State Refuses to Release Public Records-Claims Diebold Data is Proprietary

[Wilms]

Alaska: State Refuses to Release Public Records

By Kay Brown, Alaska Communications Director, DNC

January 23, 2006

State Claims Diebold Data is Proprietary


The State Division of Elections has denied access to public records that are needed to verify the accuracy of the 2004 General Election vote results, the Alaska Democratic Party (ADP) said today.

The Division of Elections claims that its electronic computer file that contains all the final vote tallies for the 2004 General Election is proprietary information belonging to its contractor, Diebold Election Systems.

Although Diebold claims that their data structure is proprietary, it is publicly available on the Internet today and has been for several years.

Numerous discrepancies are apparent in the 2004 General Election votes tallied by the state's Diebold computer system and posted on the Division of Election's web site. According to the posted “Statement of Votes Cast” by district and precinct, a far larger number of votes were cast than the official totals reported in the statewide summary. In many of the House Districts, more votes are shown in the totals than there are voters in the district. In the case of President George Bush’s votes, the district-by-district totals add up to 292,267, but his official total was only 190,889, a difference of 101,378 votes. In the U.S. Senate race, Lisa Murkowski received 226,992 votes in the district-by-district totals, but her official total was only 149,446, a difference of 77,546 votes.

snip

The GEMS software and data files are publicly available and can be viewed online. http://www.equalccw.com/dieboldtestnotes.html

snip

http://www.votetrustusa.org/index.php?option=com_content&task=view&id=826&Itemid=113

[Fredda Weinberg]

1. I've examined the code

there was nothing nefarious there. Other computer scientists have reached the same conclusion.

You're beating a dead horse.

[Wilms]

2. Pardon me?

[Fredda Weinberg]

8. I said you're beating a dead horse

Any questions?

[kster]

22. Yes, why does the

State Refuse to Release Public Records and Claim that Diebold Data is Proprietary. I would be embarrassed if I had to tell people that. Do you think that excuse is a valid one?

[Fredda Weinberg]

46. It's not just a claim

Diebold's software and database schema *are* proprietary. As noted elsewhere in this thread, you can easily extract the data and it's been offered, but refused.

[Wilms]

47. Quotes and links will help move the discussion along. n/t

[Fredda Weinberg]

49. See reply #26

[Wilms]

53. A printout?

For Sale...Cheap.

http://www.nhc.rtp.nc.us:8080/pds/gilded/progress/images/brooklynbridgelrg.jpg

[Fredda Weinberg]

55. Access data is easily exported into spreadsheets or text files

The data is physically in the election officials' office, so the extract could be supervised by objective observers, if such things exist anymore ...

[Wilms]

57. And it was made clear why that is unacceptable

http://www.votetrustusa.org/index.php?option=com_content&task=view&id=830&Itemid=113

[Fredda Weinberg]

58. Sorry, but the reasoning is flawed

For example

The database structure available for download at March's site is the same structure in the central tabular data file we are seeking.

They have no way of knowing this. They make several references to Alaska law, but no citations I can verify. And as I noted in your second thread on this topic, March's having a copy doesn't make the code or schema public domain.

[Wilms]

59. And you're protecting Diebold why?

[Fredda Weinberg]

60. Ah, the insinuations begin ...

I'm stating facts and backing up my assertions. If Diebold is being attacked for valid reasons, I'll be the first to cooperate.

[OnTheOtherHand]

62. PMJI, I notice a certain family resemblance here

I've encountered a lot of people who think that the "raw exit poll data" contain some sort of key to unlocking 2004. It's a very strange conviction from my POV. If we are looking for outliers to examine, we have quite a few based on the official election returns; we don't have to settle for exit poll scraps.

Mutatis mutandis, to make the case for releasing the .mdb file, it is important to be specific about why only the .mdb file will do. I guess the answer is this:

It has been documented that the GEMS software was programmed to contain a "double set of books" within it, and thus a print out from one file selected by Diebold will not enable a comprehensive understanding of what has caused the bizarre and inaccurate reports produced by the Diebold system for the state's 2004 election reports.

http://www.votetrustusa.org/index.php?option=com_content&task=view&id=830&Itemid=113

Based on publicly available accounts, I see no reason to assume either that a "double set of books" accounts for the inaccuracies in the reports, or that examining the .mdb will "enable a comprehensive understanding." It seems like a bit of a fishing expedition. It would be helpful if a computer scientist familiar with the structure of the GEMS database could offer a hypothesis about how the "doublt set of books" might account for the observed inaccuracies.

The general category of 'bizarre results attributable to design decisions' is not at all unfamiliar to me. For instance, we know that the NEP exit poll conducted interviews at multiple-precinct polling places and plugged in those interviews -- drawn from all precincts -- as if they were drawn from just one precinct. The designers judged that as acceptable error, given that the interview data is eventually replaced by vote returns anyway. On Election Night it probably doesn't matter much, but it creates a fog that drifts through the forensic analysis of the exit polls.

An even more counter-intuitive one that I didn't really notice until recently: the telephone version of the national exit poll incorporates questions from all four national exit poll forms. There is a variable in the database that allows analysts to select version(s) of the form. So, the designers opted to enter each telephone survey four times, once for each version, and set the weights accordingly. This means, inter alia, that the national exit poll actually contains 1500 fewer interviews than it appears to. This decision doesn't affect the results of the survey if it is used as intended, but it does mean that if you "count Bush and Kerry votes (responses) in the dataset," you will end up with more votes than there were respondents. That may, or may not, be a pretty close analogy to what happened in Alaska.

[stickdog]

71. Why the FUCK do we need to justify public examination of the

database and software being used to COUNT OUR VOTES? What the FUCK is your justification for ANY lack of transparency in this CRITICAL PUBLIC arena?

[OnTheOtherHand]

73. hello? what do you think _I_ have to do with it?

It is fun to sit around being self-righteous at each other's expense, I guess. But I don't see how it makes a tinker's dam worth of difference in the real world.

I'm curious whether there is any reason to expend special effort on this particular issue. So far, it looks like a dry well to me. But I'm not the one drilling it, so, whatever.

[stickdog]

105. My point is simply that any database used to count the votes in

a public election should be able to be examined by the public. If you disagree with me, please explain.

[OnTheOtherHand]

138. sorry, I missed the question here

Honestly, I don't know whether I even agree with that as a point of abstract principle. Certainly I care less about it than the principle that everyone in the U.S. should have access to decent health care. But frankly, it doesn't matter much how strongly I believe that everyone in the U.S. should have access to decent health care.

I think the premise that "the public" ever could examine database structures is a bit daft. But if the CS community lines up behind the idea that making DB designs open to all is the key to better security, hey, why not? I happen to have some programming background, but it doesn't include serious DB design, and it isn't my main gig. I do not assume that I understand all the relevant issues.

But again, it doesn't matter. We are not discussing whether it was a good idea for Alaska to enter into a contract that prevents them from handing over the .mdb (assuming that's what they did). At least, I'm not. That train already left the station. If we are actually worried about massive fraud in Alaska in 2004, we have other recourses. And if we aren't, well, what are we talking about, anyway?

[stickdog]

167. The train didn't leave the station. The state of Alaskan controls the

database they purchased from Diebold. Diebold's dubious contractual proprietary claims on it are completely outweighed by importance of public confidence of free and fair elections. If you don't agree, please explain yourself.

[OnTheOtherHand]

168. no, you explain yourself

First of all, can you tell us all with a straight face that if you can just look at the .mdb file, your confidence in free and fair elections will be fully restored?

Failing that, can you at least tell us that you suspect that the .mdb file contains clear evidence of fraud? If so, what might that evidence be, and why do you think that it might be there?

Failing either of those, what exactly is it that outweighs Diebold's contractual claims, however dubious you might consider them?

It doesn't really matter whether you convince me -- but if you can't convince me, I don't like your chances in the real world.

[stickdog]

169. What the fuck are Diebold's proprietary claims based on?

Is there some hot black market for bloated, flawed and insecure Access vote storage databases of which the entire IT industry is currently unaware? What exactly is Diebold afraid of? That the Alaskan Democratic Party will steal their revolutionary new method of counting and make a killing selling it to Enron accountants?

