Which is why the setup that does the scanning is all open-source, with the standardized code set published with hashes so that observers can confirm the code is the "legit stuff" that 10,000+ geek eyeballs have looked at.
The geeks can have all eyeballed some very nice-looking source code, and the observes can see that the machine spit out a hash string that the geeks said that it should, but that doesn't mean the machine is running the source code the geeks vouched for (or that it is running it corectly). It would still be far too easy to make the machine look like it was doing the right thing but contain a 'secret surprise' that wouldn't be detectable without an in-circuit hardware debugger, if then.
According to these courts, somehody like Hastert can deep-six any state election protection law that state passes and "annoint" candidates.
Indeed, that's a terrible law. But the solution to that is to change the law, not to further compromise the integrity of the elections in an attempt to conform to it.
You will be using all open source tools, freely open for community and peer review.That will include ROM burning utility.
Damn, now I'll have to put my back-door-adding code into the compiler instead, so that when it compiles the ROM burning utility it secretly adds my malicious code to it. But of course the compiler is open source too, you say? No problem, gcc also has to be compiled initially (bootstrapped) by yet another program... and so on, all the way back to the first program that was written in hand-coded assembly. And even if you've carefully checked all the source code all the way back to there, you're still running on hardware that you can only assume does what it is documented to do... for all anyone knows, someone has placed a backdoor into the CPU itself.
In my precinct, you fill out an application for a ballot. When you finally vote, that is put, in order, into what you might call a notebook
Strange... in my precinct, they just look up your name and address in a roster and cross it out. Thus there is no record of what order people arrived in (the roster is arranged alphabetically). But even in your case I think the problem could be mitigated simply by not placing the ballot applications into a well-ordered stack.
The secret ingredient is furniture varnish. When you drink it, the varnish coats your intestines with an impermeable layer of lacquer that prevents any nutrients from entering your bloodstream for the next 72 hours. The result: eat anything you like, everything will have "zero calories"! (Side effects may include projectile diarrhea and starvation)
Second: the optical scan half of this equation should scan GRAPHICS of each ballot, store them for later review, hash them to prevent later tampering and make them available by the DVD load
That's an interesting idea, but if you can't trust the machine to give you a proper vote tally, can you trust it to give you correct ballot images? If I was Mr. Evil, I might reprogram my machine to return a different set of images that was more beneficial to my pet politicians...
The problem with this is that often you know in what order people voted on the machine.
How? I suppose you could hang out all day at the polling location and write down the name of everyone who came in and voted in a notebook, but that would be rather obvious, and unless you're a poll worker you won't have access to the voting record anyway.
You can't number the ballots like that, as it leaks too much information about who voted for whom. Keep iterating!
At least here in California, ballots are already numbered like that, and it doesn't seem to bother anyone. Once you've placed your ballot into the ballot box it would be very difficult for anyone to try to trace it back to you (since there is no record of when during the day you came in to vote).
... and when it's pronounced secure etc. - burn it to a ROM and disable any access to it which doesn't require at least a crowbar
You're trusting that the ROM burning utility doesn't have any "interesting undocumented features", such as, say, adding in a back door routine to the executable image it burns to the ROM. And since you've sealed up the ROM, it will be impossible for anybody to know that the ROM's code has been tampered with.
Tell me again why we need a GHz processor with multi threaded capiblilities, several I/O ports and memory types, floating point coprocessors, a system bus, and a full HAL just to do voteCount[candidateA]++?
Like most modern apps, the actual algorithm is by far the easiest part of the task. 99.9% of the computer's power will be devoted towards presenting a user-friendly interface. Keep in mind that voting machines are expected to be able to present ballots in multiple languages, read them aloud for the blind, etc.
Imagine my shock when I moved to San Francisco and was around for my first city election - 40% turnout, popn 800,000, with a first-past-the-post system (ie tick the canddiate you want, no preferences - talk about watered-down 'democracy' for retards)
If it makes you feel any better, San Francisco has implemented Instant Runoff Voting for local races now... so they're somewhat less retarded than the rest of us already;^)
Proving code mathematically correct is a nearly pointless excercise. Who then proves the proof correct?
