I have a Verizon EVDO broadband "wireless" (basically a cellphone signal) contract.
I pay a flat rate of $60/month, probably the same as in his case. (He might be on an $80/mo plan but that's rare.)
I doubt VERY much that data transmission costs are any higher in Canada than they are in the US.
So having already paid the flat rate, it would make perfect sense to me that the add-on extra cost for data transmission in Canada would be as low as.002 cents/KB. Because that's not the TOTAL cost of the data, but a surcharge on top of the money they get off me every month regardless.
What they're trying to do is bill him another $76 - making him pay over double his normal costs when in reality, THEIR costs for that month is only fractionally higher...probably pretty damn close to 76 cents max.
So yeah, I'd buy into that original quote no problem - and it wouldn't seem like a mistake at all!
Which brings us to the next point: if he had known it was that much higher he could have throttled back his usage and probably gotten this bill down near $5/$10 or something...probably by switching to WiFi wherever possible.
Which is why the setup that does the scanning is all open-source, with the standardized code set published with hashes so that observers can confirm the code is the "legit stuff" that 10,000+ geek eyeballs have looked at.
OK. Lemme 'splain something here. Speed MATTERS when we're talking about congressional races.
There have been two recent court decisions in California (the recent CD50 case in San Diego) and Nevada in which state court judges have ruled that each house of congress has the right to approve new members based on EARLY, uncertified, unofficial election results. And in the CD50 case that I know of, this "stamp of approval" on a GOP candidate by the GOP house leadership happened before the final canvass and before the 1% hand recount spotcheck mandated by state law.
That's right, folks. According to these courts, somehody like Hastert can deep-six any state election protection law that state passes and "annoint" candidates.
If this demented thinking holds, then we don't just need to find fraud or screwups, we've got to do so FAST. Hand recounting of all the paper (or even a random selection!) will take too long. We need to throw known-good code at trustworthy data in a hurry.
The idea behind the graphic scans (about 200dpi mono will do) is to be able to throw a COMPLETELY different code set at the ballot stack quickly and cheaply. Or more than one additional code set.
The registrar of voters of Humboldt County California came up with a good idea. Please, no pot jokes:). She runs a Diebold setup doing optical scan ballots. Her plan is to spend about $15k on a new scanner, a big commercial monster with it's own integral disk burner, fast, double-sided and with at least a 500-page intake hopper. That's doable.
She wants to take every paper ballot and just feed it through that monster and produce a set of CDs with the output, and put them on the county website as ZIPs or something, or hand out CD copies to anybody.
It's the same idea: let's throw a whole 'nuther code set at the stuff, independent of every line of code to ever come out of Diebold.
Problem is, it doesn't include hashing...but hey, it's a first step.
If you go further and build the whole setup around graphic scans from the get-go, you CAN hash them. You can also burn a serial number onto the CD (or coming soon, HD-DVD, whatever, so long as it's "-R" and not "-RW").
You're envisioning CDs (or other media) that are specially marked *externally*. By marking them "internally" (via data) you can cut as many disks as you want on election night and if two of the same thing happen to get fed into the county's central tabulator software (or anybody else's!) it's easy to write code that says "if you eat serial number "x" once, and then see it again, go ahead and check that it's the same - if it is, report the dupe and don't eat it. If it does NOT match then somebody is up to no good, scream bloody murder and halt while humans sort it out".
And that's another thing: whenever the software encounters a glitch, it should say so AND record it to unalterable media - an audit trail log from hell. Upon each halt, observers representing the public, parties and/or candidates should be given the opportunity to at least view what's up, photograph or otherwise record the errors, get an explanation of what was going on then and what the plan is to recover.
We've seen counties deliberately cover up errors during elections or pre/post election "Logic and Accuracy" tests. In some cases observers were reading bluescreen text when the election officials literally yanked power plugs to blank the screen.