The state of Alaska bought their GEMS database from Diebold. Certainly, the data in it belongs to Alaska. Certainly, the "state of Alaska" must be allowed to examine its own vote storage mdb file as a matter of government. So why can't a few experts from the Alaskan Democratic party -- with the state's permission and after signing noncompetition agreements if necessary -- examine the entire database used to store Alaskan voting results to determine for themselves what role, if any, this database played in widespread reported vote total discrepancies?

Surely you must agree that widespread vote totalling discrepancies are of compelling interest to both the general public and any major political party. Surely you must agree that a transparent and trustworthy voting system is an extremely important public interest in any putatively democratic system of governance. What competing private interest does Diebold have in keeping the schema of the very possibly malfunctioning vote storing database (that they SOLD to Alaskans for a hefty sum) completely hidden from even limited and competitively protected examination?

Your mindset and that of several others on this thread is completely indefensible. We live in a country where the individual rights and separation of powers guaranteed in our Constitution can be arbitrarily usurped in the name of "keeping us safe." We live in a country where the Supreme Court just ruled that local governments can arbitrarily steal single family homes from their owners in order to satisfy those who lobby and contribute to them. But somehow a supposed contract -- that's probably unconstitutional and certainly without a shred merit in terms of public policy -- mistakenly awarded to a controversial corporation somehow grants this corporation a divine and unquestionable right to mismanage public election results in secrecy?

[kster]

171. This thread explains the secret vote counting controversy

to most people, even an eight year old would be suspicious of this proprietary bullsh*t. This is laughable, and to think some people aren't outraged about it, is just shocking, and then they try and explain the proprietary bullsh*t away, on top of it? I have to say, if I didn't believe that they were counting votes in secret this thread would turn me in to a believer.

[OnTheOtherHand]

173. thanks for sharing

Sure enough, this is exactly the same stupid argument I've gone through dozens of times on exit polls.

May I infer that (1) inspecting the .mdb file wold not reassure you; (2) you have no particular idea what might be in them that would evince fraud; and (3) you don't really care much what the contract says because you don't think the state had any right to enter into a contract that would protect the .mdb?

Would we have a "transparent and trustworthy voting system" if only Alaska turned over the .mdb? If not, what actually is your point?

You assume way too much about my "mindset," but discussing my mindset bores the hell out of me, so I will certainly spare everyone else.

[stickdog]

175. You may infer that NOT being able to inspect the mdb gives me

NO confidence in Diebold's voting systems. This statement does not imply its negation. In other words, in terms of gaining my confidence, having the right to inspect Diebold's GEMS Access database .mdb file is a necessary condition. However, it is not by any means a sufficient condition. You seem to be implying that a condition's necessity somehow depends on its sufficiency. Might I suggest a (refresher?) course in introductory logic?

I have all sorts of ideas about what might be in the Diebold's mdb database, and the more they indefensibly shield it from any and all examination, the more sinister these ideas become. Indefensible secrecy breeds untoward suspicions.

Certainly, I care about what the contact says -- primarily because I find it hard to believe that any group of people would be so moronic as to purchase a database product that didn't allow THEMSELVES to examine its design to confirm its integrity and accuracy. That said, it is clear that widespread vote totaling discrepancies are of compelling interest to both the general public and that a transparent and trustworthy voting system is an extremely important public interest in any putatively democratic system. Further, Diebold has no compelling private interest for keeping its GEMS schema completely hidden from even limited and competitively protected examination. Therefore, any clause in the contract prohibiting the database from being examined -- if, in fact, such a clause actually exists -- is not legally enforceable.

[Fredda Weinberg]

144. Should isn't the issue here

I've written several times what I would prefer - but this thread is about the Alaska 2004 election and the mdb file isn't public property.

[stickdog]

166. Should is EXACTLY the issue here. The state of Alaska SHOULD

allow the Alaskan Democratic Party to examine the mdb file. If Diebold has a problem with that, then Diebold SHOULD have to go to court and prove their dubious claim that their proprietary interests outweigh Alaskan confidence in Alaskan public elections.

[stickdog]

70. THE FACT THAT THE DATABASE IS BEING USED TO COUNT

VOTES makes it the citizenry's domain. If you disagree, please explain your justification. Do you really think DEMOCRACY is PROPRIETARY?

[stickdog]

69. It's a VOTE COUNTING database. What is your justification for

the lack of transparency? Don't you think the data structures our government uses to count votes should be transparent? If not, why not?

[Fredda Weinberg]

82. Should be? Certainly

but it ain't.

[stickdog]

97. And you're fine with that? Do you think democracy is proprietary?

[Fredda Weinberg]

112. You don't seem to be listening

Why don't you identify yourself and tell us what you've done to make the system better ... I have.

[galloglas]

121. A goddamned good reason

to be using hand counted, hand cast paper ballots!!

Is paper proprietary??????

[Fredda Weinberg]

131. New York hasn't had paper in a century

We're still using lever machines ... no paper trail and the busted ones invariably find their way into lower class neighborhoods.

You're not going to get a hand count in NYC - but it would be nice if we could use DREs to print paper ballots for our multi-lingual electorate and precinct level optical readers.

My point is simple: if there's a scandal to be uncovered in Alaska, let's get all the facts and act intelligently. Why don't see if you can get the district-by-district spreadsheet that the Democratic party there refers to and see whether or not the corrections were uniformly applied?

[stickdog]

155. Lever machines are hard to rig and very difficult to rig en masse.

You love this facile sort of comparison, don't you?

[Fredda Weinberg]

157. Man are you wrong

The fraud is systemic; the faulty machines are routinely placed in lower-class areas. But you don't seem to be concerned about the uncounted affidavit ballots, do you?

[stickdog]

160. Sure, there is systemic "fraud," just like there was with punch cards

vs. op scan ballots in Florida in 2000.

But that's not the same as fraud-o-matic DREs in which the individual machine counts can appear to be working perfectly -- and actually work perfectly during all testing stages -- at the same time that the actual election returns can be easily and globally compromised with simple hacks, worms and/or easter eggs.

[Fredda Weinberg]

161. You don't know what you're talking about, but that's ok

The Diebold system is not hooked up to the Internet - the precinct machines dial in to the accumulator. But I know this because I'VE BEEN PAYING ATTENTION.

To hack the system - and others have tried - you need privileged access: physical access to the machines.

[stickdog]

164. Do you even know what an easter egg is? Do you know what a worm is?

Do you realize that both the modems and smart card readers on these machines are unacceptable and unnecessary security risks?

[stickdog]

172. Access data is also easily manipulated and easily transferred

from one table or view to another related but not necessarily equivalent table or view. How can any observer determine anything about the accuracy of the database OR the authenticity of any specific data extract without examining all of the data structures, queries, report definitions, macros and modules contained in the Access database?

[kster]

56. This is not going to look good for Diebold

"The issue is that the (Democratic Party) is asking for a file format the state of Alaska uses but does not own." Diebold told the state it owns the format, which can't be released because it's a company secret"

[Fredda Weinberg]

83. Company property ... doesn't have to be secret but it helps

[kster]

78. OK, but

Do you think that excuse is a valid one? And should we be using these machines in public elections? If a Democrat or Republican wants to verify the election, I think they should be able and to, and if the company that builds the ballot counter does not agree I see no reason to use that company to count our votes. Plain and simple.

[Fredda Weinberg]

84. Government officials make this kind of decision every day

I'd do things differently but that's why I'm working where I do these days ... to make a positive contribution.

[galloglas]

122. Government officials make this kind of decision every day????

And that makes it OK?? Legal??

House Representatives were taking Abroamoff's payoffs everyday, also, but that does not even directly disenfranchise the public.

Why does our government, aided by MSM silence, investigate/prosecute one of situation, while ignoring the other??

[Fredda Weinberg]

132. Are you feigning outrage?

Of course governments sign contracts with private companies to deliver public services - I've worked at all nearly all levels (city, county, state) and haven't yet found one that does all or even most of their development in-house. And those contracts will, almost invariably, involve proprietary work products that the client must protect.

