Re:Private Voting, Public Counting
on
Brave New Ballot
·
· Score: 1
Thanks for replying. I have some questions abour your (Australian) system.
How many races are listed per ballot? How many different types of ballots do you have per jurisdiction? How many elections do you have per year?
I live in King County Washington (USA). We have about 1.2m registered voters in a county of 1.8m people. I think roughly 660,000 people voted during the general election last year (2005). We have three elections per year, the special, the primary, and the general. Our county has over 4,000 different types of ballots. Typical ballots have between 10 and 30 races.
I'm told that the USA and Australian have some significant differences.
We vote for every thing. Most issues, like taxation, get a vote. All the elected offices from dog catcher up through President of the USA get a vote. All changes to government get a vote.
Also, our elections are decided by a winner-takes-all voting system. So 50% + 1 vote wins outright. Whereas in Australia's proportional representation, that's a tie. So individual votes in close elections are critically important in the USA and less so in Australia.
Please correct or expand any of my statements. I really appreciate the opportunity to learn more about everyone's various systems.
Re:Private Voting, Public Counting
on
Brave New Ballot
·
· Score: 1
Thank you for fulfilling one of my predictions.
I had a lot of misconceptions about how our voting and election systems work. So last year I was poll worker. (This year I got "promoted" to poll inspector. It means you have to sign for everything and be the sucker to take everything back to the warehouse at the end of the day.)
Please work an election to get a better understanding of the current systems.
Meanwhile, you're referring to a hypothetical system. Whereas Avi Rubin, myself, and many, many election integrity activists are referring to our existing systems. These existing systems are not easily audited, if the effort is made at all. VoterAction.org's lawsuits in New Mexico has some background on it.
Similarly, Jill LaVine testified to the Election Assistive Commission about their efforts to audit a VVPAT. The only such effort that I'm aware of at this time. The 1h 15m per ballot comes from her. For a link, please read my prior comments or use google. I'm tired.
Lastly, how would you scan the VVPAT? Using a digital device? Who certified that device? HAVA 2002 mandates that all vote tabulation equipment be certified. And how would a black box VVPAT scanner be any different in nature from any other black box system used in our elections.
The idea is to make our elections fair, open, and verifiable. All these gee-whiz solutions are ocunter to that goal.
Re:The fatal flaw in "voter verifiable"
on
Brave New Ballot
·
· Score: 1
A lot of thought has been expended (wasted?) attempting to answer the question "What is a meaningful audit?" The design of the testing is extremely important. Just listening to the elections experts and statisticians, it sounds like context is very important. Meaning the ways you'd audit large and small counties would probably be very different. I've heard numbers between 1% and 4.5%. Unfortunately, I have no clue how to determine something like this for myself.
A guy named Jerry Lobdill is taking a stab at the problem. His blog is http://southpaw.goodshow.net/. He hasn't posted his draft yet. If you're curious, maybe write him to get a copy.
Private Voting, Public Counting
on
Brave New Ballot
·
· Score: 5, Informative
The United States of America uses the "Australian Ballot" form of voting. That means everyone gets a secret ballot and the ballots are counted publicly. It was then, as it remains now, the best design to accommodate our system of elections. Please understand this essential reality before suggesting "improvements" (e.g. receipts, mail ballots, etc.).
Someone in this thread is going to state that HAVA 2002 mandates the use of electronic voting machines (aka "DRE" or direct recording electronic). That is false, as thoroughly explained in Voters Unite's Myth Breakers document.
Someone in this thread will make some statement about how electronic voting devices permit the disabled to vote in private. That's not exactly true. To the best of my knowledge, the existing products do not preserve the secret ballot. Nor are they particularly accessible. Meanwhile, there are solutions which do preserve the secret ballot and are accessible to disabled voters. Such as ES&S's AutoMark, the Vote-PAD, and EqualiVote. (There are some other novel systems, too. I just haven't researched them yet.)
Someone in this thread is going to state that electronic voting is just splendid, and we can make it work, if we just try harder next time. Fine. Show me. Then let's talk. Meanwhile, all current systems suck.
Someone in this thread is going to suggest that we have all paper ballots counted manually. Like Canada. Or Germany. It's not a bad idea. But it wouldn't work in the USA with our current constraints and expectations. To contrast, in Canada, the races are very simple and so the tabulation is feasible. In Germany, they have proportional representation and rely on their superior form of exit polls. Meaning their system is very tolerant of errors. And they have legions of civil servants working weeks to get the exact manual tally. Whereas here in the USA, politicians and news networks demand results now, now, now!