For a real freakshow example of election officials behaving badly (including loading PC Anywhere on the central tabulator and hooking it up to the county intranet with no firewall!) see also my report on Memphis TN:
First, the "ES&S" machine they're talking about as a base is the Automark, which ES&S bought and is now downplaying over their less secure setup.
Second: the optical scan half of this equation should scan GRAPHICS of each ballot, store them for later review, hash them to prevent later tampering and make them available by the DVD load (or HD-DVD or whatever) as a public record. Remember, the voter name is already stripped out at this point. And if you're burning data to "-R" media of some sort, that eliminates tampering with the data once it's out of the initial box. (Right now most vendors distribute the "electronic ballot box" data on either PCMCIA or USB flash read/write media, which is insane...only Avante burns end-of-day tally data from touchscreens onto CD-R.)
Fourth: once "we the people" have the graphic scans in hand, we can tabulate them with OUR software tools versus trusting the county's tools, providing a "software check and balance". The open source "Votoscope" program written by Harri Hursti was the first attempt at this. Now that we have a decent open-source OCR package (the one HP/Google just released), we have a foundation for building more. Imagine a world in which each news agency brought their best "gamer class" powerhouse monster PC to the elections office to get the scans and do a tabulation before anybody else can, with various non-profits like BlackBoxVoting, VoteTrustUSA or local equivelents chugging along on slower gear but still able to churn out accurate numbers on election night. If the various copies don't match, OK, let's figure out why, by eyeballs on the original paper if necessary.
A good electronic voting setup can backstop against paper fraud just as much as the paper backstops electronic fraud.
The memos themselves show that the lawyers KNEW Diebold was performing "unnatural acts" with America's elections and hence hose *every* voter. That part is plain as day. Jeez, they describe budgets for contingency planning for criminal charges probably coming...this when no such charges were filed (and thanks to Bill Lockyer never were...).
So deciding that this should be run through the typing pool can be described as Pretty Damn Stupid[tm].
Let's see, it's kinda like sketching out a memo saying "to reward our new clients running a gay S&M club, we'll invite them over to anally rape the whole typing pool" THROUGH THE TYPING POOL and not expect all hell to break loose.
Quoting: "No, again, his interest was only that of any other citizen."
Exactly. We are ALL being screwed sans lube. Keep reading.
Quoting: "Only if we can accept your overwrought premise."
Heh. "Overwrought?"
Where do we even start?
Diebold's central vote tabulator was written in MS-Access. But wait, it gets worse!
One of the key developers of this "crapware" was a gent name of Jeffrey Dean.
Dean...oh my God, go grab something to munch on and follow this tale:
1986 - 1988 range, Dean worked as an accountant running the computerized bookkeeping for a Seattle law firm name of Culp, Guterson & Grader. Dean was convicted of 23 counts of computer aided embezzlement of that firm, totalling over $400k. Odd thing: he said years later in testimony in another court case that "yeah, I took the money, but it didn't go into my pocket" (paraphrase). On being questioned further, he claimed that a large group of lawyers within the firm were diverting money out, investing it, keeping the profits and returning the original cash. As the accountant, he was a key "bit player" in the scam and the only one to go down for it, four years state pen when it all blew up. This was in Seattle, all handled out of the King County (WA) prosecutor's office.
So he gets out of the pokey and a year later he owns a printing company (under his wife's name to hide his assets from the court-ordered restitution) called "Spectrum Print and Mail". He lands a printing contract with the King County elections office - to print ballots. By around 1996 - 1998 timeframe he's developing software for the county under contract, including "VoteRemote" for automating the absentee ballot checking process and the rest of the software used to sort and process absentee ballots for the county at the Spectrum printing plant in Everett, WA. (VoteRemote is now a Diebold product and Diebold now owns and runs that plant with it's procedures and software written by Dean.)
Towards the latter half of the period in which he was working literally at the King County elections office, the prosecutor's office on a different floor of the same building was allegedly looking for him due to the restitution failure. Yet despite riding up and down the elevator with the SOB daily they did nothing.