[galloglas]

139. Feigned Outrage ??

No, ma'am. The outrage is real.

You think because government outsources certain purchases (like toilet paper or pencil erasers) every day, or contracts for services with outside individuals or companies, (painting buildings or hauling trash) that you can equate those things with outsourcing MY FRANCHISE TO VOTE?.

I am sorry, ma'am. I am a citizen, I am not chattel! My government can fulfil its obligations as the citizenry permits, but it cannot fob off my vote to Diebold, or others, and then blame the catastrophe on the machine makers.

To Hell with that. I will fight it forever!

[Fredda Weinberg]

145. A state can ... and did

When I started doing software development for political entities, I was surprised by how little they did for themselves ... they barely can keep up with managing the resources they have. That's just the way it is.

Your franchise hasn't been outsourced - most election officials are elected and the rest are appointed by someone who stood for office. It may make you feel better to vent at me, but your energy would be better spent getting involved in your local political arena.

[stickdog]

174. Get off it, Fredda. Alaskan election officials are hiding what is probably

THEIR OWN incompetence behind their ludicrous claims that Diebold's dubious corporate interests outweigh Alaskans' interests in reliable and trustworthy elections.

If Diebold's database is actually well-designed and accurate, what possible harm will come to them by letting their own client -- you know, the citizens of the state of Alaska -- examine the vote counting database that THEY BOUGHT with THEIR tax dollars? This ain't a fucking operating system or even a unique enterprise application that we are talking about here -- it's nothing more than a trivially simple Access database (or at least it shouldn't be). Database vendors develop thousands upon thousands of similar database schemas every year, and only the tiniest fraction deny THEIR OWN CUSTOMERS the right to examine the back end of the database structures they purchased. It's like selling someone Windows XP but denying them Windows Explorer.

We are told we have to pee in a cup and hand it over on demand if we want many corporate or government jobs, and that if we have nothing to hide that we should have no problem with this. If Diebold has nothing to hide, then let's see the mdb file.

[Fredda Weinberg]

81. The data is public and was offered as an export

But under the laws we live by, Diebold has a right to protect their work product - the database structure.

It's not even Diebold that's making the claim ... the state licensed the software and is keeping their agreement. If they didn't, no respectable vendor would work with them again.

[Steve A Play]

102. False argument

The database format belongs to Microsoft (Access) and the scanner technology is licensed from Kodak.

I've seen the file format (the same file format Harri Hursti used to produce his VotOScope software) and there is absolutely nothing that anyone could learn from it that would violate Diebold's alleged 'intellectual property' rights.

The "State" has no inherent right to sign any agreement that violates the rights of the "people" of that "State". The right of the people to count the votes to determine who will legitimately be chosen by the majority of them is inherent to the design of any real Democracy.

Coming to a Federal courtroom near you soon! :)

BTW Fredda, no "respectable vendor" would hide the original intent of the voters behind a bogus claim of "intellectual property rights" and no state should ever work with such a vendor in the first place! :crazy:

Trust me, we're counting on Diebold never working with our state again!

Steven P. :kick:

[Fredda Weinberg]

113. You don't know what a schema is, do you?

Access is an application that manages data. The relationships among the objects is a work product and is proprietary unless rights have been granted. Deciding how to split information into manageable pieces is an art I'm proud to have mastered.

Knowledge is power and you're disarming yourself.

[stickdog]

128. I'm a data architect. I know what a schema is.

Access is a toy database. It shouldn't be used to count votes.

More importantly, RIGHTS TO THE ENTIRE SCHEMA should have been freely granted. If Diebold has a problem with this, there are thousands of other vendors who would not.

[Fredda Weinberg]

133. Thousands of qualified vendors for elections systems?

Show me.

It may be your professional opinion that Access is not a suitable DBMS for a vote tabulator - I disagree. It's a desktop application that simply aggregates data; Oracle or even SQL Server would be overkill and overly expensive. I've used it myself and only upsized when performance was a consideration or multiple users were involved.

This entire discussion has degraded into "should have" or "should be" ... but it began with a serious charge of election fraud and I haven't seen anyone so far willing to do anything to strengthen the case.

Greg Palast's initial charge, that Florida had used its central voter scrub to disenfranchise voters, got no traction until I got the raw files and ran a statistical analysis. As a professional yourself, surely you can see the reason behind my request.

Why don't you join me in asking for the district-by-district spreadsheet from the Democratic party? I can guarantee you that if the numbers back up the accusation, we will be heard.

[stickdog]

150. Come on. I could create custom Access vote counting by myself

for less than 100 billable hours.

1) Why would you want to use Oracle or SQL server when MySQL and PostgreSQL are more robust, more secure and free of charge?

2) The only reason there aren't more vendors with off-the-shelf vote tabulation products is because 99.9% of the thousands of vendors who create far more complicated, robust and secure database applications DAILY realize that you don't sell applications like this based on their merit, but based on your lobbyist budget and your inside connections. Surely, being a professional yourself, you must realize that Diebold's product is nothing special.

3) I'm not arguing that election fraud occurred on any huge scale. I'm arguing that -- regardless of any questions -- the public (and certainly the DEMOCRATIC PARTY upon their request) should be able to freely examine the entire database schema used to store the voting information of any public election.

[Fredda Weinberg]

158. Come on yourself

Every government I've worked for has insisted on Microsoft or Oracle - for the support and whatever other reasons they choose. Are connections important? Of course ... I learned that lesson a long time ago. But having learned it, I no longer rail against the system - it's exhausting and accomplishes nothing.

If you want to contribute something constructive, start a dialogue with someone like Professor Rubin, who's just gotta a ton of funding to work on this issue. Try to convince him that every component of a voting system should be in the public domain. I'll be curious as to his response.

[stickdog]

170. Fine. You can get a single user Oracle license for about $10,000.

I'm sure that's a lot less that one-tenth of what Diebold spent lobbying Alaskan politicians.

I'm sure Professor Rubin already knows that the public's interest in a transparent, accurate and secure voting system completely outweighs Diebold's interest in making sure that nobody checks under the hood of their insecure and often inaccurate vote counting systems. But I sincerely question whether you do.

[galloglas]

124. Which would leave us with

the option to hand mark and hand count ballots at precinct level.

Sounds like a deal to me!

[Melissa G]

3. This is not about code... This is about vote padding! Did you not read the

article? This is a place where there is a paper trail and it does not tally with the machines.

[Fredda Weinberg]

6. Yes, I read it and the error can arise from two possible

sources and after following this story carefully since I first opened the VB code and examined the Access database - I can conclude you're beating a dead horse.

Did you catch the HAVA hearings where Avi Rubin of John Hopkins testified? His valid criticism is that Diebold didn't use the latest encryption.

It's been years since Bev Harris phoned me breathlessly with her discovery - if you can, the threads are probably still here where I pleaded desperately for reason to prevail.

Here's the official explanation - I won't bother reformatting

During a recount, all ballots are under further review. The Accu-Vote machines are set to return any ballot that is not
properly marked. As a result, the certified recount results were higher than those previously certified. “Voters may not
have properly filled in the oval when casting their ballot or failed to request a replacement ballot if they needed to make a
correction to their ballot. Election workers may have overlooked messages displayed on the Accu-Votes, or in hand count
precincts, failed to tally and record the results correctly. This is not new to the 2004 General Election, and the recount
proves that even with ‘human error’ and Accu-Vote equipment, Alaskans can expect honest election results,” said Glaiser.

http://www.gov.state.ak.us/ltgov/elections/recount04.pdf

[Wilms]

9. I'm not following...

snip

During a recount, all ballots are under further review. The Accu-Vote machines are set to return any ballot that is not properly marked. As a result, the certified recount results were higher than those previously certified.

snip

This seems to say the Accu-Vote machines are NOT set to return any ballot that is not properly marked during the initial count, only during the recount.

What am I missing?

Nor does the link address the presidential election result, unless that wasn't dealt with during the recount.

[Fredda Weinberg]

10. The human factor

"Election workers may have overlooked messages displayed on the Accu-Votes, or in hand count precincts, failed to tally and record the results correctly."