Someone in this thread may suggest it's all about the Republicans. Or the Democrats. It hasn't proven that simple. I believe it's a fight between the people in power, who want to stay in power, and us voters. I'm a pretty progressive guy. But I readily acknowledge the bad guys (with respect to election integrity) here in King County Washington are in the Democratic leadership. (My experience is that the rank and file of both major parties are completely on board with election integrity.)
Someone in this thread may also suggest that we eliminate the need for electronic voting at poll sites by transistioning to forced mail voting (100% vote by mail). Like Oregon State has done and where most of Washington State is heading. It's terribly idea. No more secret ballot. No more public vote count. Higher error rate. Huge more expensive. Long-term decline in voter turnout. It's a big topic. We've been researching it for about 9 months and have only scratched the surface. We discuss
Someone in this thread will also exhort the necessity of using a voter verified paper audit trail. They may even encourage others to support Rush Holt's HR 550. Unfortunately, the VVPAT is a placebo. What guarantees what's recorded is what's printed? Nothing. And experiences to date demonstrate that actually auditing the VVPAT is infeasible (1h 15m per ballot cast). That said, the efforts of VerifiedVoting.org and other are not misguided. Many states already have electronic voting machines without the VVPAT. So passing HR 550 would be better than nothing.
The take away point is this:
The most reliable, secure way to vote in the USA today is to use voter-correctable precinct-based optical scanners. That means paper ballots at poll sites fed into a ballot scanner.
Please support Voter Action. They have successfully prevented the use and procurement of electronic voting machines in a few states already. They are expanding the fight as fast as they can
Remember, too, that voting legislation, in large part in response to issues in the 2000 election, designed to ensure fair, uniform, and universal access to voting for all citizens by mandating electronic voting equipment, such as HAVA (2002), were Democratic and bipartisan efforts.
HAVA does not require electronic voting equipment. This is standard FUD. Please read the legislation. Even better, read the Myth Breakers analysis posted on VotersUnite.org (for the past few years).
HAVA requires accessibility for disabled voters. There are many non-electronic solutions available, such as the Vote-PAD and EqualiVote. There are even ballot marking solutions, such as ES&S's AutoMark.
Unfortunately, between unscrupulous behavior and the Federal and State voting equipment certification process, HAVA is being used to ram electronic voting down our throats.
What we need to require is a permanent, voter-verified, auditable paper trail...
I disagree. What we need to retain (and often times regain) is paper ballots.
Voter-verified paper audit trails are a placebo. What assurance do you have that what is printed is the same as what is recorded? None.
All attempts to date to actually audit a VVPAT, to the best of knowledge, have demonstrated just how infeasible the task is. Jill LaVine, Sacramento County's Registrar of Voters testified to the EAC that their audit took 1h 15m per ballot printed on the VVPAT.
Meanwhile, many people, like VerifiedVoting.org are proponents of Rush Holt's HR 550, which would require all electronic voting machines to have a VVPAT. Even though I utterly oppose all electronic voting, I do not oppose HR 550. Why? Because HR 550 requirements would demonstrate the folly of using electronic voting machines and the voter verified paper audit trail.
I will note here that New Mexico (VoterAction.org), Connecticut (TrueVoteCT.org), and others are successfully throwing out the DREs and bringing in voter-correctable precinct-based opticals scanners. That is today's best available solution.
Thank you for the post. I'm sorry for the delay replying.
Honestly, the use of secure hash functions for protecting privacy is not intuitively obvious to me. It feels just like trying to those Zen koans (e.g. the sound of one hand clapping). I bought Peter Wayner's book Translucent Databases to learn about this stuff and get some practice. He details a bunch of ideas which are similar to yours.
When I have chance, I'll work through your idea. I wouldn't be able to prove or disprove it in any way. So don't expect much. But it's notions like yours that need to explored in order to fully understand why things are the way they are.
No it doesn't. First off the status quo of a non-verifiable method of counting votes already allows widespread voter fraud. Remember 2000? In the case of buying votes, the absentee ballot allows one to prove to others how one voted.
The title of my comment is "private voting, public counting". The verification comes from the public vote count.
Regarding absentee ballots, I agree that the loss of both the secret ballot and the public vote count opens the door to fraud. Hence my opposition to forced mail voting (my name for "100% vote by mail").
At least with this system the voter herself can verify how her vote was assigned. Some checks could be put in place that the only person that can check the list must have a receipt assigned to them and a valid ID. Maybe even biometrics. Also copying the list by those who are not mandated by the government is strictly forbidden.
Sure. Great ideas. You square that with the law and then let's talk. I can't wait to see what you come up with.
Also to claim such a system removes the secret ballot is just playing semantics. The current system is secrecy through obfuscation. A new system would be secrecy through encryption.
I cite the constitution, laws, tradition, and legal precedent and that's just "semantics" to you. Interesting.
