The first source you're quoting has nothing to do with whether or not humans cause a dominant CO2 rise in the atmosphere - it's just very long term trends: "what's going to happen to the Earth." It basically says "yah, the Earth's screwed in about half a billion years." This is ballpark what we knew already - planetary habitable zones migrate outward, and we're on the inner edge.
The second source stems all of its criticism of whether or not humans are causing the CO2 rise on a poor criticism - saying "ocean warming causes CO2 increase, so how do we know the CO2 increase is causing ocean warming, and not the reverse?" The answer to that is simple: we have a coherent model for warming due to CO2 emission. We do not have a coherent model for warming of the oceans causing CO2 emission. Occam's Razor chooses the first: the second requires an additional mechanism for generating ocean warming.
Actually, the second explanation ("ocean warming is caused by some unknown mechanism, which leads to a CO2 increase in the atmosphere") requires even more work than that: it also requires machinery to link the CO2 increase caused by ocean warming to that emitted by humans. The airborne fraction: that is, the fraction of CO2 emitted by humans that we see as an increase in atmospheric CO2 - has stayed pretty constant over 5 year averages.
Oh, and I saved my most damning point for last: you see, you can determine where the CO2 is coming from by looking at the isotopic composition of the CO2 in the atmosphere, and see if it looks like the isotopes from fossil fuels, or other sources (like the ocean). They're from fossil fuels.
The carbon dioxide rise is anthropogenic. There is no scientific debate about this fact anymore.
I don't see how, unless the systematic errors are of such low magnitude that they're negligible in practice anyway.
Depends on the level of random errors and the statistical sampling that's used. Besides, "negligible in practice"? I'd prefer to leave no tamper margin, not just negligible tamper margin.
If you assume more realistic fraudsters, who weigh the chance of getting caught against the probability that they'll affect a race, what you'll really have is no fraud (of the sort we're discussing here).
If I was a very malicious person, what you'd do is allow the electronic systems to gain credentials, and then slowly reduce the quality checks in the crosscheck to allow the random error rate to creep upwards. So long as you stay under the crosscheck error rate, there's zero risk. How can something be "suspicious" when it's identical to a nonfradulent election?
That is, imagine the electronic voting system compares to the hand count to 1 in 1000 over 10 years of elections. The malicious person simply makes it so that the electronic counting outputs a result that is different from what it actually counted by 1 in 10,000. It'd take a very, very long time to notice that tampering. Then, of course, when things aren't noticed after two or three years, you work to reduce the quality of the crosscheck, allowing your tamper margin to go up.
This is assuming a simple paper receipt that can't be tracked to actual electronic votes: that is, the "paper counting" process is separate from the "electronic counting" process. This is the bad portion I'm talking about.
Registration fraud offers a much better return for the investment.
Sure. That doesn't mean I'm not going to worry about direct vote tampering. Someone else can worry about improving registration procedures.
And how do you ensure that the voter can't be associated with the ID? The ID would have to be invisible to the voter, but available to the vote counter.
How can you ensure the voter can't be associated with a ballot? They still have fingerprints on the ballot, for instance.
The ID would just be a random number - just to associate an 'electronic vote' with a 'paper vote' to link the two counting systems. There'd be no information in the system about who the voter was which placed each 'electronic vote'. While I guess you could, I dunno, do a forensic analysis of the system or something and figure it out somehow, you could conceivably do that with a paper receipt as well. In fact, the paper receipt would be a heckuva lot easier to track.
The real concern would be trying to keep the ID invisible to voters, so that they can't prove their vote (although to me, this is less of a concern) - and that you could do by having the paper be special - revealing the ID with a UV lamp, for instance.
That would mean no nice touch screen, no easy-to-use step-by-step interface, etc
Er? No it wouldn't. I could implement those things on bare metal in less than a month. This is a very simple system. It's essentially a state machine - which means just a bunch of flip flops. The only reason it's being implemented in embedded systems right now is because Diebold is extremely lazy and cheap.
If we're going to talk about that, we should begin by discarding plurality voting and use a system that doesn't disenfranchise anyone who isn't a fan of the two most popular parties.
There is such a thing as a fair count. There is no such thing as a fair voting system. Arrow's Paradox.
The problem is that omnipotent does indeed mean able to do anything.
Common usage of omnipotent is "able to do anything."
That doesn't mean that's a usable definition. Common usage of the word "force", or even "work" doesn't lend itself to a useful physics definition. Instead those needed to be clarified significantly. Work is the clearest example.
"Able to do anything" isn't a workable definition to use in logic. It just doesn't work - it leads to things akin to Russell's paradox.
because I know that I can do anything that is possible for me to do.
I didn't say "possible for said being to do." I said possible. I can't continue living in vacuum, but it's certainly possible to do so.
It's worth pointing out, BTW, that it is far better to accept random errors than to enable deliberate errors. On average, random errors won't change the results.
My point is that random errors in your crosscheck can mask systematic errors in your primary count.
What we should have, of course, is a system in place for redoing an election in the case of too close a result. It's simply crazy to declare one person a winner if the results are within the margin of error of the counting system.
But barring that, it's dangerous to use a system with random error to check a system that can have intrinsically lower error. That allows a "tamper margin" below the random error rate in the crosscheck. That is pretty much my entire point. Note that you may think that it's silly to worry about a tampering method that might result in a negligibly small number of votes swung towards a given party, but multiplied over all of the races, and over many years, it'd be worth it.
Do we have evidence that paper ballot counting is zero error? It's simple enough to say that it is, but that's a presumption that you're making. And it's a very, very important presumption.
It's easy enough to test. Do multiple recounts. Are the results exactly the same? I didn't think they were in the Florida 2000 recounts - they were close, but they weren't identical. And that's unsurprising. People are going to make mistakes.
For each ballot, run it through the machine and get the machine result.
Right, this is pretty much what I thought you were saying before. I agree, this pretty much solves sub-threshold tampering - except that you're requiring a machine-readable paper ballot, and I don't think you need that. Just a link between the printed receipt and the count on the system.
Let me explain a bit in case you're confused - I think we were imagining different systems. What I was thinking of is a system like the Diebold system currently, but improved with a paper receipt, where the votes are totalled, and a paper receipt is printed. The votes are entered into some sort of database, or a separate count, and are nothing more than a *count*: that is "654 votes for A, 728 votes for B". The paper receipts are kept separate, and counted via random sampling to determine if that vote ("654 ballots for A, 728 ballots for B") matches the electronic vote.
I don't like that sort of system. It's crosschecking a system with very low intrinsic error with something which may have a higher error, which means you have a tamper margin.
That being said, though, I think you could fix that system by melding it with the benefits of mechanical reading.
What you do is tag each paper ballot with an ID, and instead of storing just a count, store the ID that you've printed electronically. This way you can determine any below-threshold tampering, because mangled and lost ballots are going to be lost separate from the system. This way you can't 'hide' the tampering, because you don't know what ballot is going to be mangled or lost, and you can compare each paper ballot with what the system thought the ballot was for.