The document I found dealt with recounts ... but my Google search was "alaska Division of Elections murkowski discrepancy"

Give me a moment ... I'll try to respond to all your legitimate concerns.

[Wilms]

11. Thank you.

Because it's a considerable skew.

[Fredda Weinberg]

15. There was no presidential recount

but seriously, will we call it fraud if Bush *loses* votes?

[Wilms]

17. I might.

Some, myself included, considered the possibility that results could have been tipped toward Kerry in safe * states (or even safe Kerry states) to skew the exit polls and raise questions about their veracity.

[OnTheOtherHand]

25. OK, but by 100,000 votes in Alaska?

AFAICS it makes little more sense to posit that 100,000 votes were stolen from Bush there than that 100,000 votes were stolen from Kerry. And I still don't think I've ever gotten anyone to take a simple stab at my question, 'What do you think the true vote total might have been?'

Totally separate issue from what information should be released. The Alaska Dems say they aren't alleging election theft, but they want intelligible election returns -- seems totally reasonable to me.

[OnTheOtherHand]

27. oops, let me clarify that

"aren't alleging election theft" -- I knew what I meant, but it would be asking a lot for anyone else to figure it out!

The Democrats are not asserting that anyone hacked into the computers or that anyone who lost a race really should have won, Brown said.

"We are trying to determine how many votes each candidate got in each district, and we can't tell that from the public data," she said.

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=2331628&mesg_id=2331724

I don't know when that story was dated. Things have moved on somewhat since then, I gather. The Democrats apparently still aren't asserting that anyone who lost a race really should have won (although they are in no position to rule it out), but they are insisting on more than just a report of "how many votes each candidate got in each district." So it turns into a quarrel over the database. And they are expressing doubts about the "accuracy of the 2004 General Election vote results," which certainly sounds like it could have some bearing on election theft.

[Wilms]

33. Not sure I followed all of that...

But I'd say 100,000 votes would be an awfully high number to try and mess with AND get away with it. So I don't really suspect that.

But until we know what caused this, we don't know. Diebold OpScans had tally trouble during a PA test.

I have know way of guessing the actual tally.

What I'm hoping for is an exit poll guru to tell us if Kerry and/or Bush tallies were outside the Margin of Error for the exit poll.

[Febble]

36. Looks like it.

the mean WPE was -9.6 according to the EM report, and the Best Geo Estimate was -9.4, so it was larger than the average.

The average WPE was -6.5, although WPE is a poor measure to average.

But the exit polls more redshifted than most, yes. But not as much as:

Connecticut
Delaware
Vermont
New Hampshire
Alabama
California
Mississipi
New York
North Carolina
Ohio

Connecticut wins. Any takers for massive fraud in Connecticut?

[Wilms]

41. Can you help me understand...

what WPE means

if the -9.6 was favoring Kerry (and if so, isn't that called "blue shift")

What Geo is, but at a .2 difference with WPE I probably don't need to be concerned

Finally, what would help me most is to know is the % of the tally and % of the exit poll for both Kerry and the boy wonder.

[Febble]

61. A negative WPE

is one in which the poll gave Kerry a bigger voteshare than the count. It got called "redshift" (don't know who coined that one, but it's useful - might have been Freeman) because the count was "redshifted" relative to expectations based on the poll.

OTOH can probably explain GEO better, but essentially it's the estimate made from the exit poll data alone (after adjustment for age/race/sex proportions of respondents vs non-respondents) and, I think some factors to do with the geographic selection of precincts.

According the the E-M report, the Geo estimate for Bush's voteshare in Alaska was 57.4%, and the final margin reported is -25.6, giving his vote share as 62.8%. But that ignores third party, and CNN

http://www.cnn.com/ELECTION/2004/pages/results/states/AK/P/00/

gives his share of the total vote as 61%.

But, ballpark, the exit poll underestimated Bush's share of the vote by quite a bit. Translated into votes, 57.4 % of the total vote (according to CNN) in Alaska is 178,978, CNN gives Bush 190,889. So, if you believe the exit polls ;) someone stole 11,911 votes, if I've done that right. Or switched 5,956


And the exit polls weren't likely to be that far off due to sampling error. But they could have been off for other reasons, and frankly there weren't rich pickings in Alaska. I wouldn't have thought it would have been worth the risk.

Better to steal votes in Connecticut.

[OnTheOtherHand]

63. no, not quite

You're halving where you shouldn't, I think. The error is close to 10 points on the margin, or about 30,000 votes "stolen" or 15,000 votes "switched" or some muddle in between. (15,000 is high, your 12,000 is low, I can't be bothered to figure it more precisely -- it's the third-party voters who muddle the percentages.)

It looks as if the exit poll discrepancy is similarly large in all three geostrata -- "Anchorage," "Anchorage/Fairbanks Corridor," and "Juneau/Fairbanks/Rural." So we probably don't have to fret much about what was going on out in the bush.

[Febble]

64. Sorry

Hoped you'd chip in. I find those geo tables totally confusing, ditto third party issues.

[OnTheOtherHand]

38. fair enough

I don't see how tally trouble could account for internal discrepancies within the election reports. Beyond that, I wouldna say much.

As Febble says, the Alaska exit poll is out of line, but less so than many. Based on the pre-election poll, the exit poll result seems more bizarre than the official result -- but the pre-election poll for Alaska is also very stale (from mid-September, I think), so without any first-hand knowledge of the Alaska political mood, I wouldn't pound the table on that point.

[Wilms]

40. Yes. The last thing I'd want to see is exit poll pie-fights about AK. n/t

[OnTheOtherHand]

65. the pies would be hard as rocks (unless we wait for summer) n/t

[stickdog]

74. Of course. Fraud is fraud.

Somehow I'd assumed you were DU's resident expert on exactly this subject ...

[Fredda Weinberg]

85. What I suspect is that the officials are telling the truth

and if the Democratic party published all the numbers, you'd find the results I calculated for the presidential contest were universal corrections and that all candidates had their totals reduced by the same proportion. My email to the party went unanswered ... are you willing to make the same request and put their assertion to a critical test?

[EFerrari]

23. Psuedo logic.

You say there are two possible sources of error and neglect to name them.

The Accu-Vote machines may be set up to return any ballot that isn't properly marked but the recount is further contaminated by the recounters. Not to mention, the "official" explanation blames the voter for a system deficiency.

And, lastly, you have neglected to answer my question, posed to you in multiple threads: why are you continually discounting election fraud even in the face of mountainous evidence and why are you smearing Mark Crispin Miller on this board?

[Fredda Weinberg]

51. 2 sources? Machine or man

Machines ain't infallible either, but I've always appreciated their dispassionate response.

I have been patiently watching for years to see evidence of deliberate fraud. Since my mother and I hosted the producers of "Unprecedented" in our home, I've been getting the word out about the vulnerabilities of DREs - and made a difference.

As for Professor Miller, I stand by what I've already related - call it what you will, but the fact remains that I was the one the Donahue show contacted to get a spokeman on the air and was mortified by his behavior.

[stickdog]

75. All machine counted voting systems are vulnerable without a

a full random system of manually counted audits.

Fraud-o-matic DREs are more vulnerable simply because voting into thin air makes massive, statewide fraud as easy as a click of a mouse. Are you at least clear on this concept?

[Fredda Weinberg]

86. You're not paying attention

When I had the producers of "Unprecedented" in my home, I begged them to capture LePore's demonstration of DREs in Palm Beach County ... but they were mesmerized by the experience and blew it. Ah well ...

So I kept up my local press contacts and when the Wellington debacle happened, the technology was exposed for the mistake it was and the county commission finally noticed.

As for audits, they're already part of the process - go look it up.

[stickdog]

103. I don't know WTF you are talking about. What is "Unprecedented"?

Who blew what? What's the Wellington debacle? And, most importantly, what audits are part of what process?

[Fredda Weinberg]

114. I'm tempted ...

But it would violate the board's rules. Go view

http://www.unprecedented.org/

and you'll see the right way to bring an important issue to the public's attention.

Go do a search on "wellington lepore" if you want to know more on that subject.

And check the Alaska's Division of Elections site yourself if you don't know the existing procedures. I'm not your private tutor.