The current poll site voting system in King County preserves the secret ballot and has a public vote count. We use paper ballots and voter-correctable precinct-based optical scanners. It's the cheapest, most reliable, least error-prone system currently available.
The "new" system of electronic voting and forced mail voting, being phased in over the next few years, eliminates the secret ballot and the public vote count. I oppose these "new" systems.
The "future" systems you're so enthusiastic about are currently just academic papers. I can't wait to see their real-world implementations. You're working on that, right? Please let me know when they're ready for demo.
Meanwhile, please learn something about our current system before initiating your sweeping changes. That technophilic mentality is exactly how we got in this mess in the first place.
That's a very interesting paper. Thank you for the link. I've printed it out and will try to digest the algorithm. Looks I may have to amend my "cannot be done" position to merely "shouldn't be done". Haha.
Seriously, I'd want to two things when considering the system Chaum proposes. First, I'd want to see it in action. The devil's in the details. Second, I'd want to see a list of pros and cons. One pros he lists is having a uniform means for handling normal and provisional ballots. That's a pretty good benefit. (I'm serving as a poll worker again this year. Rules and handling of provisional ballots are pretty complicated.)
That private key is your receipt. Enabling you to prove how you voted. Opening the door to coercion, vote buying, and other traditional forms of vote fraud. Ergo, no more secret ballot.
Maybe you're advocating the elimination of the secret ballot. Okay. But that's a separate discussion.
Why can't an electronic voting machine with a paper trail satisfy the private voting/public counting principle?
Damned good question.
The reason is because the VVPAT (voter verified paper audit trail) is a placebo.
What reason would anyone have to believe that the tally recorded in the memory card (and uploaded to the central tabulator) is the same as what is printed? Two different data paths. Enables two different results.
Voter Action determined that in New Mexico that Spanish language ballots were printed corrected but not recorded in memory. (Sorry, I couldn't quickly find the specific cite.)
The report from the recent botched election in Cuyahoga County Ohio had all sorts of problems related to the VVPAT. Sure, hypothetically one could design and build a VVPAT system that wasn't likely to break down, rip the paper, had good ergonomics, etc. But I prefer to talk about the actual systems we're actually using. And these actual systems actually suck.
The one attempt to audit the VVPAT that I know of resulted in the election officials quickly choosing to use PBOS over electronic voting systems with VVPAT. You can read the testimony Jill LaVine, Sacramento County's Registrar of Voters, gave to the Election Assistance Commission this last April. Brief summary: The manual recount took 1 hour and 15 minutes per ballot cast.
Lastly, your mileage may vary state to state. Some states treat the VVPAT as the legal ballot of record. Some treat the memory card as the legal record. Some don't use the VVPAT for recounts. Etc. Honestly, I don't keep close track of such things. The proponents of Holt's HR 550, like Verified Voting do a good job on that issue, if you want to know more.
Exactly right. Election integrity is a non-partisan issue.
Right now, it's us voters against the politicians and the corporations. And many of us (voters) are getting screwed.
As Andrew Gumbel details in the book "Steal This Vote", corruption is primarily a function of opportunity. Maybe the Republicans are currently better at it. But there's plenty of blame to go around.
Here in King County Washington, as elsewhere, the Democrats are the ones ushering in electronic voting and forced mail voting. And then from the right, we have our Republican Secretary of State is doing his bit to monkey with the statewide voter registration database. (The Brennan Center recently one their lawsuit against the new rules: Washington Association of Churches et. al. v. Reed.)
So, from where I'm sitting, neither political party is looking very pretty.
There's lots of good posts. I'm glad we geeks are talking about this important issue.
I spoke briefly with Bev Harris recently. See below.
I'm at work, so I need to make this brief. Just four points.
First, the two pillars of our democracy (United States of America) are private voting and public counting. We adopted the Australian Ballot (aka secret ballot) a while back. Things like electronic voting and forced mail voting (e.g. 100% vote by mail) take away the secret ballot. Here in Washington State, our constitution says we need a secret ballot. Disagree if you want. There's lots of ideas. Like voting receipts and no more secret ballots. But please start by changing our laws. Meanwhile, any attempt to take away the secret ballot (private voting) is unconstitutional.
Second, there is no technical way to have an electronic voting system which both preserves the secret ballot and the public vote count. If the ballots are secret, then there's no verifiability, meaning no public count. If the system is verifiable, then there's no secret ballot. You can have one or the other, but not both. Electronic counting, as with the precinct-based optical scanners, can be done constitutionally.