So now you've got "654 ballots for A, 728 ballots for B", and an electronic record which has a list of 654 IDs for A, and 728 IDs for B. You can't "sneak" an extra vote in there without forcing people to assume that a paper receipt got lost, which, as you noted, is preventable.
I hadn't thought of that before. That's an excellent, excellent idea. I don't think it hurts the anonymity of the process at all, because it's just an arbitrary ID. It seriously helps the credibility and that would allow a closed electronic system.
Any such machine is a closed system, because even if it runs open source there's no practical way to ensure that the code installed on the machine
No software. Why do you need software for this sort of thing? You're counting. Implement the thing on bare metal. Software just leaves yourself open for th
This study does at least show that, if whatever pertinent deity exists, it cares more about its ego than the needs of people who may die as the result of an illness.
Ah yes, once again, the ugly specter of poor wording and the Problem of Evil rears its head. (*)
You've got an implicit assumption going from this study to your conclusion: you're assuming that it was possible that these people would survive.
In fact, this is pretty common with prayer: we assume, implicitly, that the impossible is possible, and that said impossibility has no consequences other than the immediate action. That is, it's possible for said sick person to just magically become okay, with no drawbacks. This is pretty silly. The person praying doesn't know what's wrong with the sick person. They're just assuming that it would be better if they survived than if they died. That's an unfounded assumption.
Note that this isn't presuming that prayer does nothing. It just can't change the inevitable. It's only our (false) hope that allows us to believe that certain things aren't inevitable.
If I were God, I'd be incredibly insulted by your statement. You're presuming a lot of knowledge about the Universe that you simply don't have. How, precisely, do you know what the needs of the person who was ill was? And, presupposing that the person's death was inevitable, how do you know that that person's death wasn't made a ton easier by said deity?
(*: The relation to the Problem of Evil is pretty straightforward. You're presupposing that "omnipotent" strictly means "able to do anything". This, unsurprisingly, causes problems with your logic, because "able to do anything" isn't a well-formed set. If you instead say that God is omnipotent, meaning "able to do anything that is possible" that clears the situation, and the resolution that I mentioned above - that there was nothing that could be done with regard to the ill people - is pretty clear.)
Nope. Not sure where you got that idea. I don't want votes to be traceable to any granularity smaller than the voting precinct. It'd be nice if they could be more anonymous than that.
Then I'm really confused how you'd detect a tampered ballot. In the end, out of a voting precinct, all you would get are two numbers: the amount that the electronic system records, and the amount that the paper recount has.
Given only that, I can't see how the paper ballots help you, since all you can do is compare those two numbers. The paper recount isn't perfect - hell, the paper can be lost - and so if a random paper recount is compared to an electronic recount, if the error rate is like "1" or "2" votes, that probably wouldn't raise a red flag.
So the election's still manipulatable with the closed system so long as the attacker keeps the manipulation level below that of the paper recount accuracy. If the paper recount accuracy is low enough, this could be a problem.
I guess that's kindof the crux of the difference here: what is the margin of error in paper recounts? Isn't it a little unsafe to just assume it's zero?
Unless you require that the margin of error in the paper recount be extremely low. I don't know how practical that is. It certainly doesn't seem to be practical in a lot of counties in the US.
And in any case, we don't really have the provisions to say "this vote was, for all practical purposes, a tie", which means that any level of tampering has to be a worry as well. It'd be easy enough for a manufacturer of a counting process to give a bonus to, say, all Republicans which would give, in a statewide election, only a few hundred votes margin their way. Negligible for most elections, but over enough time, enough to sway a few closer ones.
I should point out that the only thing I really disagree with is the necessity for an open counting system. I think so long as hand recounts have any inaccuracy (and for any physical medium, it will) you've got to worry about very subtle vote influencing. Which is why I think any "closed" system in the loop is a bad thing.
So, how many sample ballots need to be examined to catch you at this?
Are you suggesting a system where each ballot is directly traceable to an electronic entry - like, with some sort of ID? Is that actually suggested?
Ballots should be both human and machine-readable.
Wait. If they are machine readable, that means that the counting will be done invisibly anyway.
The fact that people can read the ballots is therefore pretty much academic, isn't it? In actuality, they won't.
In some sense it would be like forgery protection with cash. In theory it's nearly impossible to forge US money, because there are about a hundred safeguards in place. In practice it's easy to slip small amounts of fake money through, because no one actually uses the safeguards.
Although I guess if the ballots were tagged with identification marks and the final record basically consisted of grouping the identification marks rather than just printing out the total, then statistical sampling can save you there.
That sort of auditability is impossible with election systems because of the requirement that votes be anonymous. Since the ultimate originator of the vote *MUST* have no way to verify that his or her individual vote was properly traced throughout the system, we can't apply the same auditing techniques.
Let me just stress this again: you're definitely right that you can't audit the end-to-end system. But you can (and must) audit everything else.
Heck, with paper ballots that wasn't done in Florida in 2000, and that caused a large portion of the confusion - that is, the paper ballots weren't up to standards in terms of the presentation of information, and to add to confusion, they were misprinted slightly as well. Now, that was almost certainly from incompetence - but it could've just as easily been done maliciously as well.
So clearly, the design of the paper ballot has to be audited as well.
However, there's an interesting problem here: we worry about things like electronic systems that are maliciously programmed or manipulated. Why couldn't we have paper ballots that are maliciously designed as well? Even psychologically maliciously designed, such that the design of the ballot leads the casual voter to choose one selection over the other.
Granted, those problems are common between the two.
The main worry I have is that some people believe that a paper ballot solves everything. It doesn't. The votes must be anonymous, which kills any ability to audit the system end-to-end. But that doesn't mean that every other step can't be audited, and it must be.
But that's better than making it possible for one person to modify *all* the votes. Much, much better.
I think you're misunderstanding the kind of system I'd be thinking of. Are you thinking of a system where the votes are electronically transferred - that is, over a network?
I agree that would be really, really stupid.
But I'd be more thinking of a "paper replacement" electronic system - like the ones that are in use in several places (including where I am) where the votes are stored physically on some medium (in several copies) and then transferred to a central facility again. It's exactly the same risk-sharing as you have with paper - just that now you've removed (or reduced) the risk of the volunteers at the voting place.
Granted, the people actually voting can't clearly tell if the volunteers are doing something wrong or not. That's definitely true. But what makes you think that paper solves that problem? The people at the polling places trust that the volunteers aren't doing something wrong. They have to. They can't be asked to police the volunteers. Heck, they barely have enough time to vote.
In fact, in the last election, there were voting errors with those machines - a few polling places had ridiculously wrong vote counts, because 'something' happened to the original that was transferred. Of course, you should already guess that they don't have strong error checking on the devices that are storing the data, which is, um, silly, but it does remain that the votes were counted properly because a copy was still stored in the voting machines.