It's even worse than that. Assume for the sake of discussion that the code has been proven correct, and the proof is also correct. How do you prove that the compiler compiled the code correctly? And if you can do that, how can you prove that the hardware will execute the compiled executable correctly? And finally, even if you can do that, how can you prove that a given machine is actually running the executable you think it is running?
It's an exercise in futility. The only real solution is one that doesn't rely on the code being correct... you have to print out the ballot and have the voter verify that the printed ballot is correct, and then have him turn the ballot in to be counted. That way it doesn't matter how buggy/evil the machine is... if it does something wrong, the voter will notice and complain.
Chernobyl gives hope to this end. 'I really expected to see a nuclear desert there,' says Ronald Chesser, an environmental biologist. 'I was quite surprised. When you enter into the exclusion zone, it's a very thriving ecosystem.'
For all you pro-nuclear guys trying to figure out what argument would bring the environmentalists around to your side, I believe this is the one. Just be sure to design the reactors to be as unsafe as possible.;^)
I'm saying I need a backlight in an e-book reader.
I'm not even sure you could make a backlight work with this kind of screen... I suspect the screen is not translucent. You'd probably need some sort of front-light instead. Presumably any solution that you would use for an old-fashioned paperback book could be applied to this also.
On the other hand, once everybody has an eBook, the printing market can become 100% bits-only, and there is no need to manufacture hardware anymore (except for the occasional replacement when someone breaks their eBook). Then you have the best of both worlds.... you don't have to chop down trees or create lots of electronics.
Re:My hands-on impressions of the Sony Reader
on
The eBook, Mark 2
·
· Score: 1
Sony really ought to take a page from the iPod's book on minimalism
Hm, I have a better idea... Aple should take a page from Sony and incorporate a nice eInk display into its next iPod. A device that delivered both audio and printed media, and didn't suck (in all the ways that Sony's products tend to, and Apple's products tend not to), would be a great thing.
I haven't actually seen one of the displays yet, but I'd agree that the refresh seems , uh, a little long.
Hopefully that will be improved in future devices, but even if it isn't, I think the limitation could be worked around fairly easily in software. For example, you could have a two-stage page-turning sequence: divide the screen into "top" and "bottom" halves, and have the machine update one half of the display while you are reading from the other half. Et voila, double-buffering to the rescue:^)
People, you might not want to hear it, and you might not agree with stupid knee-jerk reactions, but these reactions will be coming. The name "reiserfs" is tainted, whether that's rational or not.
Eh, that's easy enough to overcome. Namesys will just release the next version of ReiserFS as "ext5".
Actually, *informing* others that someone or some site has what you consider to be a "historical political bias" is not a legitimate act. It's a judgment call, and one that a news source ought not to make. As you said yourself, "They said it, and anybody can read it."
I disagree -- I think it's exactly this fear of making any judgment calls (to avoid being considered "biased") that makes the mainstream news so ineffective. Of course the news provider is going to make judgment calls about the reliability of its sources. It does that when deciding which news stories to run, and which sources to quote -- it can't then pretend to "report, and let the viewers decide", as if it was merely a passive data conduit. It's not, and presenting news without the necessary context is lying by omission.
Quoting someone about an issue in which they have an obvious conflict of interest and deliberately not informing the viewers about that conflict of interest (because they are supposed to figure it out for themselves) isn't news, it's propaganda.
While this "combats open source", it's really just the certification authority where "money = trustworthiness" stupidity applied all over.
Indeed. How long will it be before some company gets a driver signed that (intentionally or not) allows arbitrary code to be executed as a subroutine in its 'trusted' context? As soon as that happens, they're back to square one...
Here's a prediction for a long term movement. [...] Bio-diesel vehicles become popular [...]
That all sounds good, and I agree with eveything you've written above... the only question I have is whether bio-diesel (or any other oil replacement) will be able to scale up to meet the world's energy needs -- i.e. are there enough sites for algae ponds and fields for growing switchgrass to produce bio-diesel on the order of magnitude we are used to? And what will the unintended consequences (environmental and otherwise) of all this massive new agriculture be?
