LWN.net Logo

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 2:44 UTC (Wed) by drag (subscriber, #31333)
In reply to: Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org) by dskoll
Parent article: Video and photos show Linux booting on the Brazilian voting machines (BR-Linux.org)

> If you have the paper trails (ie, ballots), then why not just count the ballots and be done with it? Quicker, cheaper, more accurate, and much easier to understand. And is "provable". And "simple".

Because it's not Quicker, Cheaper, or More accurate.

It's vastly more expensive, inaccurate, and slower to count paper ballots. And much much more complicated to coordinate. And, believe it or not; prone to fraud.

The reason you have paper trails is for auditing purposes. Just like you have log files on your servers.

For a modern society you need to have electronic counting machines that are open, auditable, and simple.


(Log in to post comments)

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 2:58 UTC (Wed) by dskoll (subscriber, #1630) [Link]

It's vastly more expensive, inaccurate, and slower to count paper ballots.

No, it's NOT. We have our election results in Canada within hours. We already know who will form the government with about an hour of polls closing. 6-10 hours later, all results will be in and finalized.

The 2000 election in Canada cost about $200 million, or around $6 per Canadian citizen. Most of that cost was not related to the count, but to enumerating voters and reimbursing political parties for costs.

And much much more complicated to coordinate.

No, it is NOT. Elections Canada has been doing this successfully and efficiently for over 140 years. There has never been a federal election that was even remotely suspect in Canada, and certainly nothing like the fiasco in the United States in 2000.

And, believe it or not; prone to fraud.

You said that before. I asked for evidence. Well?

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 3:15 UTC (Wed) by drag (subscriber, #31333) [Link]

> You said that before. I asked for evidence. Well?

Every time there is a ballot recount the results come out different.

Organizations like Acorn are perpetrators of massive amounts of voter fraud involving fake people, dead people, moving people from different regions. This do it 'the hard way'.. that is getting and organizing people to commit multiple counts of fraud on a individual basis.

If your depending on individuals to do hand counts then how do you know that those people are trustworthy? It wouldn't take much orginization at all to get a few dozen people to nullify the results of many many thousands.

If you examine the sort of fraud the electronic machines are vulnerable too you'd realize that it's a game of percentages with them also. You won't be able to defraud a election more then a percent or two and get away with it. So election fraud from voting machines is only really something you have to pay attention to in very close elections. Which is the same thing from hand-counted ballots.

If you want to have the best 'proof' then you have to use both methods together.

> The 2000 election in Canada cost about $200 million, or around $6 per Canadian citizen. Most of that cost was not related to the count, but to enumerating voters and reimbursing political parties for costs.

Well in the USA we are much more hard-core about our politics. Also people are very paranoid and blow things far out of proportion. The Florida election issue from years ago is a HUGE example of normally intellegent people not understanding how election process works and problems being blown out of proportion for the purposes of bad politics.

(Hint: The president was never, ever, meant to be elected by popular vote. I'm convinced that politicians make a big big deal of the president to distract people from the elections that really matter; Senate and House of Representatives)

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 6:55 UTC (Wed) by anselm (subscriber, #2796) [Link]

If your depending on individuals to do hand counts then how do you know that those people are trustworthy? It wouldn't take much orginization at all to get a few dozen people to nullify the results of many many thousands.

The big advantage is that concerned citizens can come in and actually watch the ballots being counted. This is very different from an election official hitting a button and reading the counts off a little strip of paper that the voting machine spits out.

Counting paper ballots is something that people understand intuitively. Nobody can vouch for sure that the voting machine does what it is supposed to be doing - well, maybe under ideal conditions they can, but there are lots of incidents where voting machines are basically left sitting out in somebody's garage overnight before the actual election etc., and of course nobody knows for sure that the correct software was on the machine when it was put in the garage, let alone taken out the morning after.

Having said that, the paper ballot method seems to work very well for us here in Germany, thank you very much. But then again we tend to keep things simple; we don't elect the President and the municipal dog-catcher and everybody in between at the same time.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 7:16 UTC (Wed) by njs (subscriber, #40338) [Link]

>Organizations like Acorn are perpetrators of massive amounts of voter fraud involving fake people, dead people, moving people from different regions. This do it 'the hard way'.. that is getting and organizing people to commit multiple counts of fraud on a individual basis.

Err, you just demonstrated a huge amount of ignorance on this subject, FYI. The conversation is about voting fraud, not voter registration fraud, and the two are totally different. If you don't even distinguish between them then it rather throws the rest of your thoughts into doubt.

(Quick catchup: ACORN is a voter registration organization who pays people to bring them filled-out voter registration forms; sometimes people bring them ones full of nonsense, but they are *required by law* to file those nonsense forms anyway; invalid forms being submitted to election officials just creates more work for them screening out the nonsense, it doesn't result in any extra votes being cast; in principle clever fakes could be created and people could vote multiple times using them, but in fact no-one can find any evidence of this occurring in the US recently, despite a lot of people worrying about it very publicly and using it to justify things like voter roll purges that do, empirically, disenfranchise people. HTH. I suspect our politics differ, but that's no reason not to work from the same set of facts.)

