Well, it seems to me that what we're seeing is that it's easier to make and sell new smartphones -- and ways for them to consume bandwidth -- than it is to build out new mobile data infrastructure to grow the available bandwidth.
That is, I see this choice: Would you rather
1) that many people have smartphones but be forced to share the limited resource of bandwidth (meaning that each person gets only a certain amount); or
2) that a few people have smartphones and never encounter bandwidth constraints?
The wikipedia article for excalibur rounds says: "The weapon can make first round strikes on targets up to 20 kilometres (12 mi) away." http://en.wikipedia.org/wiki/M982_Excalibur
And those attacks were also domestic flights, not via US citizens but you don't check citizenship on domestic flights anyway
Controversial: Should we? I'd rather show my passport than go through a scanner or get a pat-down. Certainly I bet this 95-year-old woman would have preferred that.
and there have been plenty of home grown American nut jobs too.
...who have plenty of ways to carry out ill intent that have nothing to do with air travel. The TSA simply doesn't make us safer from them, even if you believe it makes air travel safer. In our non-air-travel lives, we seem pretty content with the trade off of nut job risk vs. law enforcement limitations.
From no one. They are scanning and patting down you to protect everyone else from you.
Well, OK, but I'm a *US citizen* traveling *domestically*. The law enforcement scrutiny I face at the airport is drastically disproportionate to what I face everywhere else in the country. Yet, a US citizen who intends harm to other US citizens isn't constrained to attack air travel. So I'm still scratching my head, here.
Companies should of course be free to offer compensation incentives on terms that fit their business needs. It's best for everyone if the incentives are expressed in plain language up front, so nobody feels tricked or taken advantage of later.
I also don't understand how the count can ever be verified without compromising the anonymity of the vote. If you don't trust the system you cannot trust any kind of verification it would do nor any kind of output it would produce (including any paper trail). Does anyone have any insight on the subject ?
The paper trail for a vote should be human-readable, inspected by the voter for correctness before deposit into a secure container at a polling place, and have no content that identifies the voter.
Physical pieces of paper can be physically watched (even shuffled) by multiple parties who are unlikely to collude (e.g. opposing political groups plus the press).
The 'e' part is just to speed counting. The paper makes it possible to handle claims of fraud, bugs, etc.
aside: Why does everyone keep misspelling 'rogue' in this thread? I've seen the 'rouge' reference dozens of times. That's a make-up product.
Anyway,
It sounds like the guy's able to run his own small server to meet innocuous intradepartmental needs in a basic sense. Again, in all of my posts here, I've said that the request is reasonable and hospital IT *should* be on board. The security issues are their problem; if his server breaks in other ways they can tell him it's his problem to get his service working for his department again.
It's obvious to you and me that IT ought to be on board. It clearly wasn't obvious to him. He came to us to ask about it. Why don't we just say, "yes, it's reasonable, and here are some reasons why"? Why engage in all of these characterizations of an accomplished professional who already is in touch with his IT department and is asking for advice?
It's true that Google can only divulge information known to it. However, if I were reading a hospital IT policy, I'd be in favor of one that actively discouraged sharing hospital data of any sort with Google.
It's just a calendar with shift times, as long as everyone involved doesn't mind it being on GCal, it shouldn't be a problem, but a random piece of hardware being connected to a hospital network IS a big deal. No one who's said anything is concerned about the safety of the shift times data, they ARE concerned about the HOSPITAL data that is on the same network & can be compromised by one security oversight on that server.
I agree that hardware on the hospital network is significant. In all of my comments on this thread, I've emphasized the importance of complying with IT policies, for exactly that reason.
I remain reluctant to agree that Google is a good alternate solution. I'm not trying to bash, nor to be antagonistic. I'm simply not convinced that it makes sense to use Google for hospital operations. From the OP, even the hospital's IT guys seem to be OK with the idea of hosting that service on the hospital network.
Aside from that issue: In fact I know (boy I hope it's not sacrilege to say this here) people who are gcal users, who REALLY don't like it. If this solution makes the users happy and is OK with hospital IT -- where's the beef?
It's a fair point, that Google doesn't have access to patient records. OTOH, hospital staff are probably better off relying on their internal computing resources than on Google, for their operations. Indeed, one way to *guarantee* that Google doesn't learn anything it shouldn't, is to have a policy of NOT using it for hospital operations. Furthermore, since OP is in touch with internal IT, he can make it *their* problem to ensure compliance with policies and legal requirements, either (a) by telling him not to run his own server or (b) asking that he meet certain conditions in order to do so. That also gives him an internal resource to contact if anything breaks, rather than relying on Google support.