No matter what new energy source we replace fossil fuels with, there will be one fundamental difference: we'll no longer be able to simply extract vast stores of pre-concentrated energy, like we do now. Instead, we'll have to condense more dilute forms of energy (sunlight, wind, tidal power, nuclear power, etc) together into fuel, one way or another. Perhaps there is a way to do that efficiently enough that the first world's current lifestyle can remain largely intact, but I wouldn't bet that way.
Interesting. So somewhere on Earth there is a process that creates new oil as fast as we use up the oil we've recovered. Can you describe how that process works?
And the thing is, I don't accept that there's a peak of petroleum. Heck, if it came down to it, we could make petroleum from[...]
Being able to show a proof of principle in a lab is completely different from being able to run the world on that technology. Being able to synthesize oil for $500 per barrel means that in the end-game, oil will cost at least $500 per barrel. And that is the problem. If you come up with a way to synthesize oil so that it doesn't cost (an order of magnitude) more than what oil currently costs to dig out of the ground, let me know, I will want to invest:^)
Slashdot Mirror
The geeks can have all eyeballed some very nice-looking source code, and the observes can see that the machine spit out a hash string that the geeks said that it should, but that doesn't mean the machine is running the source code the geeks vouched for (or that it is running it corectly). It would still be far too easy to make the machine look like it was doing the right thing but contain a 'secret surprise' that wouldn't be detectable without an in-circuit hardware debugger, if then.
According to these courts, somehody like Hastert can deep-six any state election
protection law that state passes and "annoint" candidates.
Indeed, that's a terrible law. But the solution to that is to change the law, not to further compromise the integrity of the elections in an attempt to conform to it.
Damn, now I'll have to put my back-door-adding code into the compiler instead, so that when it compiles the ROM burning utility it secretly adds my malicious code to it. But of course the compiler is open source too, you say? No problem, gcc also has to be compiled initially (bootstrapped) by yet another program... and so on, all the way back to the first program that was written in hand-coded assembly. And even if you've carefully checked all the source code all the way back to there, you're still running on hardware that you can only assume does what it is documented to do... for all anyone knows, someone has placed a backdoor into the CPU itself.
Strange... in my precinct, they just look up your name and address in a roster and cross it out. Thus there is no record of what order people arrived in (the roster is arranged alphabetically). But even in your case I think the problem could be mitigated simply by not placing the ballot applications into a well-ordered stack.
The secret ingredient is furniture varnish. When you drink it, the varnish coats your intestines with an impermeable layer of lacquer that prevents any nutrients from entering your bloodstream for the next 72 hours. The result: eat anything you like, everything will have "zero calories"! (Side effects may include projectile diarrhea and starvation)
That's an interesting idea, but if you can't trust the machine to give you a proper vote tally, can you trust it to give you correct ballot images? If I was Mr. Evil, I might reprogram my machine to return a different set of images that was more beneficial to my pet politicians...
How? I suppose you could hang out all day at the polling location and write down the name of everyone who came in and voted in a notebook, but that would be rather obvious, and unless you're a poll worker you won't have access to the voting record anyway.
Amen, sister! I'll be right there with you. (well, in Pasadena, but it's close enough)
At least here in California, ballots are already numbered like that, and it doesn't seem to bother anyone. Once you've placed your ballot into the ballot box it would be very difficult for anyone to try to trace it back to you (since there is no record of when during the day you came in to vote).
You're trusting that the ROM burning utility doesn't have any "interesting undocumented features", such as, say, adding in a back door routine to the executable image it burns to the ROM. And since you've sealed up the ROM, it will be impossible for anybody to know that the ROM's code has been tampered with.
Back to square one...
Like most modern apps, the actual algorithm is by far the easiest part of the task. 99.9% of the computer's power will be devoted towards presenting a user-friendly interface. Keep in mind that voting machines are expected to be able to present ballots in multiple languages, read them aloud for the blind, etc.