[stickdog]

72. If it's fully secure and completely wonderful THEN WHY THE

FUCK CAN'T WE EXAMINE IT???

[Fredda Weinberg]

87. I suspect you want an issue - not the truth

but I'll wait and see if we can get the data from the AK Dems who are making the accusation - what they published so far is only, at best, a half truth.

I asked for the entire spreadsheet - why don't you add your request to mine?

[stickdog]

104. I'm not an Alaskan. What I can't understand is why Alaskans can't

examine the database used to count their votes. That makes absolutely NO SENSE, and you've yet to address this critical issue other than to say "that's just the way it is."

[Name removed]

116. Deleted message

Message removed by moderator. Click here to review the message board rules.

[galloglas]

127. Why don't you

stop speculating on someone's motives and answer the goddamned question put to you?

[Fredda Weinberg]

134. Because you've descended into vulgarities and have ceased to make sense

I've patiently responded to every substantial response - your obscenity laced posts simply repeat irrational requests. Once again, proprietary rights are involved and legal obligations are being observed. A supervised extract of the data would provide anyone with the same information as the original file - but this offer was refused.

[galloglas]

140. Obscenity laced?

I hope you are referring to someone else. I used the word "Goddamned". Others used much worse.

Is your offense at my demanding an answer? Why/how can you use the concept of "proprietary interests" by corporations as a reason to withhold information which is in the public interest?

Figures from Alaska are one thing, but your saying that a corporation's rights to proprietary ownership supercedes my right to exercise my right to vote, and to have that vote honestly counted, is ludicrous.

Unless, of course, you are unconsciously mimicking Bob Dylan, saying,

"But the vote clerk said, "It happens every day",
As the stars fell down and the fields burned away
On Black Diamond Bay"

[Fredda Weinberg]

146. I'm telling it like it is

You don't like it - so what? Neither do I, but I'm calmer - maybe because I've seen positive changes in response to my efforts.

[galloglas]

154. I am happy

that you percieve positive changes in response to whatever efforts you have exerted. That gives you self-contained fulfillment.

However, that has nothing to do with the governmental appropriation of my voting franchise and the subletting of the government's obligation to me and my fellow citizens to corporate cronies.

So, I guess if you're happy, you must be one of them?

[Fredda Weinberg]

159. What government appropriation?

This accusation of fraud has descended into the proposition that all voting systems, regardless of contract terms, should be appropriated into the public domain. Sounds like a subplot in an Ayn Rand novel ...

"One of them"? Who was the last simpleton who raised that canard? Wait ... it's coming to me ... you don't happen to live in the White House, do you?

[galloglas]

163. No, as a matter of fact,

my ethics would prevent that. How about you?

You seem to have no moral trouble with a government agency subcontracting my voting rights to a corporate entity. By extension of your logic, would you allow the protection of my 4th and 5th Amendment rights to be subcontracted? To whom, Blackwater?

Hedda, I thought you were just "hard-Hedda-ed". But your reference to Ayn Rand makes me understand I was wrong.

Your statement,
"the proposition that all voting systems, regardless of contract terms, should be appropriated into the public domain"

assumes the alternative,

"that that all (or some) voting systems, regardless of contract terms, should be appropriated into the private domain".

So, two questions.

1) "Quo warranto??"

2) Did the Founding Fathers come back and write that into the Constitution? When did our electoral rights get transferred to "contractual agreements" with corporate entities?

Hell, with the power you concede to the corporations, they should just vote for us, tell us who won, and then we could all feel patriotic from saving the gasoline to get to the polls.

Even better, why not just have the government auction off our bodies and souls to these companies. Then there would be no reason for us to vote.

After all, we had slavery before, didn't we? Made America great, huh?

[Fredda Weinberg]

7. Sure it's about code

There's a link to the Diebold software ... if you know how, examine it yourself.

[Melissa G]

12. Okay , I'll wait for your response about the skew... Here is a link to

the original DU Alaksa article thread...
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x2331628

[Fredda Weinberg]

13. This explains the 200% of voters casting ballots

In House Districts 16 – 32 some precincts show more than 100 percent voter turnout. This is because the State of Alaska conducted a special election for the Municipality of Anchorage concurrently with the 2004 General Election. In these districts, there were two ballots, a State ballot and a Municipality ballot. The % Turnout is based on the Cards Cast number. Because each voter could have cast two ballots the turnout in these districts is inflated. In these districts look at the Times Counted number for the US Presidential race to determine the actual number of State ballots cast in each precinct. This number, divided by the number of registered voters will yield the turnout for the precinct:

http://www.ltgov.state.ak.us/elections/04genr/index.shtml

[Wilms]

16. The turnout statistics aside for the moment...

...I'm still trying to figure out the tally discrepancies.

Can you point me to it?

[Fredda Weinberg]

18. Not this late

but we can take this up again tomorrow evening. If necessary, I'll be happy to write 'em and ask.

So for now, good evening and thank you for a reasonable discussion of a serious subject.

[Wilms]

19. And thank you.

Indeed, come on back and we'll sort it out.

[Fredda Weinberg]

35. I think I've got a handle on it, but I'll ask if you're not satisfied

First off all, the unofficial numbers are all over the map, so I'm curious who did the precinct totals (I'm not willing to plug the numbers from the 40 precincts into my own spreadsheet). Here's what I came across

NYT 151,876 61.8%

http://www.nytimes.com/ref/elections2004/2004AK.html

CNN 190,889 61%

http://www.cnn.com/ELECTION/2004/pages/results/states/AK/P/00/index.html

WP 142,087 62%

http://www.washingtonpost.com/wp-srv/elections/2004/ak/prescounties/

WP 188,943 61

http://www.washingtonpost.com/wp-srv/elections/2004/ak/

But what I did a quick and dirty calculation that I think explains the presidential discrepancy ... divide the number of ballots lost in the Senatorial race by the initial total - it comes to about .34. Do the same for the presidential race ... also .34. That suggests to me that the explanation proferred covers both - there were dual ballots that had to be subtracted before the official tally could be reached.

[Wilms]

39. Wow. those are all over the map.

I wonder if it's unusually so.


I'm not following the equation you laid out.

And your conclusion, "there were dual ballots that had to be subtracted before the official tally could be reached", doesn't ring right, either.

You may, again, be confusing the turnout issue with that of the tally.

[Fredda Weinberg]

43. Mebbe, and I'm willing to ask an official but

where can I get the precinct tally you quote? The official site keeps 'em by pdf per precinct.

[Wilms]

45. I got the #'s off one of the stories that ran about this.

[Fredda Weinberg]

48. That's what I was afraid of

Professionally, I crunch gigabytes of data - that's why I had no problem running the statistical analysis on the Florida scrub files in 2001. But I'm not going to commit to an evening of parsing pdf files.

I spent the better part of two hours tonight reading news reports and haven't seen this anomaly raised as a serious issue - not even in the alternative media.

I got nothing against VoteTrustUSA ... though I think their discussion on interpreted code is bogus

http://www.votetrustusa.org/index.php?option=com_content&task=view&id=645&Itemid=26

It's a technical discussion I'll have if you want, but here's my offer: if you want, I'll ask 'em for the spreadsheet they used for their article, check it for accuracy and if it matches the published numbers, take it up with the Alaskan officials. How's that?

[Wilms]

52. Post a thread with their reply.

[Fredda Weinberg]

54. Inquiry sent ... will start a thread if I get a response

but after looking at the article again I'm not so sure. It says

... the district-by-district posted results do not show how many votes were cast for each candidate in the 40 House districts ...

so I'm curious how they got their "district-by-district totals"

And their problem with:

According to the posted “Statement of Votes Cast” by district and precinct, a far larger number of votes were cast than the official totals reported in the statewide summary. In many of the House Districts, more votes are shown in the totals than there are voters in the district.

was really easy to explain if they'd looked at the recount page or bothered to ask. But we'll see ...

[stickdog]

77. So you don't trust the media to report the numbers correctly, but you

do trust the media to report the seriousness of the issue correctly?

Why again can't we have the database file to see for ourselves?