When you say "unsupervised machine" -- in fact it's not unsupervised. He told IT about it and they (rightly IMO, and I said as much in my earlier post) requested access to it.
First, there's no reason apparent from the original post to conclude that a law has been violated. Second, your contempt is unwarranted. Third, your baseball bat proposal doesn't solve the original problem of how to migrate staff on-call schedules from paper to electronic form.
He's a doctor, a faculty member (professor), and a division head (administration/management). I promise you he's not a moron. There's a substantial amount of career achievement and hard work implicit in those credentials. Furthermore, he's actually made time to understand how to set up his own Linux server, and he's come here asking for advice. Even if you think the course he's pursuing isn't the right one, need you take this tone?
I'm surprised that you'd trust Google apps more than an internal server known to hospital IT.
The data at issue are on-call schedules for staff, not patient data.
To the original poster: I think the request is reasonable. They probably want to do simple things like verify that the server is behaving as expected, monitor what processes are running, gather usage statistics, etc., etc. Do you have a specific worry as to what might transpire as a result of giving them access?
Adhering to IT's policies *probably* also means making them the ones on the hook for legal issues.
I honestly don't understand why there is such bias against electronic voting on Slashdot since, in theory, it's a "nerd community".
Yes, e-voting, after a lot of effort can be compromised. Regular paper-ballot voting can be compromised by anyone, skilled or not, with not a lot of effort at all. Any voting system can be compromised. I don't honestly understand why the Slashdot community dislike e-voting that much.
Relatively speaking: It's easy to keep an eye on physical pieces of paper. It's tough to keep an eye on bits.
As for the idea that they would restrict content to only their channels, I dunno. I recall that a while ago Pepsi acquired some fast food franchises. So all the ones it didn't own, signed Coke contracts. Pepsi had begun competing with its customers. Maybe all the other cable providers would refuse to carry Disney content, since the cash would only go to Comcast, their competition. It's even more interesting to mix in some contemplation of Comcast's forays into the telco space, including -- I think -- MVNO arrangements with wireless carriers.
But can cable companies survive as such? As internet video gets better, what need will there be for content packaging and delivery on a geographic basis? I think all cable carriers must adapt or die. Owning content seems much more secure than owning a *legacy* means of delivering it.
I've been answering this question since I did my master's work on the subject 10 years ago. I commented in particular on an Arizona e-primary trial at one point.
As low-tech as it seems, there really are some useful properties of paper-based systems that seem hard to achieve when the physical tokens are removed.
(Note that I followed up with a correction to a researcher's name -- Cramer => Cohen.)
As you point out, receipts that allow you to verify a vote tend to provide a way to prove how you voted, and if they come in electronic form there are many large-scale exploits.
How many wonks could verify the result? Can they be trusted to cooperate but not to collude? There are crypto systems like that (such as the Cohen-type schemes I mention in my post), but the point is that a small number of people (relative to the voting population) then effectively control the outcome and secrecy of the election.
You must also contemplate what happens if a candidate or voter alleges fraud, regardless of whether the allegation is bogus.
You must also contemplate the lifetime of the encryption technology. Will ballot secrecy still be preserved a year after the election? How about 5, 10, or 20 years after it?
PKI has incidental importance, not the primary importance you seem to assign it. Of course it is nontrivial, but even assuming such infrastructure were in place and worked perfectly and were trusted by all (a big "if") and even assuming related infrastructure to function as an electronic "bulletin board" that could be trusted by all to make transactions indisputable and public (another big "if"), there are *still* significant weaknesses in every all-electronic voting scheme of which I'm aware.
Er, you seem like a true believer that "there must be a way to hold a purely electronic election." But can you cite any literature that gives specific schemes for accomplishing that, better than those I've mentioned? I admit I haven't done a recent literature review and I'd be extremely interested if you found something like that.
I skimmed it but didn't read the comments at bottom.
It's true that I don't have sales data, and I'd be curious to see it if anyone does. Fair point.
The article mentioned that some handsets in Japan have capabilities that iPhones don't. It also mentioned that some of those handsets have initial usability issues. But that spoke (however well or poorly) to nairbv's point about Japan having some "better" handsets available.
The article also pointed out various aspects of the iPhone that might not appeal to some elements of the Japanese consumer market. They seemed plausible to me.