My main point is that paper doesn't provide the security - which is what several people here are suggesting. It's just as manipulatable as any other portion of it. What you have to have is an open process, and an open design.
Let me put it this way: there were comments in this thread that suggested that a closed, proprietary system that provides a paper receipt along with an electronic record (where presumably the electronic record is being used as the primary, and the paper as the backup) is good enough, because paper ballots are somehow a perfect failsafe.
And that I heftily don't agree with. If I were malicious, all I'd need to do is make sure that the error rate in each one of the machines is small, such that you'd barely notice it with statistical sampling of the check of the paper ballots.
That's what worried me. Paper isn't a failsafe here.
With electronic vote represntations, all of that must, perforce, happen invisibly.
What do you mean by "invisibly?" It happens invisibly with paper ballots, too. Optical scan ballots, especially. Aren't most paper ballots mechanically read, anyway?
Yes, those ballots could be checked. But an electronic system could be checked, too, and both of those procedures would take the same amount of time.
You're missing the point. I'm not saying paper trails are useless. They're not. Hell, the documentation and audit logs for an electronic system would be paper.
What I'm saying is that a paper receipt for vote tracking is pointless. It doesn't tell you anything. It says "Hi, you voted for this." The only way that you're going to know if that tally is wrong is if the final result says "Person You Voted For: Zero." Then you say "hey, I know I voted for him. It said I did. And yet, there were no votes tallied there must be something wrong!"
But other than that, how does it help? It doesn't. The "paper" portion isn't the important part. It's the auditability of the system.
That auditability might include an internal paper-like trail and spot verification checks. It might even include printed paper ballots. But the way you make the system foolproof is if you open it up to external prodding and you document what you're doing.
You should check to see what I replied to originally. That person was saying "all you need is a paper trail." No. That's not true. If you do that, then all you've done is reduced the difficulty of fixing the election to that of hacking a (non-auditable) system and rigging a physical backup. This isn't safe.
Electronic systems actually give us the ability to do things we weren't able to do with pure paper ballots : like make copies, or actually use cryptography easily. If you actually use them, you can make the system much more secure.
Part of the reason I'm so adamant about this is that if you switch to electronic voting, and the paper voting becomes a "backup", security measures will get weaker about that paper voting. Probably to the point where you don't even have a chain of custody in certain cases. And then someone will take advantage of it.
Eh. I'd skip the code entirely. You don't need software. It's a simple design. Build the thing on bare metal, and you can actually guarantee that there are no bugs or problems in the logic fundamentally.
Heck, if you build it with only, say, 74 series chips, that means that in order to actually hack into it you'd essentially need to make an entire new set of chips and replace the ones that are actually there. If you put things in a PLD, FPGA, or microcontroller, you could pop out the flash/PROM and replace it. Of course, you'd make it tamper proof, but that's still not a good guarantee of what code's actually running there - i.e. a malicious manufacturer. By just using basic logic chips you can spread the vendors around and minimize risk there.
The paper trail is fine. All it does is make people believe there's a reliable backup, but it doesn't actually do anything itself. It's less reliable, after all. Printers fail. Ink runs and smudges. And paper has a nasty habit of taking up far too much space to be easily reexamined later.
I can't see how a paper trail actually helps. Surely the accuracy on counting, say, 10,000 ballots might be 1 in 10,000 - one might get lost, or torn, or is unreadable. And if you just set up a conspiracy so that the electronic counting is off by the same error that you get in the paper ballots - just in all of them, so that a random sampling won't catch it - you'll still have a problem.
The point is that paper can be verifyed by anyone.
Can it? It didn't seem that way in Florida in 2000. And in addition, the most important voter scrutiny happens after there's a problem. In most cases it's just one or two people guarding it.
Relying on a few people to guard votes is just silly. We can build the security into the system. I heftily, heftily don't agree with those who say that "older ways are better." That's a silly argument. The government doesn't rely on passing notes to communicate secret information. Banks don't rely on transferring money anymore.
What people are worried about are whether or not the machines malfunction. This is silly. What you want to be worried about is whether or not it can be tampered with - and that sort of technology exists already. And it's better for electronic voting than it is for paper.
I just don't understand what people's fascination with paper is. If you have a paper trail, you'll have a populace which thinks the vote went okay. This isn't important. What's important is actually having a vote which went okay. Whether or not the people think it went fine is pointless.
No, the old ways are the best ways here, and they're adequate.
"Adequate"?
What's "adequate"? Only "minor" voter fraud? Once every few elections? It's not like paper ballots haven't been forged before. They have. It's happened.
The real world doesn't provide those sorts of certainties.
I don't agree. Banks don't rely on four guys carrying a locked box of money when they transfer money. Neither should we.
Changing 10.000 votes is a pretty big industrial undertaking.
Uh. No? I just do it beforehand on a printer. Takes probably half a day.
Not a big deal at all.
but I haven't heard of anything real that's worthy of the name, and i have serious doubts it's even possible, given the nature of computers.
The best thing that computers can do is replicate the data many times over. You can't do that with paper ballots.
It should be noted that several electronic voting system errors were discovered by the local copies. You wouldn't've been able to find that if it was malicious with paper ballots.
The only thing I can come up with is a paper trail
How, exactly, is a paper trail more secure than advanced tamper protection schemes?
You can't copy paper easily. You can forge paper easily. Paper can be destroyed - you can't exactly do the election over, as most places don't have a provision for this. Paper isn't reliable or error-free.
Paper elections have been forged before. Why should we believe that a forgeable system makes an insecure system secure?
The fact that using a printed balot as a paper trail is such an obvious solution
What makes you think that a paper trail is a good solution?
Do you think paper ballot elections have never been forged before?
Just because people think the system is working doesn't mean it is. In fact, paper trails might be infinitely worse because people think they're a safeguard when they're not.
Think about it. The paper trail has to be anonymous, so it can't contain any identifying features. The people can't take the receipt with them when they leave (extortion/vote selling reasons, besides, it would make it useless anyway) and so they've got to relinquish control over it and trust that it's not altered. Hell, the paper could be written in an ink that's designed to evaporate, and preprinted with an invisible ink that appears in several hours afterwards.
You don't want to reassure the public that their individual vote was counted. That's easy to fake. What you want to do is reassure the public that every single vote was counted correctly. Unless the system is open and auditable, the only way to do that would be to hold another vote and compare the results. Which you obviously can't do.
Why do people think paper is a good idea for voting? We barely use it for money anymore because it's less safe and secure - try buying a car with $50K in bills. Why is it okay for voting?
Yes, we've had paper ballots and paper trails before. But if you're going to change the voting system, why not change it to actually be good?
how can they be certain that a lone partisan coder wouldn't sneak a line of code within what I'm sure are millions of lines?