Third, currently the most reliable way to vote in the USA is to use a voter-correctable precinct-based optical scanner (PBOS). Sorry, I don't have the cites handy (my bad), but dig a little and you can find the research on this. Brennan Center, GAO reports, MIT Voter Project, etc. The basic idea is that you mark a ballot and feed it into a machine. If there's a problem, the machine spits the ballot back out, giving the voter a chance to correct the problem. Yes, these machines need to be better designed, open source, yadda, yadda. But before anyone proposes a better system, please work to understand the best system currently available. (Thank you for your patience.)
Many juridictions have wisely moved away from touchscreens and other DREs and adopted PBOS systems with a low-cost, verifiable solution for disabled voting. TrueVoteCT.org just had a huge win. And Voter Action sued and got the touchscreens in New Mexico replaced with PBOS systems. (Please visit both orgs and give them cash. Activism is not cheap!)
Fourth, and lastly, Bev Harris made an incredibly important point: Our elections have to be understandable for all the voters. Blackbox Voting has spents years digging and researching. I've personally spent 2 years learning all that I can about elections, voting, and these systems. I'm a computer geek and I readily admit that I had to work pretty hard to understand stuff. Bev has a lot of contact with experts, computer scientists, security dudes, etc. Her point is that we cannot rely on those sage gurus to weigh in on our election systems. We all need to understand how our democracy works. Not just the wonks. That means our election and voting systems must be simple and straightforward.
(PS- I saw Bev during King County Washington's "logic and accuracy testing" of our new Diebold AccuVote TSx touchscreens last Tuesday. You can read "Report: Testing of Diebold AccuVote TSx" on my blog, on WashBlog, or on dailyKos. Please holler if anyone has questions. I'll do my best to reply in a timely fashion.)
Re:Not possible to take all threats seriously
on
Saving Lives with Design
·
· Score: 2, Interesting
Excellent questions.
Here's a thought: You could actually work just a wee bit to find answers some answers for yourself. The logical, and popular, places to start would be 9-11 Commission
and the important followup
Intelligence Matters.
Of course, during this period where everyone gets to choose their own facts, you can choose to accept what is obvious to rational observers. Or not.
This is only about winning, at any cost. If the GOP was actually concerned about fraud, they'd insist that electronic voting machine vendors like Sequoia open up their boxes for independent inspection. As it is today, the contracts stipulate that vote counting is a trade secret.
Why do you think Mono is now days away from a 1.0 release, while gcj and classpath are still lagging the current Java?
Excellent example. Except for one small little detail. Mono is dramatically less ambitious than gcj and classpath. Not to denigrate either effort, of course.
Is a courtesy citation so hard? The poster is referring to the article "Does Better Analysis Make Products Worse?" in Ralph Grabowski's upFront.eZine newsletter. The URL for the current issue still points to #378, but should be updated shortly.
Your requests are reasonable, but unrealistic. Sorry.
Re: #1 (micromanagement) It's all about trust. I've tried the handsoff approach. It only works with a few select high-functioning people. The rest need to be handled like kindergarteners with flamethrowers hellbent on derailing your schedule.
re: #2 (be honest) Managers like me, stuck between the upper and lower layers, don't have much wiggle room. It's pretty rare that we get an accurate picture of the overall health of the company. So cut us some slack.
re: #3 (hire adults) I wish! The biggest success factor is team building. And I've yet to have much say in who gets assigned to my teams.
re: #4 (pop-management-psychology) Most, but not all, of it is crap. Ditto methodologies and other theologies. Believe me, fending off that nonsense is a constant battle.
I sent the following reply Dare. It might be interesting to others.
---
Hi Dare-
Thanks for writing and posting your comparison. Eventually, I'll need to get some project experience using C# -- your article will surely prove helpful.
I do have an initial comment. I believe C#'s exposure of low-level machine constructs to be misguided and ultimately counterproductive.
C#'s "stack based classes" {see section 4, "Value Types (Structs)"} is one such feature. As with garbage collection, the decision to allocate objects on the stack or heap is best left to the runtime (virtual machine). Note that nothing in the Java Specs disallows allocating objects on the stack; that's an implementation detail.
Jalapeño is IBM's research Java Virtual Machine (recently released as the open source Jikes JVM). Here's an interesting paper detailing their technique for determining when objects can be allocated on the stack:
Escape Analysis for Java
http://www.research.ibm.com/jalapeno/publication.h tml#oopsla99_escape
(The paper also covers the elimination of unnecessary synchronization; also very cool.)
So if C#'s "struct" keyword (for stack-based classes) isn't necessary, neither is the "boxing" hack for wrapping "struct" objects in a class.
Escape analysis and the like could probably be applied to C# (and the CLR). But why provide the language feature if it's unnecessary?