>If your depending on individuals to do hand counts then how do you know that those people are trustworthy? It wouldn't take much orginization at all to get a few dozen people to nullify the results of many many thousands.

Which is why such counts have always been open to the public and attended by representatives of the opposing sides.

Voting is an interesting problem, with lots of tricky aspects; it's a fun literature to read up on.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 7:31 UTC (Wed) by drag (subscriber, #31333) [Link]

> Err, you just demonstrated a huge amount of ignorance on this subject, FYI. The conversation is about voting fraud, not voter registration fraud, and the two are totally different. If you don't even distinguish between them then it rather throws the rest of your thoughts into doubt.

Voter registration fraud _is_ voter fraud.

My point is that it's possible for people to orchestrate among relatively large amounts of people to defraud the voting system. In the case of Acorn it was on the voters side, not the counters side, but it's certainly possible both ways. Now they didn't get away with it in this instance, but they are not the only people doing very funny business.

It would be quite easy, if a voter official was not ethical, to 'stack' a group of counters to sway a local election. A unethical mayor or other powerful lower-level politician could sway the vote of a entire state in a closely contested area by putting like-minded people in charge and corrupting 10% or so of the counters in their county race.

And it's not something unique or new either. Some places, especially in Chicago, has been long notorious about people trying to get away with silliness on both sides. Going back many decades.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 8:32 UTC (Wed) by njs (subscriber, #40338) [Link]

>Voter registration fraud _is_ voter fraud.

The examples of registration fraud that have been cited regarding ACORN have been multiple registration forms filed by the same person (i.e., putting their own name on multiple forms), or people filling in forms with nonsense like "Mickey Mouse".

Neither of these leads to fraudulent votes on election day. (Though I'd love to see the news clips of Mickey arriving to do his patriotic duty. --on second thought he'd probably vote for someone like Rep. Berman, so never mind.) Both are screened out by the registrar of voters before election day even arrives.

Even given the level of competence we've come to expect from some Democratic-leaning political groups, as a clever conspiracy this lacks a certain something.

We know how to do voting. Computers are not necessary or even useful, except in limited cases for accessibility. Still nice to see the ideas of openness gaining currency among election officials, though.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 7:36 UTC (Wed) by nix (subscriber, #2304) [Link]

That's true only if you have dozens and dozens of 'elections' all rolled
into one, so the ballots are so complex nobody could dream of counting
them fast.

Hand counting *demonstrably works*. *Every* major industrialized democracy
uses it other than the US. Even the Swiss use it, and they're vote fiends.
It has failure modes, but they're known and everyone understands them. The
system is simple and comprehensible. Extra fraud methods are rare and very
difficult to think up. The system is not amenable to silent corruption by
remote parties without physical access to the ballot, which means
(assuming a sane system where the counts are conducted in situ) a huge
organizational problem that is highly likely to be detected.

You can't say that *any* of that is true of voting machines.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 8:56 UTC (Wed) by kragil (guest, #34373) [Link]

In germany we get very accurate projections of the vote a few seconds afer voting stopped. And we get final results the same night. And so far each test run with voting machines turned out to be vastly more expensive than paper.
The only real problem there is is having enough citizens helping with the election.

All your imaginary claims are just proven wrong by nearly every proper democracy there is.

Paper ballots then!

Posted Oct 15, 2008 13:26 UTC (Wed) by freemars (subscriber, #4235) [Link]

It's vastly more expensive, inaccurate, and slower to count paper ballots. And much much more complicated to coordinate. And, believe it or not; prone to fraud.

Optical scan paper ballots work well. A counting (not a voting) machine tabulates the ballots at the most local level. After the polls close election judges look at and sign off on the totals, and can call them in to HQ for the unofficial count. The machine, the tally sheet and the original/official ballots go back to election HQ for any recounts.

Fairly fast, not too expensive, and easy to audit.

machine counting can still be a problem

Posted Oct 15, 2008 20:05 UTC (Wed) by dlang (✭ supporter ✭, #313) [Link]

remember the florida fiasco? those were machine counted ballots.

a perfectly filled out ballot isn't the problem, the imperfectly filled out ballots are the problem

in Florida the ballots were pre-perferated to make it easier to punch out the vote that you wanted. the problem was that punching out the tab sometimes didn't completely remove it, combined with the pieces sometimes comeing loos with lots of handling of the ballot.

with optical scan ballots you have the problem of incomplely filled out circles, smudges causing non-filled out circles to be read as being filled out, etc.

I happen to agree that optically scanned ballots are probably the best bet right now for rapid and (reasonably) accurate counts. but I don't believe that they are completely reliable and error free.

Paper ballots then!