Don't have software. Implement the thing on bare metal and publish the schematics and PCBs and make the hardware available for inspection before and afterward. Sure, this is still forgable - replace the actual chips with custom-designed ones, or something like that - but you can still mitigate those risks dramatically. Much more so than you could do for paper, for instance.
Of course, we're talking about a country that doesn't even make voting a holiday. We don't exactly make the sanctity of democracy a high priority.
but have it print out a piece of paper with the voter's choices.
That's nowhere near enough.
Diebold shouldn't be worried about voters. They should be worried about volunteers who have access to the system. In that case, it's just as trivial for one of the volunteers to hack the system, and also print out fake paper trails as well.
The paper trails can't have any identification on it (vote selling/extortion reasons! votes have to be anonymous) and so they're pretty trivially forged. Sure, you could try to add, I dunno, some sort of anti-forge stuff to it, but now it's no longer just as simple as a paper trail.
The system has to be open and auditable. Paper isn't reliable. We shouldn't be switching away from paper because it's impractical: we should be switching away because it's not safe.
An electronic system can be made much safer than a paper one. Paper ballots can be lost: they can't easily be copied without logistics problems. They can be forged relatively easily - there's not exactly encryption.
You don't want a paper trail. You want an auditable system. Your instincts tell you that paper is auditable. I don't agree.
even to trust a simple hardwired system which, obviously, is still beyond the understanding of most of the population. Most people aren't EEs.
Some people can't read - they have to trust the people who can, don't they? Trust has to start somewhere. It's not important to 'reassure the public' with something that's not actually safe. The safer thing to do is reassure the public by explaining the process.
Most people might not be EEs, but they'll still believe that there's not a conspiracy between all EEs in the country to say it's safe. People will not, however, believe that there isn't a conspiracy inside one company who won't even release their source code. And they're smart not to believe that.
You might. I wouldn't. Paper and pencil isn't exactly "complicated" to fake, after all. You can fake it with... oddly enough, paper and pencil.
It's not like paper and pencil elections haven't been forged before.
The problem isn't "paper" versus "electronic". The problem is "open and auditable" versus "closed doors". In fact, in a lot of ways, paper is "closed doors" as well. How do I know that the paper ballots, when counted, are the same as when they were there? Sure, you can say that you've got a chain of possession, or a 'lock that's impossible to break' but that doesn't inspire confidence in me, especially if said process hasn't been independently tested by organizations who were trying to fraud the system, including having members of that testing organization in the chain of possession (inside jobs).
Now, with electronic, you can guarantee that: hashes of the data, multiple copies, etc. You still want an independent organization to check it out, sure, but it can be safer.
In fact, for those worried about 'software bugs', I agree. It's hard to guarantee software safety. That's why I wouldn't have software in it. It's not exactly complicated. Just implement it all (well, at least, especially the vote storage/checking stuff) on bare metal.
The Diebold system is crap because it's crap. It's just about the quickest way you could imagine building a voting setup. It's unsurprising that it's awful. You want a better way? Design it from scratch.
Paper and pencil being safe? C'mon. Banks don't transfer physical money anymore because doing it electronically is better. Voting should advance, too. Just sanely, and auditably.
Closed-cell calorimetry has been done - a ton by the people at Stanford Research Institute. That criticism was taken up pretty quickly.:) The really intriguing results were when a calorimeter was modified to allow gas spectrometry to look for helium. Granted, that sort of thing is really really hard, but they did find helium, though there were criticisms that pointed out that it could be from lab room contamination. Still, this is one of those things where you're starting to grasp at straws.
The DOE review is pretty good: here. Note that a lot of LENR-CANR is, shall we say, just a tad bit biased and crank-worthy, but they've got a lot of documentation, so they're at least useful for that.
The DOE anonymous reviewers are particularly noteworthy: it's amazing how bad some of the reviews are, which, to be honest, is fairly insulting. The DOE paid these guys. They had a responsibility to be critical and honest, and several simply weren't (specifically 1, 2, and 15 were very bad). Conversely, however, Review 4 is excellent, as was Review 7, and Review 10 and 11, and 16 (really excellent). On the whole most were pretty good, but 1 and 2 leave a definite stink. It's interesting to read, however, because it's a completely independent and unbiased set of opinions. At least one reviewer was completely convinced.
Review 12 has a great summation of the field: "The quality of work is so inconsistent in this field, including the work of some key players, which makes it difficult to clear the black cloud and to increase the credibility of the field. Repeated retractions and conflicting experimental results in the past certainly did not help. Hopefully as time on, a few careful studies will provide a definitive conclusion. Unfortunately, that has not happened yet, although some progress has been made."
However, the review does make for interesting reading. Several of the reviewers do point out that there are possibilities for an actual theoretical basis for it, in particular results from German beam tests that show that anomalous screening can occur in deuterated metal films.
My bad! We never use anything other than liquid in homes, like oh, say, freon (or any of the other cooling liquids).
Or maybe you're more worried about the fact that they'd be at high temperatures, and could maybe burst pipes? Because we never use explosive materials inside homes - like, say, natural gas, home heating oil, gasoline or kerosene?
at which point the Russians wound up sticking America with the bill for most of the Russian contribution
Do keep in mind that the US was supposed to ferry crew back and forth - when the Shuttle was grounded after Columbia, we started using the Soyuz capsules. Russia then started to say "uh, hey, we need to get paid for this sort of stuff..." and Congress starting hedging about whether they could give Russia money at all due to political issues.
It would be one heck of a coincidence if the same people who made this large number of experimental mistakes now happened to produce a valid result.
I think you'd be surprised. The number of crappy experiments which end up actually working (even if sometimes they're almost totally forged) is quite frightening. I can think of a fair handful in my own field...
Most of the criticisms you've addressed have been what several other researchers have been shoring up, and actually that's part of the problem with the DOE response. Certain techniques (calorimetry, for one thing) are much harder to do rigorously than they first appear, and so when the DOE guys say "you must've done the calorimetry wrong" the researchers were very angry, because the reviewers clearly failed to understand how difficult calorimetry is, and how careful the researchers were.
I do have to agree with the researchers on that one: calorimetry is an underappreciated discipline.
then it's a good indication that the standard models are in serious need of revision.
It could be some sort of wacko deuterium-moderated palladium fission. That wouldn't break models other than nuclear physics, and it's not like they're tremendously rigorous anyway.
Part of the problem is that Pons and Fleischmann sortof jumped on the "it's fusion!!" bandwagon to the press! when it was clear they had no idea what was going on. They even admitted it in the published paper. They simply said there was a process going on via "unknown nuclear channels", which, if their results (and the results of several others) are correct, they're right about. But calling it fusion was just retarded. That's what made them a laughingstock.
Slashdot Mirror
The first source you're quoting has nothing to do with whether or not humans cause a dominant CO2 rise in the atmosphere - it's just very long term trends: "what's going to happen to the Earth." It basically says "yah, the Earth's screwed in about half a billion years." This is ballpark what we knew already - planetary habitable zones migrate outward, and we're on the inner edge.