The worst part, though, is that C#'s stack-based classes feature focuses attention on the implementation instead of on the problem domain. I'm sure there's a quote about premature optimization that applies.
Quite a few posters criticize Sun for not refining the Java language (fast enough). I am probably the most impatient Java user on this planet. I want const, enums, overhaul of threading, tossing all the deprecated stuff, etc.
But I'm reminded by something that Bill Joy said at the first or second JavaOne conference. He was relating his experience with NFS. It took years for everyone to make compliant implementations. He concluded, probably correctly, that the most important factor to securing adoption of the Java language is to not change it.
Having watched the evolution of Java and the JDK since the alpha2 release, I'm inclined to agree. It wasn't until JDK 1.2 that WORA became a matter of discipline, versus somewhat hard. And JDK 1.4 (now in beta) is truly boss.
Slashdot Mirror
Thanks for replying. I have some questions abour your (Australian) system.
How many races are listed per ballot? How many different types of ballots do you have per jurisdiction? How many elections do you have per year?
I live in King County Washington (USA). We have about 1.2m registered voters in a county of 1.8m people. I think roughly 660,000 people voted during the general election last year (2005). We have three elections per year, the special, the primary, and the general. Our county has over 4,000 different types of ballots. Typical ballots have between 10 and 30 races.
I'm told that the USA and Australian have some significant differences.
We vote for every thing. Most issues, like taxation, get a vote. All the elected offices from dog catcher up through President of the USA get a vote. All changes to government get a vote.
Also, our elections are decided by a winner-takes-all voting system. So 50% + 1 vote wins outright. Whereas in Australia's proportional representation, that's a tie. So individual votes in close elections are critically important in the USA and less so in Australia.
Please correct or expand any of my statements. I really appreciate the opportunity to learn more about everyone's various systems.
Thank you for fulfilling one of my predictions.
I had a lot of misconceptions about how our voting and election systems work. So last year I was poll worker. (This year I got "promoted" to poll inspector. It means you have to sign for everything and be the sucker to take everything back to the warehouse at the end of the day.)
Please work an election to get a better understanding of the current systems.
Meanwhile, you're referring to a hypothetical system. Whereas Avi Rubin, myself, and many, many election integrity activists are referring to our existing systems. These existing systems are not easily audited, if the effort is made at all. VoterAction.org's lawsuits in New Mexico has some background on it.
Similarly, Jill LaVine testified to the Election Assistive Commission about their efforts to audit a VVPAT. The only such effort that I'm aware of at this time. The 1h 15m per ballot comes from her. For a link, please read my prior comments or use google. I'm tired.
Lastly, how would you scan the VVPAT? Using a digital device? Who certified that device? HAVA 2002 mandates that all vote tabulation equipment be certified. And how would a black box VVPAT scanner be any different in nature from any other black box system used in our elections.
The idea is to make our elections fair, open, and verifiable. All these gee-whiz solutions are ocunter to that goal.
A lot of thought has been expended (wasted?) attempting to answer the question "What is a meaningful audit?" The design of the testing is extremely important. Just listening to the elections experts and statisticians, it sounds like context is very important. Meaning the ways you'd audit large and small counties would probably be very different. I've heard numbers between 1% and 4.5%. Unfortunately, I have no clue how to determine something like this for myself.
A guy named Jerry Lobdill is taking a stab at the problem. His blog is http://southpaw.goodshow.net/. He hasn't posted his draft yet. If you're curious, maybe write him to get a copy.
The United States of America uses the "Australian Ballot" form of voting. That means everyone gets a secret ballot and the ballots are counted publicly. It was then, as it remains now, the best design to accommodate our system of elections. Please understand this essential reality before suggesting "improvements" (e.g. receipts, mail ballots, etc.).
Someone in this thread is going to state that HAVA 2002 mandates the use of electronic voting machines (aka "DRE" or direct recording electronic). That is false, as thoroughly explained in Voters Unite's Myth Breakers document.
Someone in this thread will make some statement about how electronic voting devices permit the disabled to vote in private. That's not exactly true. To the best of my knowledge, the existing products do not preserve the secret ballot. Nor are they particularly accessible. Meanwhile, there are solutions which do preserve the secret ballot and are accessible to disabled voters. Such as ES&S's AutoMark, the Vote-PAD, and EqualiVote. (There are some other novel systems, too. I just haven't researched them yet.)
Someone in this thread is going to state that electronic voting is just splendid, and we can make it work, if we just try harder next time. Fine. Show me. Then let's talk. Meanwhile, all current systems suck.