As I understand things, it's not simply a matter of "better". Personalization is important in Japanese culture. There's huge variety of phones and consumers can find and tweak one to be "perfect" for each individual. But Apple seems to have a diametrically opposed "one size fits all" philosophy of consumers products -- "we built the most perfect product we could, and it's the one you should use." (I type this on a MacBook Air with a piece of paper taped across the camera that has no lens cap and can't be turned off). So from what I've heard, the iPhone has done poorly in Japan and the reason is cultural mismatch. If there's a university program pushing iPhones on students, it strikes me as not unlike Apple's historical practice of using educational systems to gain footholds in markets. But good luck to them trying this in Japan...
Ah, so maybe you're right about simply "better" phones. What I wrote at top was opinion stemming from 9 years working in the telco industry, quite often closely -- sometimes in person in Tokyo -- with my then-employer's Tokyo office and Japanese customers.
1. If you are concerned about a single e-voting system corrupting the data, you could have the data passed in parallel to multiple independently developed open-source systems for recording and tallying the votes.
It's not a question of scale. It's also not a question of whether the software used is open-source. It's a question of whether a purely electronic system can do a good job of simultaneously preserving both a secret ballot and an audit trail.
2. Why should we trust the electronic financial systems that manage our bank accounts, and billions of local and international financial transactions every day, yet not trust e-voting systems? Clearly there is just as much incentive to syphon off a billion or two dollars here or there as there is to sway an election.
What property is it of the electronic financial systems that enable us to, in general, trust them (despite a few occasional fraud cases)? Why could we not build that property in to our way of conducting computerized elections?
The audit trail in finance systems connects you with your transaction.
Notably, efforts towards "electronic cash" have gone nowhere. The security challenges involved are much the same as those for e-voting. If I hand you a bill, there's proof that you have more money, and I have less money, but there's no proof that I gave you money. How to achieve that without a physical bill? Nobody has a good answer yet (and a good answer might not exist).
3. In "poorer" or disorganized countries and failed states, paper-based elections, when conducted, are generally a complete joke. There is usually profound disagreement about cheating and results claims varying by 5 or 10 per cent. The things are decided by which side has the army or judges on side.
Are you seriously claiming that an independently run, cross-national, e-voting organization could not run a fairer election in these places?
It seems to me that the major problem we would have there is that powerful interests in the country, being used to being able to rig the elections, would not accept the use of the technology, paradoxically because it might elicit the truth about the voting intentions of its public, particularly if the election was conducted, as is eminently feasible with an e-voting system, over a period of several months or even six months, to minimize the possibilities for voter intimidation.
Yes, I am claiming that.
Interestingly, the USA does not meet international standards for election rigor. Happily the various interests and watchdogs in our country balance each other pretty well and we tend not to have such blatant corruption or vote-rigging problems. But we are not free of them, and never have been. There are theoretical problems -- with some degree of practical exploitation every election -- all over, including drawing districts, to registration, proving of identity at the polling place, absentee ballot handling, and more.
Look, I grew up in Chicago...
Anyway, should people in those countries trust us or any other external interest to certify the results of their elections? (Answer: probably not.)
4. In the first US election that George Bush was declared by some arbitrary legal powr broking and arm wrestling to have won, mathematically there was no result. That is, the voting process did not reach a decision, because the difference in result was within the margin of error of the manual voting and vote counting process.
No matter what system is used, very close elections run the risk of showing results that are within the margin of error. It's impossible that both sides will be happy with whatever result emerges. Oh, well.
Incidentally, the election you mentioned was not, I think as "bad" as the Kennedy-Nixon election a few decades ago. I believe it was argued by some that organized crime had attempted to influence the results. I think
It's not really clear that "the level of techno-scrutiny we need" is even theoretically possible. Many smart people have been thinking carefully about this problem for over a decade, and their conclusions on all-electronic systems are generally "don't do it." (Incidentally this is the same conclusion I reached while doing my master's work on the subject in 1999.) Paper ballots (printing them from electronic boxes is OK, as long as voters can read them) turn out to have really crucial characteristics.
As for, "It will permit consultation of populations on a much more frequent basis," some people might argue about whether referendums-on-everything is a good model for government, and I'll leave that (probably compelling) argument to any political science types who chime in. However you have reminded me of a point I missed in my earlier post, which is that preserving audit trails and secret ballots isn't the only thing the paper ballot system gives us. It also gives us a voting booth. Why is that important, you ask? Well, so that your boss can't sit you down in the office and watch you vote the company line; so that you won't be pressured by friends at "voting parties" with a common computer at someone's home, etc., etc. The voting booth provides privacy, an essential element of the secret ballot system.
Now if as a society we were to discard the secret ballot system, I'd see many feasible technical approaches for voting systems. That would be an interesting topic for any political science types, too -- I'd be curious to see that debate.