The second source stems all of its criticism of whether or not humans are causing the CO2 rise on a poor criticism - saying "ocean warming causes CO2 increase, so how do we know the CO2 increase is causing ocean warming, and not the reverse?" The answer to that is simple: we have a coherent model for warming due to CO2 emission. We do not have a coherent model for warming of the oceans causing CO2 emission. Occam's Razor chooses the first: the second requires an additional mechanism for generating ocean warming.
Actually, the second explanation ("ocean warming is caused by some unknown mechanism, which leads to a CO2 increase in the atmosphere") requires even more work than that: it also requires machinery to link the CO2 increase caused by ocean warming to that emitted by humans. The airborne fraction: that is, the fraction of CO2 emitted by humans that we see as an increase in atmospheric CO2 - has stayed pretty constant over 5 year averages.
Oh, and I saved my most damning point for last: you see, you can determine where the CO2 is coming from by looking at the isotopic composition of the CO2 in the atmosphere, and see if it looks like the isotopes from fossil fuels, or other sources (like the ocean). They're from fossil fuels.
The carbon dioxide rise is anthropogenic. There is no scientific debate about this fact anymore.
I don't see how, unless the systematic errors are of such low magnitude that they're negligible in practice anyway.
Depends on the level of random errors and the statistical sampling that's used. Besides, "negligible in practice"? I'd prefer to leave no tamper margin, not just negligible tamper margin.
If you assume more realistic fraudsters, who weigh the chance of getting caught against the probability that they'll affect a race, what you'll really have is no fraud (of the sort we're discussing here).
If I was a very malicious person, what you'd do is allow the electronic systems to gain credentials, and then slowly reduce the quality checks in the crosscheck to allow the random error rate to creep upwards. So long as you stay under the crosscheck error rate, there's zero risk. How can something be "suspicious" when it's identical to a nonfradulent election?
That is, imagine the electronic voting system compares to the hand count to 1 in 1000 over 10 years of elections. The malicious person simply makes it so that the electronic counting outputs a result that is different from what it actually counted by 1 in 10,000. It'd take a very, very long time to notice that tampering. Then, of course, when things aren't noticed after two or three years, you work to reduce the quality of the crosscheck, allowing your tamper margin to go up.
This is assuming a simple paper receipt that can't be tracked to actual electronic votes: that is, the "paper counting" process is separate from the "electronic counting" process. This is the bad portion I'm talking about.
Registration fraud offers a much better return for the investment.
Sure. That doesn't mean I'm not going to worry about direct vote tampering. Someone else can worry about improving registration procedures.
And how do you ensure that the voter can't be associated with the ID? The ID would have to be invisible to the voter, but available to the vote counter.
How can you ensure the voter can't be associated with a ballot? They still have fingerprints on the ballot, for instance.
The ID would just be a random number - just to associate an 'electronic vote' with a 'paper vote' to link the two counting systems. There'd be no information in the system about who the voter was which placed each 'electronic vote'. While I guess you could, I dunno, do a forensic analysis of the system or something and figure it out somehow, you could conceivably do that with a paper receipt as well. In fact, the paper receipt would be a heckuva lot easier to track.
The real concern would be trying to keep the ID invisible to voters, so that they can't prove their vote (although to me, this is less of a concern) - and that you could do by having the paper be special - revealing the ID with a UV lamp, for instance.
That would mean no nice touch screen, no easy-to-use step-by-step interface, etc
Er? No it wouldn't. I could implement those things on bare metal in less than a month. This is a very simple system. It's essentially a state machine - which means just a bunch of flip flops. The only reason it's being implemented in embedded systems right now is because Diebold is extremely lazy and cheap.
If we're going to talk about that, we should begin by discarding plurality voting and use a system that doesn't disenfranchise anyone who isn't a fan of the two most popular parties.
There is such a thing as a fair count. There is no such thing as a fair voting system. Arrow's Paradox.
The problem is that omnipotent does indeed mean able to do anything.
Common usage of omnipotent is "able to do anything."
That doesn't mean that's a usable definition. Common usage of the word "force", or even "work" doesn't lend itself to a useful physics definition. Instead those needed to be clarified significantly. Work is the clearest example.
"Able to do anything" isn't a workable definition to use in logic. It just doesn't work - it leads to things akin to Russell's paradox.
because I know that I can do anything that is possible for me to do.
I didn't say "possible for said being to do." I said possible. I can't continue living in vacuum, but it's certainly possible to do so.
It's worth pointing out, BTW, that it is far better to accept random errors than to enable deliberate errors. On average, random errors won't change the results.
My point is that random errors in your crosscheck can mask systematic errors in your primary count.
What we should have, of course, is a system in place for redoing an election in the case of too close a result. It's simply crazy to declare one person a winner if the results are within the margin of error of the counting system.
But barring that, it's dangerous to use a system with random error to check a system that can have intrinsically lower error. That allows a "tamper margin" below the random error rate in the crosscheck. That is pretty much my entire point. Note that you may think that it's silly to worry about a tampering method that might result in a negligibly small number of votes swung towards a given party, but multiplied over all of the races, and over many years, it'd be worth it.
Do we have evidence that paper ballot counting is zero error? It's simple enough to say that it is, but that's a presumption that you're making. And it's a very, very important presumption.
It's easy enough to test. Do multiple recounts. Are the results exactly the same? I didn't think they were in the Florida 2000 recounts - they were close, but they weren't identical. And that's unsurprising. People are going to make mistakes.
For each ballot, run it through the machine and get the machine result.
Right, this is pretty much what I thought you were saying before. I agree, this pretty much solves sub-threshold tampering - except that you're requiring a machine-readable paper ballot, and I don't think you need that. Just a link between the printed receipt and the count on the system.
Let me explain a bit in case you're confused - I think we were imagining different systems. What I was thinking of is a system like the Diebold system currently, but improved with a paper receipt, where the votes are totalled, and a paper receipt is printed. The votes are entered into some sort of database, or a separate count, and are nothing more than a *count*: that is "654 votes for A, 728 votes for B". The paper receipts are kept separate, and counted via random sampling to determine if that vote ("654 ballots for A, 728 ballots for B") matches the electronic vote.
I don't like that sort of system. It's crosschecking a system with very low intrinsic error with something which may have a higher error, which means you have a tamper margin.
That being said, though, I think you could fix that system by melding it with the benefits of mechanical reading.
What you do is tag each paper ballot with an ID, and instead of storing just a count, store the ID that you've printed electronically. This way you can determine any below-threshold tampering, because mangled and lost ballots are going to be lost separate from the system. This way you can't 'hide' the tampering, because you don't know what ballot is going to be mangled or lost, and you can compare each paper ballot with what the system thought the ballot was for.
So now you've got "654 ballots for A, 728 ballots for B", and an electronic record which has a list of 654 IDs for A, and 728 IDs for B. You can't "sneak" an extra vote in there without forcing people to assume that a paper receipt got lost, which, as you noted, is preventable.