Someone in this thread is going to suggest that we have all paper ballots counted manually. Like Canada. Or Germany. It's not a bad idea. But it wouldn't work in the USA with our current constraints and expectations. To contrast, in Canada, the races are very simple and so the tabulation is feasible. In Germany, they have proportional representation and rely on their superior form of exit polls. Meaning their system is very tolerant of errors. And they have legions of civil servants working weeks to get the exact manual tally. Whereas here in the USA, politicians and news networks demand results now, now, now!
Someone in this thread may suggest it's all about the Republicans. Or the Democrats. It hasn't proven that simple. I believe it's a fight between the people in power, who want to stay in power, and us voters. I'm a pretty progressive guy. But I readily acknowledge the bad guys (with respect to election integrity) here in King County Washington are in the Democratic leadership. (My experience is that the rank and file of both major parties are completely on board with election integrity.)
Someone in this thread may also suggest that we eliminate the need for electronic voting at poll sites by transistioning to forced mail voting (100% vote by mail). Like Oregon State has done and where most of Washington State is heading. It's terribly idea. No more secret ballot. No more public vote count. Higher error rate. Huge more expensive. Long-term decline in voter turnout. It's a big topic. We've been researching it for about 9 months and have only scratched the surface. We discuss
Someone in this thread will also exhort the necessity of using a voter verified paper audit trail. They may even encourage others to support Rush Holt's HR 550. Unfortunately, the VVPAT is a placebo. What guarantees what's recorded is what's printed? Nothing. And experiences to date demonstrate that actually auditing the VVPAT is infeasible (1h 15m per ballot cast). That said, the efforts of VerifiedVoting.org and other are not misguided. Many states already have electronic voting machines without the VVPAT. So passing HR 550 would be better than nothing.
The take away point is this:
The most reliable, secure way to vote in the USA today is to use voter-correctable precinct-based optical scanners. That means paper ballots at poll sites fed into a ballot scanner.
Please support Voter Action. They have successfully prevented the use and procurement of electronic voting machines in a few states already. They are expanding the fight as fast as they can
HAVA does not require electronic voting equipment. This is standard FUD. Please read the legislation. Even better, read the Myth Breakers analysis posted on VotersUnite.org (for the past few years).
HAVA requires accessibility for disabled voters. There are many non-electronic solutions available, such as the Vote-PAD and EqualiVote. There are even ballot marking solutions, such as ES&S's AutoMark.
Unfortunately, between unscrupulous behavior and the Federal and State voting equipment certification process, HAVA is being used to ram electronic voting down our throats.
I disagree. What we need to retain (and often times regain) is paper ballots.
Voter-verified paper audit trails are a placebo. What assurance do you have that what is printed is the same as what is recorded? None.
All attempts to date to actually audit a VVPAT, to the best of knowledge, have demonstrated just how infeasible the task is. Jill LaVine, Sacramento County's Registrar of Voters testified to the EAC that their audit took 1h 15m per ballot printed on the VVPAT.
Meanwhile, many people, like VerifiedVoting.org are proponents of Rush Holt's HR 550, which would require all electronic voting machines to have a VVPAT. Even though I utterly oppose all electronic voting, I do not oppose HR 550. Why? Because HR 550 requirements would demonstrate the folly of using electronic voting machines and the voter verified paper audit trail.
I will note here that New Mexico (VoterAction.org), Connecticut (TrueVoteCT.org), and others are successfully throwing out the DREs and bringing in voter-correctable precinct-based opticals scanners. That is today's best available solution.
Thank you for the post. I'm sorry for the delay replying.
Honestly, the use of secure hash functions for protecting privacy is not intuitively obvious to me. It feels just like trying to those Zen koans (e.g. the sound of one hand clapping). I bought Peter Wayner's book Translucent Databases to learn about this stuff and get some practice. He details a bunch of ideas which are similar to yours.
When I have chance, I'll work through your idea. I wouldn't be able to prove or disprove it in any way. So don't expect much. But it's notions like yours that need to explored in order to fully understand why things are the way they are.
Thanks.
The title of my comment is "private voting, public counting". The verification comes from the public vote count.
Regarding absentee ballots, I agree that the loss of both the secret ballot and the public vote count opens the door to fraud. Hence my opposition to forced mail voting (my name for "100% vote by mail").
Sure. Great ideas. You square that with the law and then let's talk. I can't wait to see what you come up with.
I cite the constitution, laws, tradition, and legal precedent and that's just "semantics" to you. Interesting.
The current poll site voting system in King County preserves the secret ballot and has a public vote count. We use paper ballots and voter-correctable precinct-based optical scanners. It's the cheapest, most reliable, least error-prone system currently available.
The "new" system of electronic voting and forced mail voting, being phased in over the next few years, eliminates the secret ballot and the public vote count. I oppose these "new" systems.