By 2000 he sold Spectrum to Global Elections Systems, the voting machine manufacturer later bought by Diebold in 2002 and now known as Diebold Election Systems. Immediately upon joining Global he was put in charge of development for that crappy MS-Access-based central tabulator, GEMS...whose Windows program icon is a fist holding a globe I $hit you not. Dean was writing add-ons to GEMS while still at Spectrum. Diebold Corporate kept almost the entire Global management structure in place in the new elections division.
One of Spectrum's top employees was John Elder, somebody Dean met in the pokey. A coke dealer and no, I don't mean the kind you drink.
OK, back to King County. Who was still a major customer of Global after their buyout of Spectrum and where Dean frequently supported the elections division until in 2002 elections supervisor Julie Ann Kemf (may have that spelling wrong) got one too many "bad vibes" off of Dean and did some checking, found out about his massive felony embezzlement conviction and booted him out of the building for good, and started the process of ripping out every voting machine the funky SOB had ever touched. At which point Kemf was arrested for DUI and assault on a local cop, a lurid story made the local paper and she was fired...after which all charges were dropped. Her punishment for trying to clean up that county's election process.
Right, let's switch gears to Global, the chaps.
Founded in 1988 in Vancouver BC by:
Norton Cooper - jail for a year mid-1980s for fraud against the Canada government; ordered out of stock pitch schemes and was part of the collapse of the Vancouver stock exchange - ordered by decree not to pitch stock after 1992 o
Would it change your opinion to know that large portions of the file show Jones Day actively helping Diebold cover up ongoing violations of election law?
Understand something: as a voter, Heller was a direct *victim* of the misconduct being jointly committed by Diebold and Jones Day. So he had more than a passing interest in the matter, he was directly being screwed over by these actions and through these documents.
That's part of what sets this completely apart from most other breaches of this nature. It's as if a rape victim working for a lawyer found out that the lawyer was concealing the identity of her assailant.
Screaming bloody murder about it is a natural result.
...I can tell y'all that the most important thing in it is what's NOT in it: any indication that Diebold was fixing the actual problems, or for that matter even interested in doing so.
There are also several places where Jones Day was actively involved in covering up ongoing illegalities by Diebold, which is probably what got Heller as upset as he was.
Jim March Staffer, Black Box Voting http://blackboxvoting.org/ Lead plaintiff, March v. Diebold, the suit that got California that $2.6mil refund...which believe me, had a TON more support than just the Heller material and would have come out the same way regardless, basically quashed by California Attorney General Lockyer before we could even get to discovery. It should have been a refund of all Diebold contracts in California, somewhere over $100mil.
The pattern of date discrepancies does NOT look like "pure machine glitch" (hardware issues like a CMOS battery failure or corruption) and also doesn't look like the possible result of an OS bug.
The way they're mostly "clustered" in a limited date period of Oct. 13th - 20th of the correct year says to me "human intervension". It's not "randomized" the way most computer glitches are.
Next: by way of Jeremiah Akin, Riverside County elections staff have said that the PS/2 keyboard port on the back of each touchscreen terminal is used for, among other things, "to change the date and time".
We know from the logs on the serial numbers of the machines affected that the dates were accurate during the "logic and accuracy test" typically performed up to a month before the election.
OK, let's assume the Riverside folks are right about the keyboard being required for manual date/time changes.
Standard practice in the elections biz is to do the L&A then shut the machine down and DON'T mess with it until election morning. This is basic across all voting machines and has been since the lever days going back to the 19th century.
If the date was messed with by a human with a keyboard between the time of the L&A and the time of the election, well...what the holy hell were they doing? Once the keyboard is in you can tweak the boot order in ROM, loading new code off of new media, or maybe individual programs. (We know little about the OS on these but the boot ROM system is basically same as any laptop.)
In other words, it's not that radical a guess to say that somebody was up to something no good and the date weirdness was just a side effect.