I did my master's work on e-voting in 1999, and followed it since.
You're right, but nailed only half of the issue (audit).
The other half is that we expect our elections to employ secret ballots. With paper, you can physically watch the ballots, even though you've dissociated voters from votes. The voter can see that the paper is marked as the voter intended, but not with anything that identifies the voter, and deposit it in a ballot box. The voter can further have confidence that, as you say, many interested parties who are unlikely to collude will then watch that ballot box carefully; and that the votes it contains will be counted. The press can witness the physical process of vote retrieval and counting.
But in all electronic systems, it seems impossible to provide both an audit trail and a secret ballot. Various schemes that attempt to deliver both properties one way or another require trust in a rather small set of entities whose actions are not very transparent to observers and who cannot necessarily be deemed "unlikely to collude". Those trusted parties have the power to control the outcome of the election, subvert the secret ballots, or both -- which effectively means that they, and not the voters, pick the winners of the election.
I've written application software for telco/carrier back ends, that's been deployed at carriers on every continent.
My layer sat on top of the infrastructure layer for SMS, but here's my best recollection from the internal training I got years ago.
Just like the article said, SMS was carried over the control path for call setup/takedown.
That path had very low bandwidth compared to the data (voice data, not IP data) path for calls. It was a control path and didn't need much. It was a limited resource with dedicated protocols and channels. Particularly in the early days of SMS, voice data was THE source of revenue for cell phones.
So using up the bandwidth in the control channel meant not being able to connect calls. There was a shortage of a precious resource whose over-use could strangle the primary flow of revenue to the carrier. So while the message itself was not expensive per se, it definitely was not without cost to the carrier. Carriers wanted to start exploring non-voice data services, and this gave them a way to get started without changing their existing network protocols and data flow, but they definitely had to be careful about how it was deployed and priced.
Of course telco pricing for ANYTHING is such a complicated market-driven competitive mess that there is an entire sub-industry of companies who keep track of all the special offers, plans, promotions, and whatnot that a given company has issued, knows which of its customers are in which plan, analyzes all the customer activity, and figures out who should be billed for what and how much. You're crazy if you think you can work out a "fair" price for a single element of telco service based on a guess of what it costs the telco to provide it.
At any rate, over time the carriers built up infrastructure around SMS, added IP gateways for the data that went over those dedicated control channels, added to their control channel capacity, etc., etc. Plus, in the US at least, SMS became and remained one of the most popular and steady data services among mobile users. All the major carriers in the US face significant business problems right now, and if they've got a small bright spot with predictable revenue at decent margins for a service folks like to use -- well, you can bet they're happy to have it and will charge what the market will bear (one way or another, in one plan or another, etc., etc.).
Wait, so why does interferon seem like a good treatment? From what you say it would perhaps delay progression of the virus's spread, but not fundamentally disrupt its operation.
My musing here is that maybe it's worked out that it's more effective to spread directly by cell-to-cell over the long term than to dump out a ton of virons and explode the cell.
Now the immune system is capable of identifying infected cells, but that's a lot harder to identify than a viron. So that is probably working to its advantage also. The white blood cells probably also have an easier time identifying a cel that's not doing its job anymore and is swollen with virons
Well, so, again, why isn't it a better research goal to force the virus to speed up viron production (thus defeating the cleverness you highlight), rather than helping it to slow down viron production (which enhances the cleverness you highlight)?
Slashdot Mirror
Well, it seems to me that what we're seeing is that it's easier to make and sell new smartphones -- and ways for them to consume bandwidth -- than it is to build out new mobile data infrastructure to grow the available bandwidth.
That is, I see this choice: Would you rather
1) that many people have smartphones but be forced to share the limited resource of bandwidth (meaning that each person gets only a certain amount); or
2) that a few people have smartphones and never encounter bandwidth constraints?
The wikipedia article for excalibur rounds says:
"The weapon can make first round strikes on targets up to 20 kilometres (12 mi) away."
http://en.wikipedia.org/wiki/M982_Excalibur
Isn't violation of attorney-client privilege kind of a big deal?
And those attacks were also domestic flights, not via US citizens but you don't check citizenship on domestic flights anyway
Controversial: Should we? I'd rather show my passport than go through a scanner or get a pat-down. Certainly I bet this 95-year-old woman would have preferred that.
and there have been plenty of home grown American nut jobs too.
...who have plenty of ways to carry out ill intent that have nothing to do with air travel. The TSA simply doesn't make us safer from them, even if you believe it makes air travel safer. In our non-air-travel lives, we seem pretty content with the trade off of nut job risk vs. law enforcement limitations.