The "future" systems you're so enthusiastic about are currently just academic papers. I can't wait to see their real-world implementations. You're working on that, right? Please let me know when they're ready for demo.
Meanwhile, please learn something about our current system before initiating your sweeping changes. That technophilic mentality is exactly how we got in this mess in the first place.
That's a very interesting paper. Thank you for the link. I've printed it out and will try to digest the algorithm. Looks I may have to amend my "cannot be done" position to merely "shouldn't be done". Haha.
Seriously, I'd want to two things when considering the system Chaum proposes. First, I'd want to see it in action. The devil's in the details. Second, I'd want to see a list of pros and cons. One pros he lists is having a uniform means for handling normal and provisional ballots. That's a pretty good benefit. (I'm serving as a poll worker again this year. Rules and handling of provisional ballots are pretty complicated.)
That private key is your receipt. Enabling you to prove how you voted. Opening the door to coercion, vote buying, and other traditional forms of vote fraud. Ergo, no more secret ballot.
Maybe you're advocating the elimination of the secret ballot. Okay. But that's a separate discussion.
Damned good question.
The reason is because the VVPAT (voter verified paper audit trail) is a placebo.
What reason would anyone have to believe that the tally recorded in the memory card (and uploaded to the central tabulator) is the same as what is printed? Two different data paths. Enables two different results.
Voter Action determined that in New Mexico that Spanish language ballots were printed corrected but not recorded in memory. (Sorry, I couldn't quickly find the specific cite.)
The report from the recent botched election in Cuyahoga County Ohio had all sorts of problems related to the VVPAT. Sure, hypothetically one could design and build a VVPAT system that wasn't likely to break down, rip the paper, had good ergonomics, etc. But I prefer to talk about the actual systems we're actually using. And these actual systems actually suck.
The one attempt to audit the VVPAT that I know of resulted in the election officials quickly choosing to use PBOS over electronic voting systems with VVPAT. You can read the testimony Jill LaVine, Sacramento County's Registrar of Voters, gave to the Election Assistance Commission this last April. Brief summary: The manual recount took 1 hour and 15 minutes per ballot cast.
Lastly, your mileage may vary state to state. Some states treat the VVPAT as the legal ballot of record. Some treat the memory card as the legal record. Some don't use the VVPAT for recounts. Etc. Honestly, I don't keep close track of such things. The proponents of Holt's HR 550, like Verified Voting do a good job on that issue, if you want to know more.
Again, great question. Keep 'em coming.
Exactly right. Election integrity is a non-partisan issue.
Right now, it's us voters against the politicians and the corporations. And many of us (voters) are getting screwed.
As Andrew Gumbel details in the book "Steal This Vote", corruption is primarily a function of opportunity. Maybe the Republicans are currently better at it. But there's plenty of blame to go around.
Here in King County Washington, as elsewhere, the Democrats are the ones ushering in electronic voting and forced mail voting. And then from the right, we have our Republican Secretary of State is doing his bit to monkey with the statewide voter registration database. (The Brennan Center recently one their lawsuit against the new rules: Washington Association of Churches et. al. v. Reed
So, from where I'm sitting, neither political party is looking very pretty.
There's lots of good posts. I'm glad we geeks are talking about this important issue.
I spoke briefly with Bev Harris recently. See below.
I'm at work, so I need to make this brief. Just four points.
First, the two pillars of our democracy (United States of America) are private voting and public counting. We adopted the Australian Ballot (aka secret ballot) a while back. Things like electronic voting and forced mail voting (e.g. 100% vote by mail) take away the secret ballot. Here in Washington State, our constitution says we need a secret ballot. Disagree if you want. There's lots of ideas. Like voting receipts and no more secret ballots. But please start by changing our laws. Meanwhile, any attempt to take away the secret ballot (private voting) is unconstitutional.
Second, there is no technical way to have an electronic voting system which both preserves the secret ballot and the public vote count. If the ballots are secret, then there's no verifiability, meaning no public count. If the system is verifiable, then there's no secret ballot. You can have one or the other, but not both. Electronic counting, as with the precinct-based optical scanners, can be done constitutionally.
Third, currently the most reliable way to vote in the USA is to use a voter-correctable precinct-based optical scanner (PBOS). Sorry, I don't have the cites handy (my bad), but dig a little and you can find the research on this. Brennan Center, GAO reports, MIT Voter Project, etc. The basic idea is that you mark a ballot and feed it into a machine. If there's a problem, the machine spits the ballot back out, giving the voter a chance to correct the problem. Yes, these machines need to be better designed, open source, yadda, yadda. But before anyone proposes a better system, please work to understand the best system currently available. (Thank you for your patience.)