I hadn't thought of that before. That's an excellent, excellent idea. I don't think it hurts the anonymity of the process at all, because it's just an arbitrary ID. It seriously helps the credibility and that would allow a closed electronic system.
Any such machine is a closed system, because even if it runs open source there's no practical way to ensure that the code installed on the machine
No software. Why do you need software for this sort of thing? You're counting. Implement the thing on bare metal. Software just leaves yourself open for th
This study does at least show that, if whatever pertinent deity exists, it cares more about its ego than the needs of people who may die as the result of an illness.
Ah yes, once again, the ugly specter of poor wording and the Problem of Evil rears its head. (*)
You've got an implicit assumption going from this study to your conclusion: you're assuming that it was possible that these people would survive.
In fact, this is pretty common with prayer: we assume, implicitly, that the impossible is possible, and that said impossibility has no consequences other than the immediate action. That is, it's possible for said sick person to just magically become okay, with no drawbacks. This is pretty silly. The person praying doesn't know what's wrong with the sick person. They're just assuming that it would be better if they survived than if they died. That's an unfounded assumption.
Note that this isn't presuming that prayer does nothing. It just can't change the inevitable. It's only our (false) hope that allows us to believe that certain things aren't inevitable.
If I were God, I'd be incredibly insulted by your statement. You're presuming a lot of knowledge about the Universe that you simply don't have. How, precisely, do you know what the needs of the person who was ill was? And, presupposing that the person's death was inevitable, how do you know that that person's death wasn't made a ton easier by said deity?
(*: The relation to the Problem of Evil is pretty straightforward. You're presupposing that "omnipotent" strictly means "able to do anything". This, unsurprisingly, causes problems with your logic, because "able to do anything" isn't a well-formed set. If you instead say that God is omnipotent, meaning "able to do anything that is possible" that clears the situation, and the resolution that I mentioned above - that there was nothing that could be done with regard to the ill people - is pretty clear.)
Nope. Not sure where you got that idea. I don't want votes to be traceable to any granularity smaller than the voting precinct. It'd be nice if they could be more anonymous than that.
Then I'm really confused how you'd detect a tampered ballot. In the end, out of a voting precinct, all you would get are two numbers: the amount that the electronic system records, and the amount that the paper recount has.
Given only that, I can't see how the paper ballots help you, since all you can do is compare those two numbers. The paper recount isn't perfect - hell, the paper can be lost - and so if a random paper recount is compared to an electronic recount, if the error rate is like "1" or "2" votes, that probably wouldn't raise a red flag.
So the election's still manipulatable with the closed system so long as the attacker keeps the manipulation level below that of the paper recount accuracy. If the paper recount accuracy is low enough, this could be a problem.
I guess that's kindof the crux of the difference here: what is the margin of error in paper recounts? Isn't it a little unsafe to just assume it's zero?
Unless you require that the margin of error in the paper recount be extremely low. I don't know how practical that is. It certainly doesn't seem to be practical in a lot of counties in the US.
And in any case, we don't really have the provisions to say "this vote was, for all practical purposes, a tie", which means that any level of tampering has to be a worry as well. It'd be easy enough for a manufacturer of a counting process to give a bonus to, say, all Republicans which would give, in a statewide election, only a few hundred votes margin their way. Negligible for most elections, but over enough time, enough to sway a few closer ones.
I should point out that the only thing I really disagree with is the necessity for an open counting system. I think so long as hand recounts have any inaccuracy (and for any physical medium, it will) you've got to worry about very subtle vote influencing. Which is why I think any "closed" system in the loop is a bad thing.
So, how many sample ballots need to be examined to catch you at this?
Are you suggesting a system where each ballot is directly traceable to an electronic entry - like, with some sort of ID? Is that actually suggested?
Ballots should be both human and machine-readable.
Wait. If they are machine readable, that means that the counting will be done invisibly anyway.
The fact that people can read the ballots is therefore pretty much academic, isn't it? In actuality, they won't.
In some sense it would be like forgery protection with cash. In theory it's nearly impossible to forge US money, because there are about a hundred safeguards in place. In practice it's easy to slip small amounts of fake money through, because no one actually uses the safeguards.
Although I guess if the ballots were tagged with identification marks and the final record basically consisted of grouping the identification marks rather than just printing out the total, then statistical sampling can save you there.
That sort of auditability is impossible with election systems because of the requirement that votes be anonymous. Since the ultimate originator of the vote *MUST* have no way to verify that his or her individual vote was properly traced throughout the system, we can't apply the same auditing techniques.
Let me just stress this again: you're definitely right that you can't audit the end-to-end system. But you can (and must) audit everything else.
Heck, with paper ballots that wasn't done in Florida in 2000, and that caused a large portion of the confusion - that is, the paper ballots weren't up to standards in terms of the presentation of information, and to add to confusion, they were misprinted slightly as well. Now, that was almost certainly from incompetence - but it could've just as easily been done maliciously as well.
So clearly, the design of the paper ballot has to be audited as well.
However, there's an interesting problem here: we worry about things like electronic systems that are maliciously programmed or manipulated. Why couldn't we have paper ballots that are maliciously designed as well? Even psychologically maliciously designed, such that the design of the ballot leads the casual voter to choose one selection over the other.
Granted, those problems are common between the two.
The main worry I have is that some people believe that a paper ballot solves everything. It doesn't. The votes must be anonymous, which kills any ability to audit the system end-to-end. But that doesn't mean that every other step can't be audited, and it must be.
But that's better than making it possible for one person to modify *all* the votes. Much, much better.
I think you're misunderstanding the kind of system I'd be thinking of. Are you thinking of a system where the votes are electronically transferred - that is, over a network?
I agree that would be really, really stupid.
But I'd be more thinking of a "paper replacement" electronic system - like the ones that are in use in several places (including where I am) where the votes are stored physically on some medium (in several copies) and then transferred to a central facility again. It's exactly the same risk-sharing as you have with paper - just that now you've removed (or reduced) the risk of the volunteers at the voting place.
Granted, the people actually voting can't clearly tell if the volunteers are doing something wrong or not. That's definitely true. But what makes you think that paper solves that problem? The people at the polling places trust that the volunteers aren't doing something wrong. They have to. They can't be asked to police the volunteers. Heck, they barely have enough time to vote.
In fact, in the last election, there were voting errors with those machines - a few polling places had ridiculously wrong vote counts, because 'something' happened to the original that was transferred. Of course, you should already guess that they don't have strong error checking on the devices that are storing the data, which is, um, silly, but it does remain that the votes were counted properly because a copy was still stored in the voting machines.
My main point is that paper doesn't provide the security - which is what several people here are suggesting. It's just as manipulatable as any other portion of it. What you have to have is an open process, and an open design.