From no one. They are scanning and patting down you to protect everyone else from you.
Well, OK, but I'm a *US citizen* traveling *domestically*. The law enforcement scrutiny I face at the airport is drastically disproportionate to what I face everywhere else in the country. Yet, a US citizen who intends harm to other US citizens isn't constrained to attack air travel. So I'm still scratching my head, here.
...from who are the TSA protecting me when they scan me or pat me down?
Companies should of course be free to offer compensation incentives on terms that fit their business needs. It's best for everyone if the incentives are expressed in plain language up front, so nobody feels tricked or taken advantage of later.
I also don't understand how the count can ever be verified without compromising the anonymity of the vote. If you don't trust the system you cannot trust any kind of verification it would do nor any kind of output it would produce (including any paper trail). Does anyone have any insight on the subject ?
The paper trail for a vote should be human-readable, inspected by the voter for correctness before deposit into a secure container at a polling place, and have no content that identifies the voter.
Physical pieces of paper can be physically watched (even shuffled) by multiple parties who are unlikely to collude (e.g. opposing political groups plus the press).
The 'e' part is just to speed counting. The paper makes it possible to handle claims of fraud, bugs, etc.
aside: Why does everyone keep misspelling 'rogue' in this thread? I've seen the 'rouge' reference dozens of times. That's a make-up product.
Anyway,
It sounds like the guy's able to run his own small server to meet innocuous intradepartmental needs in a basic sense. Again, in all of my posts here, I've said that the request is reasonable and hospital IT *should* be on board. The security issues are their problem; if his server breaks in other ways they can tell him it's his problem to get his service working for his department again.
It's obvious to you and me that IT ought to be on board. It clearly wasn't obvious to him. He came to us to ask about it. Why don't we just say, "yes, it's reasonable, and here are some reasons why"? Why engage in all of these characterizations of an accomplished professional who already is in touch with his IT department and is asking for advice?
It's true that Google can only divulge information known to it. However, if I were reading a hospital IT policy, I'd be in favor of one that actively discouraged sharing hospital data of any sort with Google.
It's just a calendar with shift times, as long as everyone involved doesn't mind it being on GCal, it shouldn't be a problem, but a random piece of hardware being connected to a hospital network IS a big deal. No one who's said anything is concerned about the safety of the shift times data, they ARE concerned about the HOSPITAL data that is on the same network & can be compromised by one security oversight on that server.
I agree that hardware on the hospital network is significant. In all of my comments on this thread, I've emphasized the importance of complying with IT policies, for exactly that reason.
I remain reluctant to agree that Google is a good alternate solution. I'm not trying to bash, nor to be antagonistic. I'm simply not convinced that it makes sense to use Google for hospital operations. From the OP, even the hospital's IT guys seem to be OK with the idea of hosting that service on the hospital network.
Aside from that issue: In fact I know (boy I hope it's not sacrilege to say this here) people who are gcal users, who REALLY don't like it. If this solution makes the users happy and is OK with hospital IT -- where's the beef?
It's a fair point, that Google doesn't have access to patient records. OTOH, hospital staff are probably better off relying on their internal computing resources than on Google, for their operations. Indeed, one way to *guarantee* that Google doesn't learn anything it shouldn't, is to have a policy of NOT using it for hospital operations. Furthermore, since OP is in touch with internal IT, he can make it *their* problem to ensure compliance with policies and legal requirements, either (a) by telling him not to run his own server or (b) asking that he meet certain conditions in order to do so. That also gives him an internal resource to contact if anything breaks, rather than relying on Google support.
When you say "unsupervised machine" -- in fact it's not unsupervised. He told IT about it and they (rightly IMO, and I said as much in my earlier post) requested access to it.
First, there's no reason apparent from the original post to conclude that a law has been violated. Second, your contempt is unwarranted. Third, your baseball bat proposal doesn't solve the original problem of how to migrate staff on-call schedules from paper to electronic form.
He's a doctor, a faculty member (professor), and a division head (administration/management). I promise you he's not a moron. There's a substantial amount of career achievement and hard work implicit in those credentials. Furthermore, he's actually made time to understand how to set up his own Linux server, and he's come here asking for advice. Even if you think the course he's pursuing isn't the right one, need you take this tone?
I'm surprised that you'd trust Google apps more than an internal server known to hospital IT.
The data at issue are on-call schedules for staff, not patient data.
To the original poster:
I think the request is reasonable. They probably want to do simple things like verify that the server is behaving as expected, monitor what processes are running, gather usage statistics, etc., etc. Do you have a specific worry as to what might transpire as a result of giving them access?