If they were doing a very serious hack involving loading new code, it's possible that what they did hosed the date and they needed to reset it by hand...and in 40 or so cases they forgot that part?
Under this hypothesis the range of dates from the 13th to the 20th is maybe the time the "midnight black hat crew" spent touching each machine. The number of days involved is about right.
Again, this is speculation. We need the manuals on these things to understand the date function in detail. And the process by which new code or data is loaded, probably via PCMCIA card.
We need to replicate ALL these various errors and figure out how they happened, what could cause them and whether or not they're "pointers" to deeper problems, whether that's just "bad gear" or somebody actually loading a vote-shaving routine of some sort.
Sorry I can't respond to posts scattered throughout, I'm kinda busy right now:).
But here's some general info not found in the story:
1) Glade County Florida gladly handed us at Black Box Voting a copy of their GEMS data file (the MS-Access abortion). Diebold didn't do squat to 'em. So the people saying Alaska's elections office is to blame are dead on. What are they hiding? They're among dozens of other jurisdictions also refusing these data files across the nation. Diebold has been distributing a memo asking them not to but legally it isn't worth it's weight in broken video card parts.
2) If y'all want to see the cease'n'desist from Diebold to me asking me to take my site down (containing these same types of files) in 2003, it's still online at:
(If you see a black Buell S3 motorcycle running around the Seattle area with the words "LIEBOLD" on both sides of the gas tank, wave, that's me:).)
The point is, they've known the files are out there, I dared them to sue me via a DMCA counter-notification including giving them my home address for process service and they backed down. There's no more "secret sauce" here as the trademark lawyers put it.
(The files on my site are being moved this week over to http://www.blackboxvoting.org/ with pointers to the new locations as I'm now paid staff at BBV. That's a fairly recent development but immaterial to this situation.)
3) The MS-Access data files contain a "double set of books"...all of the vote data is duplicated in two tables. If you query the Diebold-written front end ("GEMS") for data on the whole county's election totals, those numbers come out of one table. If you query for any one precinct or a group of same, the numbers come out of the other table. By default they match. To hack an election, you rig the numbers that provide the whole county totals via MS-Access itself or VB scripts or Java or whatever tweaking on the Jet database engine. That way, the hapless clueless honest GEMS user at the county elections office who can't tell there's two tables is hosed. IF they suspect trouble at all, they spot-check individual precincts, hand-counting the totals and matching them to the individual precinct totals in GEMS. Do that a few times, they'll think it's all cool. They have no way of knowing there's "two sets of books" in the damned thing unless they print out EVERY precinct and add them up on a hand calculator.
4) If Diebold concerts tables to Excel, y'all REALLY think they'll export both if somebody hacked one? Riiiight. Hence the need for the raw file. (Oh yeah. There's a THIRD table. We don't know what it's for. Yet.)
Now look, it's not certain this was done in Alaska, OK? Actually, this whole thing in Alaska doesn't really look like a deliberate vote hack. We've seen some already, they're slicker than this...like James Bond (well except for that idiocy in Volusia County 2000 but nevermind). Whatever happened in Alaska was more "Inspector Clouseau". Probably just a dumb screwup on the part of elections officials.
But "we the people" (or at least the geekier among us) damned well have the right to sort it out, and that's why this is going to get pushed to a lawsuit, if not in Alaska, somewhere else. There are other states like Washington and Colorado where there are cash penalties for wrongfully denying public records so they're reaaaally tempting targets if the Alaska Democrats drop this ball. But...having talked to them, I don't think they will, I think they're going to follow this all the way to court and win.
One way or another, we're going to get access to these data files, it's a no-brainer.
Due to security design issues and contractual non-performance, Leon County (Florida) supervisor of elections Ion Sancho told Black Box Voting that he will never use Diebold in an election again. He has requested funds to replace the Diebold system from the county. He will issue a formal announcement to this effect shortly.
