encompasses _hacking and recompiling_ C driver source code? Sorry, that's laughable. By the standards of past FCC declaratory rulings, that's about as much a user-accessible a "control" as a radio's solder joints are. But it's pretty much the assertion I expected to hear, once some industry spokesman finally bothered to specify what "the FCC won't let us" means. It basically means "Well, if you stand on your head the right way and cross your eyes, you _might_ be able to argue that the FCC could conceivably disallow anyone outside the firm to see how we do this, so screw the open-source community. They'll never even ask for specifics, anyway." [later:]
Does the HAL have to be proprietary? This is the subject of some debate.
I would be very interested in seeing a polite and well-informed discussion on this topic. Maybe at a LinuxWorld BoF? Part of the problem is that, much as you say Atheros (and possibly other companies) has extensively analysed what may be done within the confines of regulatory law, outsiders have not been able to see or assess any of that.
I'm not speaking for Atheros here. I think the regs are clear. You may be unaware that the FCC has fined people for merely showing up at a tradeshow with non-authorized equipment (even non-intentional radiators) without the requisite sticker afixed to the bottom of the unit.
No, I'm well aware of such things. (I was assistant chief engineer for WPRB-FM in Princeton, New Jersey, once upon a time -- a commercial 17kW college station.) I'm just saying that I've read quite a number of FCC declaratory rulings and rulemakings (administrative law), and think their classing driver source code as a "user accessible control" is a ludicrously far-fetched interpretation. Did Heathkit Company give me unlawful access to a "user accessible control" when it sold the plans and build instructions to amateur radio transceivers? Because there was more than enough information in their detailed kit instructions and schemtics to modify them to intrude on the adjacent aeronautical navigation bands. Hell, with their plans and parts for even the Heathkit model FM-3, I could probably rig up a modification to that receiver unit to listen _and transmit_ on TV channel 6, without a lot of trouble.
Nothing would stop anyone even half motivated to change the hal source such that "underused" spectrum could (and would) be used. Atheros went as far as they could. You want more? Change the law.
The notion that the law objectively required this is basically what I class as a knee-jerk reaction from a company lawyer who thinks "It's not a _credible_ risk, but I'm covering our asses when I include this in my report to management -- and who's paying us to even have to think about this, anyway?" [still later:]
Your amateur radio transceiver is not a part 15 device. Amateur Radio is regulated under Part 97, not Part 15.
You seem to be going rather far out of your way to evade my point: Schematics and plans for Part 15 devices from any firm at all would suffice. (And many kits in sundry FCC regulatory categories from Heathkit Company and others are unlicensed devices.)
OTOH, the source code for the HAL is the >only< (and I do emphasize**only**) control available for the Atheros parts.
Again, you are ignorning the point: You are stretching credulity past reason to suggest that hackable source code is a "user accessible" control. Hacking C source is about as much consumer accessible as radio frequency crystals or solder joints. If you think otherwise, please show me an FCC docket where they sanctioned somebody for making available source code and declared it a "user accessible" control.
Hell, with their plans and parts for even the Heathkit model FM-3, I could probably rig up a modification to that receiver unit to listen _and transmit_ on TV channel 6, without a lot of trouble.
Yes, and this is specifically covered by the regulations, (73.525) but you (or rather the chief engineer) were in-charge of making sure that your station was in compliance. If, on the other hand, you were going to runs this under your HAM license, you'd have a different license violation on your hands, since TV ch6 is centered at 83.25MHz (82.0 MHz - 86.75 MHz, actually if you include the audio component), and meanwhile the HAM bands jump from 50-54MHz (6m band) to 144.0-148.0 MHz (2 meter), so you'd be in violation of your (HAM) license if you so much as keyed up the transmitter on 82.25MHz.
And, of course, if you were going to do it with no license, the FCC has you covered there, too.
Once again, you are completely ignoring my point: FCC would not held Heathkit to have violated regulations covering any of these spectrum segments by merely selling me schematics and plans, no matter how detailed. No matter how... how is it you say in English... hackable.
Of course, your apparent willingness to do this is exactly the problem. Far too many people would simply recompile the source code to "allow" themselves or others to transmit anywhere between 2.3GHz and 2.6GHz and/or the upper band I described before.
And _they_ would be violating FCC regs. Thank you for making my point for me. You're welcome to draw the obvious analogy to DeCSS, any time. ;->
The lawyer-in-question (at Atheros) is actually very "open source" friendly.
Suggestion: He or someone else familiar with Atheros's perspective should write up the matter for the benefit of interested laymen -- "outsiders" in the exact sense used earlier.
And yes, the law requires it. I've already cited chapter and verse.
And a beautiful non-sequitur citation it was, indeed! Fans of legalistic analysis the world over will appreciate it. ---<end long dialogue>---