Adhering to IT's policies *probably* also means making them the ones on the hook for legal issues.
I honestly don't understand why there is such bias against electronic voting on Slashdot since, in theory, it's a "nerd community".
Yes, e-voting, after a lot of effort can be compromised. Regular paper-ballot voting can be compromised by anyone, skilled or not, with not a lot of effort at all. Any voting system can be compromised. I don't honestly understand why the Slashdot community dislike e-voting that much.
Relatively speaking: It's easy to keep an eye on physical pieces of paper. It's tough to keep an eye on bits.
As someone else noted, this is not the first time they've thought of this (with Disney in particular):
http://news.slashdot.org/comments.pl?sid=1337863&cid=29085761
As for the idea that they would restrict content to only their channels, I dunno. I recall that a while ago Pepsi acquired some fast food franchises. So all the ones it didn't own, signed Coke contracts. Pepsi had begun competing with its customers. Maybe all the other cable providers would refuse to carry Disney content, since the cash would only go to Comcast, their competition. It's even more interesting to mix in some contemplation of Comcast's forays into the telco space, including -- I think -- MVNO arrangements with wireless carriers.
But can cable companies survive as such? As internet video gets better, what need will there be for content packaging and delivery on a geographic basis? I think all cable carriers must adapt or die. Owning content seems much more secure than owning a *legacy* means of delivering it.
I've been answering this question since I did my master's work on the subject 10 years ago. I commented in particular on an Arizona e-primary trial at one point.
As low-tech as it seems, there really are some useful properties of paper-based systems that seem hard to achieve when the physical tokens are removed.
Here are some recent and not-so-recent posts:
http://slashdot.org/comments.pl?sid=6507&cid=940549 Re:Hrmph. Voting unsafe? July 12th, 2000
http://slashdot.org/comments.pl?sid=27682&cid=2975240 Re:All the arguments against online elections *02:06 PM February 8th, 2002
http://slashdot.org/comments.pl?sid=53211&cid=5263219 Re:Verifiable vote swapping is and should be illeg *05:17 AM February 9th, 2003
http://slashdot.org/comments.pl?sid=70945&cid=6434503 Re:There is always a Way *12:06 PM July 14th, 2003
http://slashdot.org/comments.pl?sid=77420&cid=6901725 Re:Why not use digital cash-like protocols? *01:49 PM September 8th, 2003
http://slashdot.org/comments.pl?sid=605423&cid=24086593 Re:The problem *01:56 PM July 7th, 2008
Finally, I recently had a several post long discussion with a fellow slashdotter underneath this May 29 2009 post:
http://slashdot.org/comments.pl?sid=1249937&cid=28144379
Hi,
I understand that you believe that.
I refer you to a post I made about 9 years ago, here:
http://slashdot.org/comments.pl?sid=6507&cid=940549
(Note that I followed up with a correction to a researcher's name -- Cramer => Cohen.)
As you point out, receipts that allow you to verify a vote tend to provide a way to prove how you voted, and if they come in electronic form there are many large-scale exploits.
How many wonks could verify the result? Can they be trusted to cooperate but not to collude? There are crypto systems like that (such as the Cohen-type schemes I mention in my post), but the point is that a small number of people (relative to the voting population) then effectively control the outcome and secrecy of the election.
You must also contemplate what happens if a candidate or voter alleges fraud, regardless of whether the allegation is bogus.
You must also contemplate the lifetime of the encryption technology. Will ballot secrecy still be preserved a year after the election? How about 5, 10, or 20 years after it?
PKI has incidental importance, not the primary importance you seem to assign it. Of course it is nontrivial, but even assuming such infrastructure were in place and worked perfectly and were trusted by all (a big "if") and even assuming related infrastructure to function as an electronic "bulletin board" that could be trusted by all to make transactions indisputable and public (another big "if"), there are *still* significant weaknesses in every all-electronic voting scheme of which I'm aware.
Er, you seem like a true believer that "there must be a way to hold a purely electronic election." But can you cite any literature that gives specific schemes for accomplishing that, better than those I've mentioned? I admit I haven't done a recent literature review and I'd be extremely interested if you found something like that.
Cheers,
-[Zappo]
I skimmed it but didn't read the comments at bottom.
It's true that I don't have sales data, and I'd be curious to see it if anyone does. Fair point.
The article mentioned that some handsets in Japan have capabilities that iPhones don't. It also mentioned that some of those handsets have initial usability issues. But that spoke (however well or poorly) to nairbv's point about Japan having some "better" handsets available.