Finnish security expert Harri Hursti proved that Diebold lied to Secretaries of State across the nation when Diebold claimed votes could not be changed on the memory card.
A test election was run in Leon County today with a total of eight ballots - six ballots voted "no" on a ballot question as to whether Diebold voting machines can be hacked or not. Two ballots, cast by Dr. Herbert Thomson and by Harri Hursti voted "yes" indicating a belief that the Diebold machines could be hacked.
At the beginning of the test election the memory card programmed by Harri Hursti was inserted into an Optical Scan Diebold voting machine. A "zero report" was run indicating zero votes on the memory card. In fact, however, Hursti had pre-loaded the memory card with plus and minus votes.
The eight ballots were run through the optical scan machine. The standard Diebold-supplied "ender card" was run through as is normal procedure ending the election. A results tape was run from the voting machine.
Correct results should have been:
Yes:2 No:6
However the results tape read:
Yes:7 No:1
The results were then uploaded from the optical scan voting machine into the GEMS central tabulator. The central tabulator is the "mothership" that pulls in all votes from voting machines. The results in the central tabulator read:
Yes:7 No:1
This proves that the votes themselves were changed in a one-step process that would not be detected in any normal canvassing procedure - using only a credit-card sized memory card.
Diebold Elections Systems head of research and development Pat Green specifically told the Cuyahoga County board of elections that votes could not be changed on the memory card.
According to Public Records responses obtained by Black Box Voting in response to our requests shows that Diebold promulgated this misrepresentation to as many as 800 state and local elections officials.
In other news, according to Bradblog a stockholder suit was filed today against Diebold by the law offices of Scott and Scott:
Six months ago Leon County elections administrator Ion Sancho asked us (Black Box Voting) to "test hack" his Diebold optical scan system. We brought Finnish security expert Harri Hursti and Dr. Hugh Thomson from Florida along.
Dr. Thomson proved that the central tabulator's database (in MS-Access of all things) can be hacked without a retail copy of MS-Access present. He used Visual Basic to control the MS Jet database engine directly, using very small script files...small enough to be typed in via MS-Windows Notepad at the tabulator console. We already knew the MS-Access database was tamper-friendly but this was real-world proof that you didn't need to bring in and load a copy of Access to tamper. The same things can almost certainly be done in Java and probably other ways as well.
Harri Hursti pulled off something new.
The report co-written with Bev Harris proved it's possible to doctor the poll tapes. These are the end-of-day printouts showing the number of votes for each candidate or issue taken in on that machine. It's basically
1) Ethics matter when we're dealing with our democracy. If you can't understand that, well...
2) Diebold has a specific history of withholding modified code from the test labs and lying to the labs. There is every reason to believe they'll do the same to the government. See also these files for documented case histories of such fraud against the testing labs:
But our anon buddy here has a good point. Not only is it pretty odd to go hire these former cyber-crooks, it kinda turns them into "superstars" who all the little idiot script-kiddies want to grow up to be.
In other words, if the normal path to "superstardom in computer security" with a big fat paycheck is seen to run through the world of black-hat stuff, that's a bad thing for society.
Countering that though is how Godawful annoying modern "black hat hacking" is compared to the "hackerz" of 1970s through mid-90s. Nobody is going to hire phishers, bot-net artists and the like due to public outcry, we're WAY more pissed off about them than we ever were at Mitnick.
Little known fact: the source code for WinCE is fully known to the hardware vendors.
It's unique among Windows versions in that it's not a finished product - each hardware vendor has to finish it for their own weird gear. WinCE was made to run on hardware that is NOT industry standard, everything from PDAs to TV set-top boxes.
Up through CE 3.0 you could download the entire source code from Microsoft's website. I think once they included the.NET stuff they stopped doing that but I could be wrong.
At the central vote tally box, the Diebold GEMS central tabulator runs on top of WinNT/2000 series so they can't put THAT source in escrow.
Fun fact about GEMS: not only was convicted embezzler and admitted murderer Jeffrey Dean in charge of development for at least a couple of years, the program icon is a hoot. It's a fist holding a globe, basically a day-glow-colors version of the corporate logo for Dr. Evil in the Austin Powers movies:).
We should prowl around Diebold HQ looking for midgets, bald cats and sharks with unusual head prosthetics...
Mention of both are extensive in the various online databases of Global/Diebold's internal memos between 1998 and early 2003. Go google:
"Jeffrey dean" diebold...and you'll get about 350 hits, so this is real well known among people paying attention to this stuff.
To be fair, at the time Diebold bought Global Dean was moved to consultant status, possibly to avoid the Diebold corporate background check. They damned well know about him NOW of course ever since Bev Harris broke the news.
Look, Global was based out of Vancouver BC. Bev and others have gone up there to talk to current and former employees...a LOT appeared to be "coked up" or talked about rampant drug abuse up there. If what we're hearing is anywhere close to accurate, Global acted like the set of a John Belushi movie or something.
Trust me on this: ain't no WAY Diebold will want to publish lists of programmers.
Notice how Diebold talks about source code escrow as the issue in NC? It's a red herring. Diebold does source code escrow in California no problem.
The issue is the programmer names. Major-grade doom involved.
It's true that getting a total list of programmers in an open-source system would be impossible.
But as a practical matter it's impossible to name all of the Windows programmers either. The court wouldn't expect that of Diebold any more than they'd require a total list of Linux programmers from an open-source voting project.
What Diebold could easily do is name their own programmers.
Except there's no way in hell they'd want to do that.
In 2002 Diebold bought Global Election Systems, which became the Diebold Election Systems unit. Global was founded under another name in 1988 by Norton Cooper, Michael K. Graye and Charles Hong Lee...all with damned interesting resumes (footnote 1):
Norton Cooper - jail for a year mid-1980s for fraud against the Canada government; ordered out of stock pitch schemes and was part of the collapse of the Vancouver stock exchange - ordered by decree not to pitch stock after 1992 or so because he caused havoc every time. Written up by Barron's and Forbes as a "hazard to avoid at the golf course". First convicted of political corruption in 1974 - look up a Canadian case titled "The Queen v. Norton Cooper" 1977 Canadian Supreme Court.
Charles Hong Lee - stock schemes; Cooper's partner pitching deals. Defrauded Chinese immigrants, $600,000(Can) court-ordered restitution mid-90s. Sold "real estate" which was actually the bail for the third partner below to the tune of about $300,000(can) circa 1995ish.
Michael K. Graye - nailed for stealing $18mil from three companies in the '88-'89 era, caught in '94, jailed in the US for stock fraud around '94 re: Vinex wines, released around 2000 - 2002(3?) in the US, brought back to Canada, still in jail there. Arrested for tax evasion and money laundering circa '94.
Those three in turn hired even more "colorful" staff:
John Elder was a cocaine trafficker, in a WA prison early/mid 1990s...fellow inmate was Jeffrey Dean (see next entry). Handled ballot printing for Global late 1990s. Seems to have been the one to bring Dean into Global.
Jeffrey Dean was convicted early '90s of 23 counts of computer-aided embezzlement. He was a computer consultant for a large Seattle law firm and defrauded them of about $450,000 in what US courts called a "sophisticated computer-aided scheme". In a statement to Seattle PD, he claimed he needed the money because Canadians were blackmailing him; in that country, he'd gotten into a fistfight and the other guy had died. (Yes, I've seen the police report.) He joined Elder in the Global ballot printing business late '90s, and with Global's introduction was doing computer consulting with the King County WA elections division - they had no idea of his criminal record. By 2000 he was doing programming for Global and by early Oct. of 2000 he was a full employee and lead programmer for the GEMS vote-tally product still in use. By late Oct. 2000 and shipping in time for the November election, GEMS ver.1.17.5 contains the first "double set of books" problem where all votes are recorded twice internally and don't need to match...long story but it apparantly hides some forms of vote fraud. At the time Diebold bought Global in 2002, Dean quit and was immediately hired back as a consultant via management decision made within the division. This appears to be an attempt to keep Dean's criminal past out of Diebold corporate head office's scrutiny.