[Fredda Weinberg]

88. The media is publishing the Democratic party's numbers

accurately, but the letter they've made public doesn't contain Kerry's totals - I suspect that if they did, we'd find that all the candidates were corrected uniformly and there's nothing here but an average election.

[stickdog]

98. Which brings us to the point -- if everything's fine, why can't we see

the database?

[Fredda Weinberg]

147. And my original point ... because the state signed a contract

that committed them to protect Diebold's proprietary work product. The data is public - the database is not.

[kster]

149. Which is why States should stop buying these machines

immediately.

[stickdog]

151. If the state signed such a bs and probably unconstitutional contract,

then the state needs to tear it up or at least modify it.

[stickdog]

76. And this is a GOOD reason NOT to examine the data structures

further?

[Fredda Weinberg]

89. The structure shouldn't be the issue

You can get a supervised export of all the data from all the tables - which has been offered. Insisting on the proprietary work product seems to be a way to keep this dead horse running.

I'd think differently if my query to the party had gotten a response. Why don't you try?

[stickdog]

99. Don't you think that the database that COUNTS Alaskans' votes

should be open to examination by the citizens of Alaska? If not, why not?

[Fredda Weinberg]

115. The database doesn't count anything

and a supervised export would satisfy any honest critics concerns - which is why this issue is bogus.

[stickdog]

129. How do you know the database doesn't count anything unless

you examine it? You keep avoiding the point and presenting fraudulent "justifications," Fredda. I'm a data architect, and I've developed dozens of full scale Access projects -- typically for small clients who couldn't afford decent databases. An Access database project typically includes far more than table definitions, and the public should have full rights to examine all the data structures, VB Code modules, queries, forms, reports and macros in the entire database.

[Fredda Weinberg]

135. You've never protected your source code?

I find that incredible, although it happens to be my practice.

I haven't said anything fraudulent yet, but find it amusing that you can claim in one post that Access is a "toy" (reply #128) and here boast that you've developed dozens of Access projects yourself. No wonder you don't reveal your true identity!

The presence of code isn't relevant here - an extract would allow a comparison between the ballots cast and the totals, which is the point.

Once again, you're demanding what "should be" ... but the fact remains that a contract exists. So if the point is to uncover fraud, why not accept a supervised data extract and begin a serious examination?

[stickdog]

152. Sure, I've "protected" my source code. What does that

have to do with an Access mdb file?

Do you really protect your mdb files from your own clients? How and why?

If you need to protect your VB Code, Fredda, I have a rudimentary tip for you. Don't make it part of an Access project.

[Fredda Weinberg]

162. My schemas are as valuable as my source code

Databases are one of my specialties. But I have a personal policy against closed systems - I always include source in my deliverables and relational diagrams in my documentation.

On the other hand, I'm not a corporation that needs to protect its intellectual property to survive. I recognize that IT wouldn't get far if we relied exclusively on open source products. And I know officials are reluctant to work with anyone who isn't large enough to support them with dedicated staff.

I'm not responsible for the way things are but am sanguine because I've got my long-term plan to address this issue. But as far as this "scandal" in Alaska goes, I think at this point it's bogus - and if you're not willing to verify the validity of the charge by requesting the spreadsheet the Democratic party refers to - you're not worth another keystroke.

[stickdog]

165. How can one guarantee that any specific spreadsheet is accurate

without inspecting the entire database?

And of course you have personal policy against closed systems. So do 99% of educated buyers. Funny how our state governments somehow aren't included in the number, isn't it?

[alittlelark]

5. Uh, yeah...............

[stickdog]

24. So that's why we can't review this public information?

Because YOU have concluded releasing this public data for public inspection isn't important and -- like Bush -- are enjoining us to trust YOU?

[Fredda Weinberg]

37. Nuts ... there's plenty of data

and I didn't ask you to trust me - Professor Avi Rubin testified before the HAVA commission and the only thing he wanted to do was tighten the encryption standard. He didn't even complain about the Access database.

You call Diebold's internal tables "public data" but that doesn't make it so. I wish things were different, but we live in a proprietary world and schemas are intellectual property.

[Wilms]

42. Where are you coming from? n/t

[Fredda Weinberg]

44. I'm afraid I don't understand the question

[stickdog]

68. DEMOCRACY IS NOT PROPRIETARY!!!

It's a database that's being used to count our votes. So the fuck is your justification for keeping it secret? Don't you think the data structures our government uses to count our votes should be transparent? If not, why not?

[kster]

80. "

:thumbsup:

[eomer]

79. Professor Rubin did a study on Diebold touchscreen software...

and complained not just about the encryption standard but rather about a very long list of security flaws, just one of which was the encryption standard. Here is the report.

MS Access is not used in the touchscreen implementation that Rubin studied -- I believe it's used in the GEMS tabulation software. If Rubin didn't mention it in his public testimony my guess is it's because he hadn't studied GEMS.

I did a decent amount of searching trying to find Rubin making a comment about GEMS and couldn't find it. If I'm wrong and Rubin has commented specifically about GEMS and, in that context, was only worried about encryption, can you point me to where I can read something about that?

[Fredda Weinberg]

91. Talk about apples and oranges

What you linked to wasn't Rubin's testimony before HAVA. If you weren't concerned enough about the topic to watch the hearings, then don't blame me.

I've reviewed the paper and it's consistent with what I've stated so far - he found the system vulnerable but nothing nefarious.

[stickdog]

100. THEN WHY CAN'T WE SEE THE DATABASE???

You keep avoiding the point. This is a the database used to count the votes in a PUBLIC ELECTION! Do you or do you not agree that the public should have the right to examine it? If you don't, please explain.

[eomer]

109. I found a transcript of Rubin's testimony 5/5/2004 before the EAC.

Is that the testimony you're talking about? Here is the link.

Here's an excerpt:

<snip>

My primary concerns with today’s DREs are:
- There is no way for voters to verify that their votes were recorded correctly.
- There is no way to publicly count the votes.
- In the case of a controversial election, meaningful recounts are impossible.
- The machines must be completely trusted. They must be trusted not to fail, not to
have been programmed maliciously, and not to have been tampered with at any point
prior to or during the election. We have techniques for building secure systems, and
they are not being utilized.
- With respect to the Diebold Accuvote TS and TSx, we found gross design and
programming errors, as outlined in our attached report. The current certification
process resulted in these machines being approved for use and being used in
elections.
- We do not know if the machines from other vendors are as bad as the Diebold ones
because they have not made their systems available for analysis.

Since our study came out, three other major studies often referred to as the SAIC report, the Ohio
reports, and the RABA report, all cited serious security vulnerabilities in DREs. RABA, which is
closely allied with the National Security Agency, called for a “pervasive rewrite” of Diebold’s
code. Yet, the vendors, and many election officials, such as those in Maryland and Georgia
continue to insist that the machines are perfectly secure. I cannot fathom the basis for their
claims. I do not know of a single computer security expert who would testify that these machines
are secure. I personally know dozens of computer security experts who would testify that they are
not.

I have been disappointed that the policy community did not reach out to the computer security
community when making decisions about voting technology, and when my community came to
the table, they said it was too late. At first I was puzzled by the lack of attention to the security
critiques of DREs. Today I am outraged. At this point the failures of current DREs have been
documented in four major studies by leading computer security experts, and we have ample field
experience documenting failures at the polling place. Yet computer security experts, myself
included, find ourselves routinely referred to as luddites and conspiracy theorists. Failing to
confer with computer security experts in decisions about voting technology was a mistake. Given
the gravity of the security failings the computer security community has documented in current
DRE systems it is irresponsible to move forward without addressing them.

Addressing the problems I and others have documented with DREs requires more than just fixing
the machines. We must reform the process for establishing voting technology to provide
transparency. Vendors are not subject to public code review. In the one instance where
independent security experts had an opportunity to examine a voting system, the results proved
that the current process results in machines being deployed with unacceptable lack of quality
control. We cannot achieve perfectly secure systems; such things do not exist. But on the
spectrum of terrible to very good, we are sitting at terrible. Not only have the vendors not
implemented security safeguards that are possible, they have not even correctly implemented the
ones that are easy.

<snip>

He says he is "outraged", quotes others as calling for "a pervasive rewrite" and uses the phrase "gross design and programming errors". Your implication that he didn't really find too much wrong other than the encryption standard is a gross mischaracterization.

And his testimony is, as I said, about Diebold touchscreens -- not about Diebold GEMS. That would be why he doesn't mention Access. Access is used in GEMS, not in the touchscreens.

[OnTheOtherHand]

110. but AFAICT Alaska didn't vote on DREs

Mostly op-scan, some hand-counted paper.

(I have no idea what testimony Fredda is referring to.)

[Fredda Weinberg]

118. The HAVA hearings were covered by C-Span

I was curious what Professor Rubin would have to say and particularly, because he was sitting next to another professor who'd gotten all sorts of grief in Georgia - another hyped "scandal" that proved less than it appeared.

[Fredda Weinberg]

117. Nope, you didn't watch him

and just posted his written statement, which was entered into the record. I cared enough to see the whole thing.

[eomer]

119. Actually I did watch that hearing.

But that has nothing to do with the question of what Rubin thinks about election software.

His written statement is, if anything, an even better indication of what he thinks than his spoken remarks since he was able to consider carefully exactly what he wanted to say and how he wanted to say it. The statement shows clearly that he is not concerned only about the encryption standard. He has numerous other concerns about the touchscreen software that he reviewed.

[Fredda Weinberg]

120. Any mention of malicious code?

Nope, just vulnerabilities, which is what I reported. The recent experiment showed that hackers needed privileged access to manipulate the vote count, which is what Rubin already observed.

So I've made my point and you've substantiated it. Thank you.

[eomer]

123. Agreed that Rubin didn't say he had found any nefarious code --

he just found gross design and programming errors that left the touchscreen system totally unreliable.

That has nothing to do with the issue being discussed here since these Alaska votes were not on Diebold touchscreens.

[stickdog]

130. Who cares if Rubin found malicious code? The Alaskan Democratic

Party wants to see the database used to store the voting information. Why can't they?

[Fredda Weinberg]

136. Answered a dozen times

You won't accept it, but that's beyond my control.

[stickdog]

153. What I won't accept is your blithe acceptance of bs proprietary claims

on one of our most fundamental public rights.

[alittlelark]

4. deleted - response not to OP.

[FogerRox]

14. Violation of OPRA ? Sue the fookers. like yesterday

[alittlelark]

20. R U a funny sort a fellow?

:evilgrin:

[farmbo]

21. I wonder if Alaska's highways are property of their asphalt contractors?

Unless there is a specific statute exempting Diebold from the Alaska Public Records Law, this statement is BS:

"The Division of Elections claims that its electronic computer file that contains all the final vote tallies for the 2004 General Election is proprietary information belonging to its contractor, Diebold Election Systems."

The software needs to be open for public inspection by qualified experts. If the ALL POWERFUL DIEBOLD feels it needs additional protection, the inspections can be conducted "in Camera"...under a Judge's supervision.

Trust But Verify.

[Blue_In_AK]

26. Thanks for posting this.

Here are snips from this morning's Anchorage Daily News article. I'll post the link, but you may need a subscription to read it.

http://www.adn.com/news/alaska/story/7386582p-7298824c.html
<snip>

"Basically what they say is they want to give us a printout from the (electronic) file. They don't want to give us the file itself. It doesn't enable us to get to the bottom of what we need to know," said Kay Brown, spokeswoman for the party.

<snip>

"The issue is not about whether public information can be released, because the Division of Elections has already offered to provide the information requested by the (Alaska Democratic Party)," elections director Whitney Brewster said in a written statement. "The issue is that the (Democratic Party) is asking for a file format the state of Alaska uses but does not own." Diebold told the state it owns the format, which can't be released because it's a company secret.

<snip>

Questions still hound the company. Some elections officials in other states are questioning whether its electronic machines are secure. Investors have sued the Ohio-based parent company, Diebold Inc., over whether it concealed problems with its voting machines, among other issues. Its chief executive, who once vowed to deliver Ohio electoral votes to President Bush, recently stepped down. The latest controversy concerns the database holding the results of Alaska's 2004 general election. Democrats say it's important for them to see the database in its original structure ---- the format in which the data was created and now is stored and reported. That's how they hope to figure out if the votes were registered and reported accurately.

Documents provided by the Democrats show that Brewster contacted Diebold and was told the public data can be released only after being transferred to a common format such as Microsoft Excel. In a Jan. 6 e-mail, Diebold's lawyer, Charles R. Owen, told Brewster that "the structure of the database file ... is proprietary information."

<snip>

"The results from the 2004 election in Alaska just plain look squirrelly," March said. For instance, district-by-district vote totals add up to 292,267 votes for President Bush, but his official total was only 190,889.

[riqster]

28. Puh-Leeeze

All one would have to do is export the data to another db type like Access, Oracle, compare it with the numbers in the State's system, and you have a valid analysis. If the numbers do not agree, then you perform an RCA to see why.

And if the root cause is in Diebold's system, or if it is in the export routine, or if it shows fraud, proceed accordingly. Quite frankly, I'd rather NOT use Diebold's app to audit itself.

IIRC, Diebold's database is bog-standard if a bit kludgy. Should be a piece of cake to get an export.

[Blue_In_AK]

29. I don't know what any of that means...

but I'll take your word for it. :)

[riqster]

30. Sorry for the geekspeak

What I said boils down to, get the data. Then, safely outside of Dieboldland, you can examine it and see what's up.


To use Diebold's own tool to validate Diebold data is a shortcut, and an inadvisable one at that.

[OnTheOtherHand]

31. but I thought I had read

-- perhaps from following the link in the OP? -- the director of the Election Board saying that s/he had already offered a data export (I think the story said 'something like Excel,' groan), but the Alaska Democratic Party had insisted on the Diebold format.

My first thought was much the same as yours: there shouldn't be any Special Magic in the Diebold format that accounts for numbers adding up or not adding up.

[riqster]

34. Actually, there may be something in the front end

and if we get raw data, we can get around that. Worth a shot.

[roseBudd]

32. Alaska is the state that Cheryl Belluci of Triad Elections bragged about

A Better Way?
This response submitted by Cheryl B. on 5/25/99.


Hmmm... I've been reading this thread about numbers on ballots with interest. You see, I work for a company which counts ballots for county boards of elections in five different states. In fact, I got to count ballots in an election in Alaska last month... uh, maybe I shouldn't bring that up.

The problem with ballots is that they must be unique, that's where the numbers come into play.

With the system we use, the ballots have unique stub numbers that are assigned to each voter. Those stubs are torn off and at that point the ballot is "disassociated" with the number... BUT, that is done at the polling place where someone has verified your identity. That is very hard to do when mailing comes in to play.

[Melissa G]

50. Kickin for Wilms and Fredda to finish their discussion.....

and some pretty 50 post flames..

[riqster]

66. One thing to keep in mind

Is that due to the privatization of a public trust, the citizens cannot view their own voting information. In effect, Diebold is holding this data hostage, with the blessing of the government.

[Melissa G]

67. Exactly, Private Companies using Private Code to count our votes

Our democracy being sold and i expect in some cases outsouced..
Insert Funeral Music here... Time to bring back Transparent Verifiable Elections...
Time to get rid of the Monarchy (aka imperial pResidency)again...

[Bill Bored]

90. This thread is bizarre!

Privatized elections, GEMS, someone named Fredda, double sets of books, EXIT POLLS?!?!?!?!?!?!?!!?

Let's get real.

Look, if Fredda knows what she's talking about, and I suspect she might, I think she's implying that this is a simple case of Op Scanners with the Ballot Reject Settings turned off in GEMS. This causes all ballots with mistakes on them NOT to be counted and the voters don't get a chance to correct them on every optical scanner in the jurisdiction, or perhaps a subset of scanners.

Is this it or am I missing something?

Now, if the AK Dems want to view the Op Scan settings to see if the reject settings were On (they ARE in fact turned OFF by default in GEMS, which by default will turn them OFF on every Op Scanner in the jurisdiction) then they should have a right to do this some way somehow without violating intellectual property rights. The question is how to do it under AK law.

Finally, what did the recount of the paper ballots show? Were the paper ballots used as the official count? If so, then there is no case for election fraud at the moment unless ballots mysteriously appeared or vanished. However, it should be possible to have some quality control procedure in place to see that the ballot definitions and ballot reject settings for the scanners are set correctly and Diebold should be forced to prove this to ALL parties in the election before it even begins.

What am I missing here? (And please make no references to exit polls in any response to me in this thread. Thanks.)

[Fredda Weinberg]

92. There was no manual recount

Here's a description of the recount done in the senatorial race.

http://www.gov.state.ak.us/ltgov/elections/recount04.pdf

As a check, I calculated how those numbers compared with the presidential figures as reported by the Democratic party. The adjustment was the same within 2 significant digits - close enough for the real world.

If I could have gotten the Kerry figures, as I requested from the party last night, I'm pretty sure that I would find the same ratio ... proving that there was no fraud.

[Bill Bored]

94. This is even more bizarre than I thought.

They did a MACHINE Count to verify a HAND Count?

[Fredda Weinberg]

95. 2 machine counts

[Bill Bored]

96. OK, but how do we know the machines were programmed the same way

for the recount as they were for the election?

All this recount was, was a test of a scanner or scanners with a test deck that happened to be ballots from an actual election. And the scanners came very close to the hand count, right? So what does that prove? That you can program scanners to count votes? How were the originally-scanned ballots recounted?

I'm still missing a lot of information here. What was the reason for the 100,000-vote discrepancy in the Presidential race between district totals and the statewide totals?

Same question for the 77,000-vote discrepancy in the Senate race?

And thanks for not mentioning the exit polls! :)

[Fredda Weinberg]

148. There happened to be a dual ballot election in most of the counties

I've posted this above and don't feel like searching through this 100 plus thread at this hour, but the elections officials explained the discrepancy for the recounted senate race and the presidential correction was the same proportionately. What I want to know is ... were the Kerry numbers adjusted the same way and in that case, there's no here here.

But I'm not willing to pick through 40 pdf files. The Democratic party claims to have a spreadsheet and I asked politely for a copy now two days ago without a reply. No one else here seems to be asking, so I guess they really don't care whether this is all a mistake by people who didn't bother to check the Division of Elections website for a plausible explanation.

[Bill Bored]

93. Oh, and if the Bev H./Howard D. GEMS hack is suspected,

all they need to do is a precinct canvass.

So what's the big deal?

[stickdog]

101. Democracy is NOT proprietary!

This database is being used to count votes in a public election. How in God's name can ANYONE argue for the supremacy of supposed "intellectual property rights" in this situation?

Are you really saying that you believe that "intellectual property rights" trump election transparency? If so, what bizarrely warped kind of public policy justification are you invoking? Unrestrained capitalism above democracy itself?

[Bill Bored]

106. I'm saying I don't get this story...yet.

I know the idea is not to privatize elections. I get that.

But they had paper ballots! If they want a recount, let them have it. Why do they need the database if they have the actual paper ballots? Who cares what Diebold said the vote count was? Not me. Just COUNT the paper ballots!

Now, if you want to prove that there was a programming problem, either deliberate or accidental, that's another story. Then you will need to see the programming. But as long as you have the voter-verified paper ballots, all you have to do to confirm the outcome of the elections, is to count them. And there's no reason for Diebold, or any other private company to have anything to do with that.

So, I'm sorry, but there's something about this story that still doesn't add up for me.

[stickdog]

107. Sure, count the ballots. AND look at the database. AND figure

out with total transparency where all the discrepancies came from. Why is that too much to ask?

[Bill Bored]

108. It's not. I frankly don't think intellectual property should come into

play here at all. Someone had to program GEMS. It seems it was Diebold in this case, but it could just as easily have been a Board of Elections employee. Looking at that should not be an issue. It's not source code.

The state is probably to blame for outsourcing their elections. They may have signed away their rights, and of course the argument can be made that the state had no right to do so and the contract is null and void.

[riqster]

111. Bingo

Democracy should NOT be outsourced. And it has been,which leads us to the question...is it still a democracy, when private corporations (the biggest of which are controlled by one family of RW extremists) are in charge of a public function?

I'd say not. This is a corporatocracy.

[Bill Bored]

125. I agree with all that, but then the fault lies with the state.

Real paper ballots do mitigate this somewhat however, so it's still a bit of a gray area in my mind in this particular case, unless I'm missing something.

It looks like a screw up in the results reporting between overlapping district lines. But there were also some discrepancies in the absentee count too.

Using the scanners to audit the hand counts is ludicrous although I guess since they agreed, it's not so bad. It certainly does NOT prove that the scanners worked properly where they were used for the first count.

So it's a complicated story.

If they want the .mdb file to check precinct totals, there are other ways of doing that (poll tapes), and that IS the actual precinct level vote count.

And of course, they can hand count those ballots to rule out the "Hursti hack" and even bad ballot definition programming.

So despite the presence of Diebold, I'm having a hard time seeing the lack of transparancy.

That said, the state should NOT have outsourced their elections.

[riqster]

126. Excessive complexity...

creates more places to hide things. Back in the day, people counted ballots. A recount involved doing it over in a more stringent fashion. Now, of course, there are multiple computers creating various files, stored in different formats and locations, all of which have to be brought together to ascertain accuracy. Quite difficult, and without understanding the counting process at a granular level (which is hidden within proprietary software now), you can't be sure that everything was done properly.

Let me try another analogy or two. Let's say, instead of a proper financial report, the treasurer of your state decided to put out a spreadsheet listing all of the dollars in your state budget, all of the ones and fives, tens and twenties, etc. Perhaps he or she will even list those bills by department. And you could get a final total dollar amount too, on another report. BUT, the accounting methods would be secret, so you wouldn't really know HOW the money was used nor for what reason or by whom, only that it existed at one time.

Or, another one. You own a quantity of gold coins that your forefathers passed down to you. They are stored in a safe-deposit box to which you cannot have access, since the box is on private property (the bank). The coins ARE yours, of course, but....

Neither of these examples are perfect, of course, but they point up the essential problem-the votes and voting process, which are the property of the citizenry, are not accessible to them. In days gone by, that was not the case. Now it is not possible to prove that the election was legitimate.

Oh, and the recount of ballots? Hand recounts are very rare, and IIRC one was not done in this case. Rather, the machines were used to recount the ballots, so whatever flaws existed in the original count process would exist the second time around. Hand recounts are expensive, so it takes a lot to get one done. Just because one CAN do a recount of paper ballots doesn;t mean one will be done.

And, of course, in Ohio, many thousands of ballots were never counted at all.

[Bill Bored]

142. It's a tough problem

but the bottom line is simply the vote counts.

In this case, it should be possible to reconstruct the election with poll tapes and paper ballots, as necessary, without the use of GEMS.

If this were a paperless DRE jurisdiction, that's another story, but even then, there may be ways of viewing GEMS without copying the files, by simply sitting in front of the GEMS PC for a day or so and viewing the configuration.

I agree that this makes Diebold look like a bunch of crooks, but that's not unusual, is it?

[riqster]

156. That would be simplest

In Ohio, if you want a manual recount, you have to petition for it, and pay for it. Plus, certain ballots are not available, being locked down by county and state officials. So, you can never get a perfect number.

Is that also the case in WA? How does one go about getting a manual recount?

[Melissa G]

141. Exit polls... Exit polls .....Exit polls....!!!!

okay, now that we have that out of the way and i have re included your favorite topic:evilgrin: I think it was a case of many people looking at describing an elephant which happens fairly frequently on this board. Also there are so many different levels of assumptions of knowledge and history... plus it is a weird conversation..

[Bill Bored]

143. I agree. See post 142. nt

[kster]

137. This would solve the proprietary problem

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=406688&mesg_id=406688

[riqster]

176. Well, Diebold has given in. Sorta.

They are releasing the mdb file and the accompanying GEMS db file, AFTER Diebold removes anything that might compromise "security". Make sure you read the entire doc... http://bbvdocs.org/diebold/alaska-surrenders.pdf