Let me put it this way: there were comments in this thread that suggested that a closed, proprietary system that provides a paper receipt along with an electronic record (where presumably the electronic record is being used as the primary, and the paper as the backup) is good enough, because paper ballots are somehow a perfect failsafe.
And that I heftily don't agree with. If I were malicious, all I'd need to do is make sure that the error rate in each one of the machines is small, such that you'd barely notice it with statistical sampling of the check of the paper ballots.
That's what worried me. Paper isn't a failsafe here.
With electronic vote represntations, all of that must, perforce, happen invisibly.
What do you mean by "invisibly?" It happens invisibly with paper ballots, too. Optical scan ballots, especially. Aren't most paper ballots mechanically read, anyway?
Yes, those ballots could be checked. But an electronic system could be checked, too, and both of those procedures would take the same amount of time.
You have to produce documentation.
You're missing the point. I'm not saying paper trails are useless. They're not. Hell, the documentation and audit logs for an electronic system would be paper.
What I'm saying is that a paper receipt for vote tracking is pointless. It doesn't tell you anything. It says "Hi, you voted for this." The only way that you're going to know if that tally is wrong is if the final result says "Person You Voted For: Zero." Then you say "hey, I know I voted for him. It said I did. And yet, there were no votes tallied there must be something wrong!"
But other than that, how does it help? It doesn't. The "paper" portion isn't the important part. It's the auditability of the system.
That auditability might include an internal paper-like trail and spot verification checks. It might even include printed paper ballots. But the way you make the system foolproof is if you open it up to external prodding and you document what you're doing.
You should check to see what I replied to originally. That person was saying "all you need is a paper trail." No. That's not true. If you do that, then all you've done is reduced the difficulty of fixing the election to that of hacking a (non-auditable) system and rigging a physical backup. This isn't safe.
Electronic systems actually give us the ability to do things we weren't able to do with pure paper ballots : like make copies, or actually use cryptography easily. If you actually use them, you can make the system much more secure.
Part of the reason I'm so adamant about this is that if you switch to electronic voting, and the paper voting becomes a "backup", security measures will get weaker about that paper voting. Probably to the point where you don't even have a chain of custody in certain cases. And then someone will take advantage of it.
Without paper, you have absolutely no proof that what the electronic system says was voted is what actually was voted.
Why does paper provide that proof?
It doesn't. It just says "yah, here's a piece of paper. This is what I recorded. Really. Swear. Trust me."
The proof is in the fact that the design is open and has been audited. That's the only proof you can have.
Eh. I'd skip the code entirely. You don't need software. It's a simple design. Build the thing on bare metal, and you can actually guarantee that there are no bugs or problems in the logic fundamentally.
Heck, if you build it with only, say, 74 series chips, that means that in order to actually hack into it you'd essentially need to make an entire new set of chips and replace the ones that are actually there. If you put things in a PLD, FPGA, or microcontroller, you could pop out the flash/PROM and replace it. Of course, you'd make it tamper proof, but that's still not a good guarantee of what code's actually running there - i.e. a malicious manufacturer. By just using basic logic chips you can spread the vendors around and minimize risk there.
The paper trail is fine. All it does is make people believe there's a reliable backup, but it doesn't actually do anything itself. It's less reliable, after all. Printers fail. Ink runs and smudges. And paper has a nasty habit of taking up far too much space to be easily reexamined later.
I can't see how a paper trail actually helps. Surely the accuracy on counting, say, 10,000 ballots might be 1 in 10,000 - one might get lost, or torn, or is unreadable. And if you just set up a conspiracy so that the electronic counting is off by the same error that you get in the paper ballots - just in all of them, so that a random sampling won't catch it - you'll still have a problem.
The point is that paper can be verifyed by anyone.
Can it? It didn't seem that way in Florida in 2000. And in addition, the most important voter scrutiny happens after there's a problem. In most cases it's just one or two people guarding it.
Relying on a few people to guard votes is just silly. We can build the security into the system. I heftily, heftily don't agree with those who say that "older ways are better." That's a silly argument. The government doesn't rely on passing notes to communicate secret information. Banks don't rely on transferring money anymore.
What people are worried about are whether or not the machines malfunction. This is silly. What you want to be worried about is whether or not it can be tampered with - and that sort of technology exists already. And it's better for electronic voting than it is for paper.
I just don't understand what people's fascination with paper is. If you have a paper trail, you'll have a populace which thinks the vote went okay. This isn't important. What's important is actually having a vote which went okay. Whether or not the people think it went fine is pointless.
No, the old ways are the best ways here, and they're adequate.
"Adequate"?
What's "adequate"? Only "minor" voter fraud? Once every few elections? It's not like paper ballots haven't been forged before. They have. It's happened.
The real world doesn't provide those sorts of certainties.
I don't agree. Banks don't rely on four guys carrying a locked box of money when they transfer money. Neither should we.
Changing 10.000 votes is a pretty big industrial undertaking.
Uh. No? I just do it beforehand on a printer. Takes probably half a day.
Not a big deal at all.
but I haven't heard of anything real that's worthy of the name, and i have serious doubts it's even possible, given the nature of computers.
The best thing that computers can do is replicate the data many times over. You can't do that with paper ballots.
It should be noted that several electronic voting system errors were discovered by the local copies. You wouldn't've been able to find that if it was malicious with paper ballots.
The only thing I can come up with is a paper trail
How, exactly, is a paper trail more secure than advanced tamper protection schemes?
You can't copy paper easily.
You can forge paper easily.
Paper can be destroyed - you can't exactly do the election over, as most places don't have a provision for this.
Paper isn't reliable or error-free.
Paper elections have been forged before. Why should we believe that a forgeable system makes an insecure system secure?
The fact that using a printed balot as a paper trail is such an obvious solution
What makes you think that a paper trail is a good solution?
Do you think paper ballot elections have never been forged before?
Just because people think the system is working doesn't mean it is. In fact, paper trails might be infinitely worse because people think they're a safeguard when they're not.
Think about it. The paper trail has to be anonymous, so it can't contain any identifying features. The people can't take the receipt with them when they leave (extortion/vote selling reasons, besides, it would make it useless anyway) and so they've got to relinquish control over it and trust that it's not altered. Hell, the paper could be written in an ink that's designed to evaporate, and preprinted with an invisible ink that appears in several hours afterwards.
You don't want to reassure the public that their individual vote was counted. That's easy to fake. What you want to do is reassure the public that every single vote was counted correctly. Unless the system is open and auditable, the only way to do that would be to hold another vote and compare the results. Which you obviously can't do.
Why do people think paper is a good idea for voting? We barely use it for money anymore because it's less safe and secure - try buying a car with $50K in bills. Why is it okay for voting?
Yes, we've had paper ballots and paper trails before. But if you're going to change the voting system, why not change it to actually be good?
how can they be certain that a lone partisan coder wouldn't sneak a line of code within what I'm sure are millions of lines?