Posted Oct 16, 2008 9:41 UTC (Thu) by forthy (guest, #1525) [Link]

The "counting machines" have again been found to count wrong. These machines are broken by design, anyway. Counting votes starts with sorting votes - you look at the paper, and put it on an appropriate stack. You can identify dubious votes (e.g. cross not in the circle, but at the side, candidates stroked through, handwritten explanations on the paper like "I promised to vote for Osama to my grandchilds, but because he's a black terrorist, I vote for McSame", inappropriate pens causing marks elsewhere, etc.), and after a debate decide on which stack they go. You can recheck the stacks again and again, and by nature, your result will stabilize.

The result of a counting machine which doesn't sort however will never stabilize. Each time, it will interpret dubious votes differently, and therefore it looks like a Rübezahl effort. It's just because the machine is broken by design, because the designers never saw a real ballot counting procedure. The only risk of fraud is when votes are replaced during the counting process.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 15, 2008 19:56 UTC (Wed) by man_ls (subscriber, #15091) [Link]

A "me too" from Spain: simple paper ballots, manual count and recount, complete results in about 5 hours after polls close. Cheap and well coordinated. Since all political parties audit the results, fraud is unheard of.
For a modern society you need to have electronic counting machines that are open, auditable, and simple.
No, what you need to have is an election mechanism which is open, auditable and simple. Counting can be done by drunken donkeys as far as the system is concerned -- but in fact is done by randomly selected officials, for efficiency.

Why don't you think about simplifying your 18th century election systems instead of making it even more complex? In those days you had to carry the results by horse across vast expanses, so it probably made sense to ask a lot of things at a time. Not now.

Video and photos show Linux booting on the brazilian voting machines (BR-Linux.org)

Posted Oct 16, 2008 13:14 UTC (Thu) by Hanno (guest, #41730) [Link]

Hi,

you may want to read this

http://www.wijvertrouwenstemcomputersniet.nl/other/es3b-e...
(in English)

It is an instructing read an worth every page

> it's Quicker

True.

> Cheaper

False. All in all, a human ballot count is cheaper than a machine count.

Voting machines are prohibitively expensive. Buying _and_ maintaining voting machines costs a lot of money.

A machine count is only cheaper if (!) the machine is used with little change or maintenance for several votes over several years or decades. The widely used Nedap voting computers are based on technology that was "modern" 25 years ago.

People who opt for voting machines usually argue with the cost and then forget maintenance. This includes servicing outdated technology (see above) as well as exploring and patching security issues of voting machines once they are known.

Just two quick examples.

The ultra-modern "voting pen", a counting assistant device with a paper trail, was designed for the Hamburg elections of 2008. It has cost the city 4.5 million euros.

This money has been spent - and then the CCC demonstrated several simple hacks on how its vote can be rigged. One hack used a simple manipulation invisible to human vote ballot observers and voters.

The hack was demonstrated shortly before the vote and the digital pen was mothballed. Much ado was made by the city officials that they now have to find the needed volunteer ballot counters in a very short time and that the new voting rules may make the human count weeks to complete. However, enough volunteers signed up instantly and the count took a few hours for the party results and 1.5 more days for the regional candidate results. The regional press claims that the regional count was stretched by officials who wanted to discredit the human counting process.

The city of Hamburg has been widely criticized in German media a few days ago for flushing away this much taxpayers' money down the toilet.

http://www.abendblatt.de/daten/2008/10/10/950382.html?cmf=1

According to critics, the digital vote machine used in Amsterdam made a ballot more than 1 million Euro more expensive _per_ vote - the vote costs rose from 1.6 to 2.7 million. 0.9 million Euro of this went directly to the maintenance service parter for the digital vote machine - 3 Euro per voter.

http://www.heise.de/ct/06/23/036/

> more accurate

No. As others pointed out, the problem with vote counting is not those that can be counted easily. Human counters are better in this regard, as I can attest from my own experience as a ballot counter.

> [paper ballots are]
> much much more complicated to coordinate.

I have been a human ballot counter and will sign up to help for the next vote, again. It was not a complicated process, it was well coordinated and the count was fast enough.

> And, believe it or not; prone to fraud.

To rig a manual vote count, you have to conspire with several dozen people. All of them have to remain quiet after your candidate won.

To rig an electronic vote, you have to conspire with one or two specialists for software and electronics. (The Nedap hack and the Hamburg digital vote pen crack were done by small teams of volunteers.)

Which conspiracy is more likely?

> For a modern society you need to have electronic
> counting machines that are open, auditable, and simple.

Using "modern" as a reason for voting machines is a fallacy. As one critic puts it, voting machines are a solution to a problem we do not have. "Slow result count" is not a problem in votes.

Human ballot counts usually take a few hours or (max) days. There is no need to press a "result" button a minute after the vote ended. The sensationalist media should have no problem waiting for this result before they crown the new mayor, chancellor, prime-minister or president.

Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds