Volkswagen Passat Forum banner

1 - 20 of 23 Posts

·
Registered
Joined
·
4,095 Posts
Discussion Starter #1
http://www.avirubin.com/judge.html

I sure hope this guy (and people like him who DO have a clue) get to review things before the real election.

already it seems there is a vote of no-confidence in diebold.

and of course, diebold has gone on record saying 'we are committed to deliverying the election to bush' (paraphrasing).

anyone else feel outraged at the sloppiness of diebold's whole setup?
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #3
part of the point of the expose is that THERE IS ZERO CHANCE OF ANY MEANINGFUL RECOUNT!

you fark with the 'zero machine' and all bets are off (so to speak). and what does a recount mean, if the data is munged before the 'recount' can be done?

there's no paper trail that is meaningful. there is no open code review of the diebold machines. and the ceo of diebold has already revealed his hand.

lets not worry about messy things like recounts - our new super-schnell service totally eliminates all chances of asking for a recount - since ITS IMPOSSIBL TO _DO_ A RECOUNT!

doesn't this worry anyone??
 

·
Registered
Joined
·
12,116 Posts
I think there is a serious cause for concern no matter what your party affiliation (or lack there-of) not necessarily because the potential for tampering by the vendor but more because of certain state's (like California) rushing into it too quickly IMO. I would have preferred to see dual-counts for at least a few elections. Diebold argues that no voting system is completely secure and I agree but the question is whether their e-voting system offers any value over traditional paper votes other than sending a message to voters that "something" is being done to help prevent another Florida debacle.

Rubin's report is exactly what we need but it is obviously hyper-critical. Diebold's comments are of course on the the other end of the spectrum. I'd probably feel worse if the major independant analysis came back and said everything was perfect.

I don't think the lack of a recount potential is meaningful because a tempered election is a tampered election. 2000's results showed us that recounts don't prove anything when every recount comes back with different results. I honestly believe that it's better to put all your effort into making the first result go smoothly because it's easier to tamper with votes once they've been in a box for six months.

Diebold's responses to Rubin's report aren't as contrived as he makes them out to be: http://www2.diebold.com/checksandbalances.pdf
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #5
its pretty simple, really - when deploying 'new systems' (doesn't matter what discipline) you ALWAYS RUN THE OLD AND NEW IN PARALLEL for long enough for your target users to feel comfortable, that the two systems continue to get the same answer. but we're not doing that. we're DUMPING one and JUMPING to another. big big mistake!

the point isn't to count votes more efficiently. its to make LESS ERRORS. one simple way is to run parallel processes and checkpoint them periodically. you see if electronic(voter)=paper(voter). if not, then an alarm goes off and you have to 'deal with it' someway. but the point is that the parity error was caught - and that's a Good Thing.

its very clear that the current crop of evote-machines are not ready for prime time.

and so, jumping from a flawed system to a more-flawed system makes less than no sense to me ;( ;(
 

·
Registered
Joined
·
12,116 Posts
I agree running in parallel is a good idea but I haven't yet seen evidence to call the new system "more-flawed." They should be run in parallel so we can gather that evidence if it exists because invariably the dual counts would not match up.

They do have an argument against it in their rebuttal linked above- if you use e-voting machines simply as ticket printers then you haven't gained anything. If they are tampered with, they will just print bad tickets. The point of change is to improve the system. Otherwise you have to ask the voters to vote twice, introducing human error ("damn, what did I vote on prop 56?")

There are systems where you wouldn't want to deploy in parallel.. for example not asking people to continue using radium treatments when they were found to be harmful just as a measuring stick against replacement treatments. But is this one of those systems? Would asking voters to vote twice defeat the attempt to improve voter confidence and introduce too much opportunity for human error? Maybe..
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #7
no, by running in parallel, I meant simply that the single machine would print out a voter-verified paper ticket that was used as one system of actual counting; and the electronic stuff would go thru an entirely different path.

and actually, it should STAY this way. one side parity's the other. for something as important (and recently, highly questioned) as our vote, we deserve to use quality systems.

I would also add that since these serve the public, and public funds paid for them, they should be like other government-paid-for code - it should be visable and examinable by ANYONE who wants. we all should be able to check-up on our voting system, to our heart's content. its my understanding that there is zero initiative for diebold to release their code. that, in itself, is suspect behavior.
 

·
Registered
Joined
·
4,721 Posts
if the evil machines scare you, just go absentee, you get a paper ballot, and can vote from nearly anywhere in the world.

plus you can mail your ballot in early or late if you hope to skew the results... :suspicio: :crazy: :lol:
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #9
jimbob said:
if the evil machines scare you, just go absentee, you get a paper ballot, and can vote from nearly anywhere in the world.

plus you can mail your ballot in early or late if you hope to skew the results... :suspicio: :crazy: :lol:
exactly what leo on tech-tv says ;)

no proof that my ballots are going anywhere but /dev/null (circular file) anyway. where's the feedback that my vote was REGISTERED and COUNTED?

of course we never had that. never. but with today's tech, we could.

I am sad for the HUGE potential that e-voting could bring. accountability, for one. and its not unreasonable for each voter to get a receipt with a serial # on it and be able to check it, at will, online.

but it all comes down to 'I did stuff to a UI and I sure hope the data goes somewhere and is used'. beyond that, its all a blackbox. blackboxes don't inspire confidence. at least not in the minds of computer scientists.

I think the e-voting is used to cast an emotional appeal to voters. reminds me of an incident in high school (some 30 yrs ago). computers were still kind of new in school - and one day we played a trick on someone. we entered some data we commonly knew about this girl and just created 'PRINT"blah blah blah" statements. nothing really more than that. just waiting for someone to hit enter and it would spit all that back. so we told this girl that we had access to FBI files and if we entered her name, all kinds of info would come back. so we 'entered' her name and hit return. the PRINT's came out. she was blown away. and probably scared, too.

well, things haven't changed. grandmas are still convinced that if its on computer printout, its accurate. they don't KNOW the GIGO principle. they think that if its electronic and OKd by our government, that it was properly tested and its going to work.

sigh...

we dont' know if there are haywire PRINT statements in there (so to speak) or not. we just don't know. but its all TOO EASY to fark things up and since there's no notion (at all) of a recount, we'll NEVER KNOW if we've been duped or not.

we are far far worse off now than when we had hanging chads.
 

·
Registered
Joined
·
12,116 Posts
linux-works said:
no, by running in parallel, I meant simply that the single machine would print out a voter-verified paper ticket that was used as one system of actual counting; and the electronic stuff would go thru an entirely different path.
Right, and if the machine has been tampered with it will give a correct receipt to the voter and transmit false results to the central office. Nothing has changed. It could also print out a paper that could be stored by the poling place to be counted later and ask the voter to confirm that it's what they wanted, while transmitting false results to the central office. Now the paper and electronic counts don't match- who wins? How do you identify which system was flawed, the human count or electronic one? Again you've gained nothing.

I'm just playing devil's advocate here. I think they should have done a few voting cycles completely discarding the elctronic results just to build up some experience with them, but that idea does not give the voter confidence they wanted to build for THIS election.

I don't know how you justify the statement that we are now worse off. Just because the system is not perfect (none are) does not automatically mean it's worse than human counts. All you can hope for is a smaller margin for error and this system may give us that.
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #11
my point is that collectively, the experts could design a damned near bulletproof and confidence-inspiring system. in full daylight. not in the dark, like diebold did.

they didn't. the gov creapt around in the dark like they always do. they farked up this whole election already. I have zero confidence in the voting system now. and I'm not alone in how I feel, either. congrats, guys... ;(

evoting can be a really cool thing. but it needs to follow an OPEN design and implementation model if it has any hopes of being secure (I know it sounds counter-intuitive, but its correct, nonetheless)
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #12
Macabre said:
I don't know how you justify the statement that we are now worse off. Just because the system is not perfect (none are) does not automatically mean it's worse than human counts. All you can hope for is a smaller margin for error and this system may give us that.
we're worse off because we've DISCARDED the old system (de-installed it) and went with the new. before it has proven itself and before it was ready.

there's no recount now, so whatever we get, we get...

that disturbs me highly.
 

·
Registered
Joined
·
12,116 Posts
linux-works said:
my point is that collectively, the experts could design a damned near bulletproof and confidence-inspiring system. in full daylight. not in the dark, like diebold did.

they didn't. the gov creapt around in the dark like they always do. they farked up this whole election already. I have zero confidence in the voting system now. and I'm not alone in how I feel, either. congrats, guys... ;(

evoting can be a really cool thing. but it needs to follow an OPEN design and implementation model if it has any hopes of being secure (I know it sounds counter-intuitive, but its correct, nonetheless)
Huh, what? Going to a vendor and purchasing services isn't "creaping around in the dark." It's the way large entities like governments purchase things. I'd have to assume that Diebold went up to bid using the state mandated bidding process which varies by state of course and won, then followed a normal development cycle to deliver a product. It is up to each state to determine how many third-party audits are necessary and some were performed (I found reference to at least two).

I have a feeling that you're trying to imply that only open source products can be secure and that is simply ludicrous. Completely closed designs are inherently insecure but there's a big difference there, and that is not the case with this product.
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #14
Macabre said:
Huh, what? Going to a vendor and purchasing services isn't "creaping around in the dark." It's the way large entities like governments purchase things. I'd have to assume that Diebold went up to bid using the state mandated bidding process which varies by state of course and won, then followed a normal development cycle to deliver a product. It is up to each state to determine how many third-party audits are necessary and some were performed (I found reference to at least two).
and I do blame diebold more than the US gov. the US doesn't know any better (sic) but the vendor is supposed to be 'an expert'. and this expert will not tell us (voters) how their machines work. I just find that suspect.

I have a feeling that you're trying to imply that only open source products can be secure and that is simply ludicrous. Completely closed designs are inherently insecure but there's a big difference there, and that is not the case with this product.
opensource isn't the only way. but its the most trusted way. it is. it allows any of us to examine what is going on. easiest way to shut up the opposition is to let them have read access to the source code. if/when we find no problems, cool. but let us have the chance. too much is riding on this for 'trust us - we tested it and its ok'. from the vendor itself?? yeah, right.

instead of the gov paying money for a closed system, they could have had a collaborative effort from the experts (who would probably even volunteer their time, just as a civic duty) for free. and you'd get better code, more 'eyes' seeing/checking it and less worries since its no longer a black box that we can't look inside.

is closed-source always insecure? no. is it as secure as open code? not usually. really - not usually.
 

·
Registered
Joined
·
12,116 Posts
I'd have to wonder what you're basing that on. There's a certain type of product that works in an open source environment and I doubt this would be one of them. Products with a specific use with *very* strict requirements do not fit into an open source development process well at all. The place for open source development is similar to the place for public research - as a foundation for product to be built upon. Products themselves don't work very well in that model.

However it's not relevant because although you are an exception, the public at large would not have confidence in an open source product of this nature and the whole purpose of the endeavor is to build confidence.
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #16
right now, you have folks like avi crying 'foul'.

opensource it and once all the checkmarks (fixes) are in, people like he and I stop calling 'foul'.

oh, and compare the stability and bug-count of, say, freebsd to anything MS has. anything. opensource has faster fixes and is more verifyable that there are no glaring security issues inherent in the design or even the implementation.

I use SSH as an example of 'better life thru opensource'. the commercial version has had more bugs (and nastier ones) than the opensource version! by your logic, that shouldn't ever happen. ssh is an application, like you said. yet the free/open one is better and more stable. why is that?

apps vs. o/s - that's not the issue. what is the issue is inspectability. and design. and neither is verifyable in the diebold case. that's bad.
 

·
Registered
Joined
·
12,116 Posts
Ahh, I thought this thread was about voting machines, but apparently it's an open-source soapbox. I'll leave you to it! :D
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #18
no, I just suggested that something as important as our voting should NOT be done 'in secret'.

and especially NOT by a company who has publicly announced their agenda to do whatever they can to assure one particular candidate gets in office.

opensource is one way. hell, hire avi and let him create an indy team of experts and I'll trust HIM. but I need independant verification from SOMEONE that the voting system works.

right now, I dont' see any double-verify and that worries me a lot.
 

·
Registered
Joined
·
4,095 Posts
Discussion Starter #19
http://www.cleveland.com/editorials/plaindealer/index.ssf?/base/opinion/1062322233282931.xml

In a perfect world, companies that make voting equipment would be apolitical. But it's not a perfect world.

Still, you would think that the CEO of a company that wants to make a lot of money selling voting machines to Ohio would see the value of at least pretending impartiality. Instead, Diebold Chief Executive Walden O'Dell committed himself in a recent fund-raising letter to work to "deliver [Ohio's] electoral votes" to President George W. Bush.
 

·
Registered
Joined
·
12,116 Posts
Well keep in mind that Rubin does *not* have the actual source code used in the systems, or any specs on the hardware. They found some code on Diebold's public website and did their analysis on that. Diebold claims that only a small part of the code they did their analysis on is actually used in the voter kiosks.

Here is the results of the first third-party audit done on the complete system, paid for by the state of Maryland: http://www.dbm.maryland.gov/dbm_search/technology/toc_voting_system_report/votingsystemreportfinal.pdf and http://www.dbm.maryland.gov/dbm_search/technology/toc_voting_system_report/votingsystemreportappb.pdf

The following was written by the independant auditor, SAIC, at the top of the second document
The following table is a brief analysis of statements made by Professor Rubin, et al, in their report on the Diebold source code entitled
“Analysis of an Electronic Voting System”, July 23, 2003. In general, SAIC made many of the same observations, when considering
only the source code. While many of the statements made by Mr. Rubin were technically correct, it is clear that Mr. Rubin did not
have a complete understanding of the State of Maryland’s implementation of the AccuVote-TS voting system, and the election process
controls or environment. During this assessment, SAIC had access to system and election documentation, personnel and equipment.
Applying the NIST Risk Assessment methodology to the evaluation of the equipment in its operational environment and the totality of
the management, operational, and technical controls, SAIC reached many different conclusions. Indeed, Professor Rubin states
repeatedly in his paper that he does not know how the system operates in an election and he further identifies the assumptions that he
used to reach his conclusions. In those cases where these assumptions concerning operational or management controls were incorrect,
the resultant conclusions were, unsurprisingly, also incorrect.


Here is the details of the second third-party audit: http://www.raba.com/press/TA_Report_AccuVote.pdf

Both of these have their share of very negative comments, and I don't want to come off as a Diebold fanboy here, but my point is that they are very much open to independant verification (as I'm sure was part of their contract with various state agencies). This process is not even remotely out of the ordinary. There's lots more to how product development works in public and private sector than what slashdot tells you :)

There have been other audits done but I do not have the details.
 
1 - 20 of 23 Posts
About this Discussion
22 Replies
5 Participants
F. Monger
Volkswagen Passat Forum
Passatworld is a forum dedicated to Volkswagen Passat enthusiasts to discuss mods, Quattro, Turbo Diesel, reviews and more!
Full Forum Listing
Top