Don't have software. Implement the thing on bare metal and publish the schematics and PCBs and make the hardware available for inspection before and afterward. Sure, this is still forgable - replace the actual chips with custom-designed ones, or something like that - but you can still mitigate those risks dramatically. Much more so than you could do for paper, for instance.
Of course, we're talking about a country that doesn't even make voting a holiday. We don't exactly make the sanctity of democracy a high priority.
but have it print out a piece of paper with the voter's choices.
That's nowhere near enough.
Diebold shouldn't be worried about voters. They should be worried about volunteers who have access to the system. In that case, it's just as trivial for one of the volunteers to hack the system, and also print out fake paper trails as well.
The paper trails can't have any identification on it (vote selling/extortion reasons! votes have to be anonymous) and so they're pretty trivially forged. Sure, you could try to add, I dunno, some sort of anti-forge stuff to it, but now it's no longer just as simple as a paper trail.
The system has to be open and auditable. Paper isn't reliable. We shouldn't be switching away from paper because it's impractical: we should be switching away because it's not safe.
An electronic system can be made much safer than a paper one. Paper ballots can be lost: they can't easily be copied without logistics problems. They can be forged relatively easily - there's not exactly encryption.
You don't want a paper trail. You want an auditable system. Your instincts tell you that paper is auditable. I don't agree.
even to trust a simple hardwired system which, obviously, is still beyond the understanding of most of the population. Most people aren't EEs.
Some people can't read - they have to trust the people who can, don't they? Trust has to start somewhere. It's not important to 'reassure the public' with something that's not actually safe. The safer thing to do is reassure the public by explaining the process.
Most people might not be EEs, but they'll still believe that there's not a conspiracy between all EEs in the country to say it's safe. People will not, however, believe that there isn't a conspiracy inside one company who won't even release their source code. And they're smart not to believe that.
You might. I wouldn't. Paper and pencil isn't exactly "complicated" to fake, after all. You can fake it with... oddly enough, paper and pencil.
It's not like paper and pencil elections haven't been forged before.
The problem isn't "paper" versus "electronic". The problem is "open and auditable" versus "closed doors". In fact, in a lot of ways, paper is "closed doors" as well. How do I know that the paper ballots, when counted, are the same as when they were there? Sure, you can say that you've got a chain of possession, or a 'lock that's impossible to break' but that doesn't inspire confidence in me, especially if said process hasn't been independently tested by organizations who were trying to fraud the system, including having members of that testing organization in the chain of possession (inside jobs).
Now, with electronic, you can guarantee that: hashes of the data, multiple copies, etc. You still want an independent organization to check it out, sure, but it can be safer.
In fact, for those worried about 'software bugs', I agree. It's hard to guarantee software safety. That's why I wouldn't have software in it. It's not exactly complicated. Just implement it all (well, at least, especially the vote storage/checking stuff) on bare metal.
The Diebold system is crap because it's crap. It's just about the quickest way you could imagine building a voting setup. It's unsurprising that it's awful. You want a better way? Design it from scratch.
Paper and pencil being safe? C'mon. Banks don't transfer physical money anymore because doing it electronically is better. Voting should advance, too. Just sanely, and auditably.
Closed-cell calorimetry has been done - a ton by the people at Stanford Research Institute. That criticism was taken up pretty quickly. :) The really intriguing results were when a calorimeter was modified to allow gas spectrometry to look for helium. Granted, that sort of thing is really really hard, but they did find helium, though there were criticisms that pointed out that it could be from lab room contamination. Still, this is one of those things where you're starting to grasp at straws.
The DOE review is pretty good: here. Note that a lot of LENR-CANR is, shall we say, just a tad bit biased and crank-worthy, but they've got a lot of documentation, so they're at least useful for that.
The DOE anonymous reviewers are particularly noteworthy: it's amazing how bad some of the reviews are, which, to be honest, is fairly insulting. The DOE paid these guys. They had a responsibility to be critical and honest, and several simply weren't (specifically 1, 2, and 15 were very bad). Conversely, however, Review 4 is excellent, as was Review 7, and Review 10 and 11, and 16 (really excellent). On the whole most were pretty good, but 1 and 2 leave a definite stink. It's interesting to read, however, because it's a completely independent and unbiased set of opinions. At least one reviewer was completely convinced.
Review 12 has a great summation of the field: "The quality of work is so inconsistent in this field, including the work of some key players, which
makes it difficult to clear the black cloud and to increase the credibility of the field. Repeated retractions and conflicting experimental results in the past certainly did not help. Hopefully as time on, a few careful studies will provide a definitive conclusion. Unfortunately, that has not happened yet, although some
progress has been made."
However, the review does make for interesting reading. Several of the reviewers do point out that there are possibilities for an actual theoretical basis for it, in particular results from German beam tests that show that anomalous screening can occur in deuterated metal films.
But for *home* use?
My bad! We never use anything other than liquid in homes, like oh, say, freon (or any of the other cooling liquids).
Or maybe you're more worried about the fact that they'd be at high temperatures, and could maybe burst pipes? Because we never use explosive materials inside homes - like, say, natural gas, home heating oil, gasoline or kerosene?
at which point the Russians wound up sticking America with the bill for most of the Russian contribution
Do keep in mind that the US was supposed to ferry crew back and forth - when the Shuttle was grounded after Columbia, we started using the Soyuz capsules. Russia then started to say "uh, hey, we need to get paid for this sort of stuff..." and Congress starting hedging about whether they could give Russia money at all due to political issues.
We're not exactly blameless in this.
It would be one heck of a coincidence if the same people who made this large number of experimental mistakes now happened to produce a valid result.
I think you'd be surprised. The number of crappy experiments which end up actually working (even if sometimes they're almost totally forged) is quite frightening. I can think of a fair handful in my own field...
Most of the criticisms you've addressed have been what several other researchers have been shoring up, and actually that's part of the problem with the DOE response. Certain techniques (calorimetry, for one thing) are much harder to do rigorously than they first appear, and so when the DOE guys say "you must've done the calorimetry wrong" the researchers were very angry, because the reviewers clearly failed to understand how difficult calorimetry is, and how careful the researchers were.
I do have to agree with the researchers on that one: calorimetry is an underappreciated discipline.
then it's a good indication that the standard models are in serious need of revision.
It could be some sort of wacko deuterium-moderated palladium fission. That wouldn't break models other than nuclear physics, and it's not like they're tremendously rigorous anyway.
Part of the problem is that Pons and Fleischmann sortof jumped on the "it's fusion!!" bandwagon to the press! when it was clear they had no idea what was going on. They even admitted it in the published paper. They simply said there was a process going on via "unknown nuclear channels", which, if their results (and the results of several others) are correct, they're right about. But calling it fusion was just retarded. That's what made them a laughingstock.
The whole contraption operates at atmospheric pressure, so what you get is at best steam at 100 deg. C or 370K.
Why, precisely, would you have to use water?
Last time I checked, there are plenty of other liquids (with higher boiling points) which would be perfectly suitable.