The article also pointed out various aspects of the iPhone that might not appeal to some elements of the Japanese consumer market. They seemed plausible to me.
Finally, I'm guessing that *you* may not have read the editor's note about how the article changed after its initial publication (I read only the updated version):
http://www.wired.com/gadgetlab/2009/02/why-the-iphone/#editnote
As I understand things, it's not simply a matter of "better". Personalization is important in Japanese culture. There's huge variety of phones and consumers can find and tweak one to be "perfect" for each individual. But Apple seems to have a diametrically opposed "one size fits all" philosophy of consumers products -- "we built the most perfect product we could, and it's the one you should use." (I type this on a MacBook Air with a piece of paper taped across the camera that has no lens cap and can't be turned off). So from what I've heard, the iPhone has done poorly in Japan and the reason is cultural mismatch. If there's a university program pushing iPhones on students, it strikes me as not unlike Apple's historical practice of using educational systems to gain footholds in markets. But good luck to them trying this in Japan...
After I wrote the above, I found the following story on "Why the Japanese Hate the iPHone:"
http://www.wired.com/gadgetlab/2009/02/why-the-iphone/
Ah, so maybe you're right about simply "better" phones. What I wrote at top was opinion stemming from 9 years working in the telco industry, quite often closely -- sometimes in person in Tokyo -- with my then-employer's Tokyo office and Japanese customers.
1. If you are concerned about a single e-voting system corrupting the data, you could have the data passed in parallel to multiple independently developed open-source systems for recording and tallying the votes.
It's not a question of scale. It's also not a question of whether the software used is open-source. It's a question of whether a purely electronic system can do a good job of simultaneously preserving both a secret ballot and an audit trail.
2. Why should we trust the electronic financial systems that manage our bank accounts, and billions of local and international financial transactions every day,
yet not trust e-voting systems? Clearly there is just as much incentive to syphon off a billion or two dollars here or there as there is to sway an election.
What property is it of the electronic financial systems that enable us to, in general, trust them (despite a few occasional fraud cases)?
Why could we not build that property in to our way of conducting computerized elections?
The audit trail in finance systems connects you with your transaction.
Notably, efforts towards "electronic cash" have gone nowhere. The security challenges involved are much the same as those for e-voting. If I hand you a bill, there's proof that you have more money, and I have less money, but there's no proof that I gave you money. How to achieve that without a physical bill? Nobody has a good answer yet (and a good answer might not exist).
3. In "poorer" or disorganized countries and failed states, paper-based elections, when conducted, are generally a complete joke. There is usually profound disagreement about
cheating and results claims varying by 5 or 10 per cent. The things are decided by which side has the army or judges on side.
Are you seriously claiming that an independently run, cross-national, e-voting organization could not run a fairer election in these places?
It seems to me that the major problem we would have there is that powerful interests in the country, being used to being able to rig the elections,
would not accept the use of the technology, paradoxically because it might elicit the truth about the voting intentions of its public, particularly if the
election was conducted, as is eminently feasible with an e-voting system, over a period of several months or even six months, to minimize the
possibilities for voter intimidation.
Yes, I am claiming that.
Interestingly, the USA does not meet international standards for election rigor. Happily the various interests and watchdogs in our country balance each other pretty well and we tend not to have such blatant corruption or vote-rigging problems. But we are not free of them, and never have been. There are theoretical problems -- with some degree of practical exploitation every election -- all over, including drawing districts, to registration, proving of identity at the polling place, absentee ballot handling, and more.
Look, I grew up in Chicago...
Anyway, should people in those countries trust us or any other external interest to certify the results of their elections? (Answer: probably not.)
4. In the first US election that George Bush was declared by some arbitrary legal powr broking and arm wrestling to have won, mathematically there
was no result. That is, the voting process did not reach a decision, because the difference in result was within the margin of error of the manual
voting and vote counting process.
No matter what system is used, very close elections run the risk of showing results that are within the margin of error. It's impossible that both sides will be happy with whatever result emerges. Oh, well.
Incidentally, the election you mentioned was not, I think as "bad" as the Kennedy-Nixon election a few decades ago. I believe it was argued by some that organized crime had attempted to influence the results. I think
Hi,
Please see also my comment above:
http://news.slashdot.org/comments.pl?sid=1249937&cid=28147257
It's not really clear that "the level of techno-scrutiny we need" is even theoretically possible. Many smart people have been thinking carefully about this problem for over a decade, and their conclusions on all-electronic systems are generally "don't do it." (Incidentally this is the same conclusion I reached while doing my master's work on the subject in 1999.) Paper ballots (printing them from electronic boxes is OK, as long as voters can read them) turn out to have really crucial characteristics.
As for, "It will permit consultation of populations on a much more frequent basis," some people might argue about whether referendums-on-everything is a good model for government, and I'll leave that (probably compelling) argument to any political science types who chime in. However you have reminded me of a point I missed in my earlier post, which is that preserving audit trails and secret ballots isn't the only thing the paper ballot system gives us. It also gives us a voting booth. Why is that important, you ask? Well, so that your boss can't sit you down in the office and watch you vote the company line; so that you won't be pressured by friends at "voting parties" with a common computer at someone's home, etc., etc. The voting booth provides privacy, an essential element of the secret ballot system.
Now if as a society we were to discard the secret ballot system, I'd see many feasible technical approaches for voting systems. That would be an interesting topic for any political science types, too -- I'd be curious to see that debate.
I did my master's work on e-voting in 1999, and followed it since.
You're right, but nailed only half of the issue (audit).
The other half is that we expect our elections to employ secret ballots. With paper, you can physically watch the ballots, even though you've dissociated voters from votes. The voter can see that the paper is marked as the voter intended, but not with anything that identifies the voter, and deposit it in a ballot box. The voter can further have confidence that, as you say, many interested parties who are unlikely to collude will then watch that ballot box carefully; and that the votes it contains will be counted. The press can witness the physical process of vote retrieval and counting.
But in all electronic systems, it seems impossible to provide both an audit trail and a secret ballot. Various schemes that attempt to deliver both properties one way or another require trust in a rather small set of entities whose actions are not very transparent to observers and who cannot necessarily be deemed "unlikely to collude". Those trusted parties have the power to control the outcome of the election, subvert the secret ballots, or both -- which effectively means that they, and not the voters, pick the winners of the election.
How on earth would the Verizon employee know that its really a cop calling?
Some of your points are valid, but as for knowing who's calling... umm, they're the *phone company*.
As for the rest, there's a pretty good post on Verizon's actual procedures for dealing with it all, here: http://news.slashdot.org/comments.pl?sid=1242757&cid=28063387
I've written application software for telco/carrier back ends, that's been deployed at carriers on every continent.
My layer sat on top of the infrastructure layer for SMS, but here's my best recollection from the internal training I got years ago.
Just like the article said, SMS was carried over the control path for call setup/takedown.
That path had very low bandwidth compared to the data (voice data, not IP data) path for calls. It was a control path and didn't need much. It was a limited resource with dedicated protocols and channels. Particularly in the early days of SMS, voice data was THE source of revenue for cell phones.
So using up the bandwidth in the control channel meant not being able to connect calls. There was a shortage of a precious resource whose over-use could strangle the primary flow of revenue to the carrier. So while the message itself was not expensive per se, it definitely was not without cost to the carrier. Carriers wanted to start exploring non-voice data services, and this gave them a way to get started without changing their existing network protocols and data flow, but they definitely had to be careful about how it was deployed and priced.
Of course telco pricing for ANYTHING is such a complicated market-driven competitive mess that there is an entire sub-industry of companies who keep track of all the special offers, plans, promotions, and whatnot that a given company has issued, knows which of its customers are in which plan, analyzes all the customer activity, and figures out who should be billed for what and how much. You're crazy if you think you can work out a "fair" price for a single element of telco service based on a guess of what it costs the telco to provide it.
At any rate, over time the carriers built up infrastructure around SMS, added IP gateways for the data that went over those dedicated control channels, added to their control channel capacity, etc., etc. Plus, in the US at least, SMS became and remained one of the most popular and steady data services among mobile users. All the major carriers in the US face significant business problems right now, and if they've got a small bright spot with predictable revenue at decent margins for a service folks like to use -- well, you can bet they're happy to have it and will charge what the market will bear (one way or another, in one plan or another, etc., etc.).
Wait, so why does interferon seem like a good treatment? From what you say it would perhaps delay progression of the virus's spread, but not fundamentally disrupt its operation.
My musing here is that maybe it's worked out that it's more effective to spread directly by cell-to-cell over the long term than to dump out a ton of virons and explode the cell.
Now the immune system is capable of identifying infected cells, but that's a lot harder to identify than a viron. So that is probably working to its advantage also. The white blood cells probably also have an easier time identifying a cel that's not doing its job anymore and is swollen with virons
Well, so, again, why isn't it a better research goal to force the virus to speed up viron production (thus defeating the cleverness you highlight), rather than helping it to slow down viron production (which enhances the cleverness you highlight)?