If it makes you feel any better, San Francisco has implemented Instant Runoff Voting for local races now... so they're somewhat less retarded than the rest of us already
It's even worse than that. Assume for the sake of discussion that the code has been proven correct, and the proof is also correct. How do you prove that the compiler compiled the code correctly? And if you can do that, how can you prove that the hardware will execute the compiled executable correctly? And finally, even if you can do that, how can you prove that a given machine is actually running the executable you think it is running?
It's an exercise in futility. The only real solution is one that doesn't rely on the code being correct... you have to print out the ballot and have the voter verify that the printed ballot is correct, and then have him turn the ballot in to be counted. That way it doesn't matter how buggy/evil the machine is... if it does something wrong, the voter will notice and complain.
For all you pro-nuclear guys trying to figure out what argument would bring the environmentalists around to your side, I believe this is the one. Just be sure to design the reactors to be as unsafe as possible.
If you believe that, I've got a bridge to sell you. No, you can't see it. Trust me, it's really nice.
I'm not even sure you could make a backlight work with this kind of screen... I suspect the screen is not translucent. You'd probably need some sort of front-light instead. Presumably any solution that you would use for an old-fashioned paperback book could be applied to this also.
On the other hand, once everybody has an eBook, the printing market can become 100% bits-only, and there is no need to manufacture hardware anymore (except for the occasional replacement when someone breaks their eBook). Then you have the best of both worlds.... you don't have to chop down trees or create lots of electronics.
Hm, I have a better idea... Aple should take a page from Sony and incorporate a nice eInk display into its next iPod. A device that delivered both audio and printed media, and didn't suck (in all the ways that Sony's products tend to, and Apple's products tend not to), would be a great thing.
Hopefully that will be improved in future devices, but even if it isn't, I think the limitation could be worked around fairly easily in software. For example, you could have a two-stage page-turning sequence: divide the screen into "top" and "bottom" halves, and have the machine update one half of the display while you are reading from the other half. Et voila, double-buffering to the rescue
Eh, that's easy enough to overcome. Namesys will just release the next version of ReiserFS as "ext5".
You realize that somebody is bound to write a word processor that runs on a GPU now, just to prove it can be done?
I disagree -- I think it's exactly this fear of making any judgment calls (to avoid being considered "biased") that makes the mainstream news so ineffective. Of course the news provider is going to make judgment calls about the reliability of its sources. It does that when deciding which news stories to run, and which sources to quote -- it can't then pretend to "report, and let the viewers decide", as if it was merely a passive data conduit. It's not, and presenting news without the necessary context is lying by omission.
Quoting someone about an issue in which they have an obvious conflict of interest and deliberately not informing the viewers about that conflict of interest (because they are supposed to figure it out for themselves) isn't news, it's propaganda.
Indeed. How long will it be before some company gets a driver signed that (intentionally or not) allows arbitrary code to be executed as a subroutine in its 'trusted' context? As soon as that happens, they're back to square one...
That all sounds good, and I agree with eveything you've written above... the only question I have is whether bio-diesel (or any other oil replacement) will be able to scale up to meet the world's energy needs -- i.e. are there enough sites for algae ponds and fields for growing switchgrass to produce bio-diesel on the order of magnitude we are used to? And what will the unintended consequences (environmental and otherwise) of all this massive new agriculture be?
No matter what new energy source we replace fossil fuels with, there will be one fundamental difference: we'll no longer be able to simply extract vast stores of pre-concentrated energy, like we do now. Instead, we'll have to condense more dilute forms of energy (sunlight, wind, tidal power, nuclear power, etc) together into fuel, one way or another. Perhaps there is a way to do that efficiently enough that the first world's current lifestyle can remain largely intact, but I wouldn't bet that way.
Interesting. So somewhere on Earth there is a process that creates new oil as fast as we use up the oil we've recovered. Can you describe how that process works?
Being able to show a proof of principle in a lab is completely different from being able to run the world on that technology. Being able to synthesize oil for $500 per barrel means that in the end-game, oil will cost at least $500 per barrel. And that is the problem. If you come up with a way to synthesize oil so that it doesn't cost (an order of magnitude) more than what oil currently costs to dig out of the ground, let me know, I will want to invest