Many juridictions have wisely moved away from touchscreens and other DREs and adopted PBOS systems with a low-cost, verifiable solution for disabled voting. TrueVoteCT.org just had a huge win. And Voter Action sued and got the touchscreens in New Mexico replaced with PBOS systems. (Please visit both orgs and give them cash. Activism is not cheap!)
Fourth, and lastly, Bev Harris made an incredibly important point: Our elections have to be understandable for all the voters. Blackbox Voting has spents years digging and researching. I've personally spent 2 years learning all that I can about elections, voting, and these systems. I'm a computer geek and I readily admit that I had to work pretty hard to understand stuff. Bev has a lot of contact with experts, computer scientists, security dudes, etc. Her point is that we cannot rely on those sage gurus to weigh in on our election systems. We all need to understand how our democracy works. Not just the wonks. That means our election and voting systems must be simple and straightforward.
(PS- I saw Bev during King County Washington's "logic and accuracy testing" of our new Diebold AccuVote TSx touchscreens last Tuesday. You can read "Report: Testing of Diebold AccuVote TSx" on my blog, on WashBlog, or on dailyKos. Please holler if anyone has questions. I'll do my best to reply in a timely fashion.)
Here's a thought: You could actually work just a wee bit to find answers some answers for yourself. The logical, and popular, places to start would be 9-11 Commission and the important followup Intelligence Matters.
Of course, during this period where everyone gets to choose their own facts, you can choose to accept what is obvious to rational observers. Or not.
Evidence Of Election Irregularities In Snohomish County, Washington, General Election, 200
This is only about winning, at any cost. If the GOP was actually concerned about fraud, they'd insist that electronic voting machine vendors like Sequoia open up their boxes for independent inspection. As it is today, the contracts stipulate that vote counting is a trade secret.
That's just lovely.
Excellent example. Except for one small little detail. Mono is dramatically less ambitious than gcj and classpath. Not to denigrate either effort, of course.
Is a courtesy citation so hard? The poster is referring to the article "Does Better Analysis Make Products Worse?" in Ralph Grabowski's upFront.eZine newsletter. The URL for the current issue still points to #378, but should be updated shortly.
Your requests are reasonable, but unrealistic. Sorry.
Re: #1 (micromanagement) It's all about trust. I've tried the handsoff approach. It only works with a few select high-functioning people. The rest need to be handled like kindergarteners with flamethrowers hellbent on derailing your schedule.
re: #2 (be honest) Managers like me, stuck between the upper and lower layers, don't have much wiggle room. It's pretty rare that we get an accurate picture of the overall health of the company. So cut us some slack.
re: #3 (hire adults) I wish! The biggest success factor is team building. And I've yet to have much say in who gets assigned to my teams.
re: #4 (pop-management-psychology) Most, but not all, of it is crap. Ditto methodologies and other theologies. Believe me, fending off that nonsense is a constant battle.
I sent the following reply Dare. It might be interesting to others.
h tml#oopsla99_escape
---
Hi Dare-
Thanks for writing and posting your comparison. Eventually, I'll need to get some project experience using C# -- your article will surely prove helpful.
I do have an initial comment. I believe C#'s exposure of low-level machine constructs to be misguided and ultimately counterproductive.
C#'s "stack based classes" {see section 4, "Value Types (Structs)"} is one such feature. As with garbage collection, the decision to allocate objects on the stack or heap is best left to the runtime (virtual machine). Note that nothing in the Java Specs disallows allocating objects on the stack; that's an implementation detail.
Jalapeño is IBM's research Java Virtual Machine (recently released as the open source Jikes JVM). Here's an interesting paper detailing their technique for determining when objects can be allocated on the stack:
Escape Analysis for Java
http://www.research.ibm.com/jalapeno/publication.
(The paper also covers the elimination of unnecessary synchronization; also very cool.)
So if C#'s "struct" keyword (for stack-based classes) isn't necessary, neither is the "boxing" hack for wrapping "struct" objects in a class.
Escape analysis and the like could probably be applied to C# (and the CLR). But why provide the language feature if it's unnecessary?
The worst part, though, is that C#'s stack-based classes feature focuses attention on the implementation instead of on the problem domain. I'm sure there's a quote about premature optimization that applies.
Again, thank you for posting your comparison!
Cheers, Jason
But I'm reminded by something that Bill Joy said at the first or second JavaOne conference. He was relating his experience with NFS. It took years for everyone to make compliant implementations. He concluded, probably correctly, that the most important factor to securing adoption of the Java language is to not change it.
Having watched the evolution of Java and the JDK since the alpha2 release, I'm inclined to agree. It wasn't until JDK 1.2 that WORA became a matter of discipline, versus somewhat hard. And JDK 1.4 (now in beta) is truly boss.