At the time Diebold bought Global, Dean owned 10% of Global's stock.
We don't know how many other lower-level progammers within Global/Diebold have criminal records. It's rather obvious that Diebold sure as hell doesn't want us finding out.
Footnote 1 - see also "Black Box Voting: Ballot Tampering In The 21st Century" by Bev Harris, esp. the "Diebold" section at the end of Chapter 8. Free PDF downloads can be found at: http://blackboxvoting.org/
Look, speaking personally as Republican with strong Libertarian leanings (a "Ron Paul Republican"), I voted Bush over Kerry in '04. I'm not all that enthralled with Dubya, far from it, but I hate Kerry's guts.
So I'm not saying Kerry probably should have won Ohio because I enjoy saying it. Far from it, the words stick in my throat. (It looks to me like it was a combination of electronic vote fraud and "disenfranchisement fraud", messing with voter registration rolls and not putting enough voting stations in college and minority areas with high Democratic turnouts.)
The fact is, we had more election-related violence before and during the 2004 election than any other that I can recall (almost age 40). If public confidence in the vote collapses, it'll be civil war within 10 or 20 years no matter WHO is running things.
I mentioned a bit ago a link to an article I wrote citing and debunking R. Doug Lewis' dismissal of voter verified paper trails.
I didn't know it at the time, but Dr. David Jefferson had already seen that same article by Lewis and did his own debunking of it.
Jefferson is a very capable computer security expert and one of the better academics trying to do watchdogging on all this. He's actually gotten better of late at being willing to blow the whistle on various election systems fouls although he could have done better early on.
In any case, here's what Jefferson thought of Lewis:
So yeah, Lewis is one person I have just about zero respect for. All the worst stuff happened on his watch. The entire process he screwed up has been taken away from him very publicly and is being given to the new EAC.
The other thing is, hey, this is Slashdot:). We're geeks. We gotta have at least some place where we can tell it like it is, right?:)
A lot of this is about borrowing the "technologies" to do proper accounting from the world of CPAs, banks and financial accounting systems. It's necessary to track "who did what" in great detail. When massive amounts of money flow through a bank, better believe they know every human being inside and out of the bank who had a hand in the transfer...and they keep a very non-erasable copy of that data.
Many of those ideas can be transferred to tracking the processing of votes. True, the name of the voter gets stripped from the vote real early in the process, and that's one difference from accounting practice, but from that point on "bank grade tracking" is not only possible it's damned necessary.
This isn't all about electronic controls, either. Wells Fargo Bank had proper transaction processing as far back as the 1860s...Lloyd's of London had it right going back to the 1600s.
One KEY element: say you take something away from a tally. You don't erase and toss the data! Instead you record that a deduction was made from the total, who did it, when and a note on why. Then if it turns out the deduction was in error it's fixable. NONE of the major-vendor voting systems act like this. Need a record gone? Erase it. Ghaaa.
Avanti and OVC both have an interesting take on the audit records at the voting terminal: record everything to CD-R as a series of sessions. You end up with a fixed non-erasable record of votes. THAT media is what gets tallied back at county elections HQ. Diebold and the rest use PCMCIA cards or similar read-WRITE (and erasable) media.
It is NOT easy to arrange a "test hack" (red team attack) on real live voting machines.
Ion Sancho in Leon County FL took *massive* political flak for allowing us to do one there. One of my posts below I describe why Diebold has been protected for so long ("Re:Just wondering...some partial answers."). This is NOT like the general PC biz where you can buy systems or components to test...you've got to be allowed access to systems that are under lock and key...
1) To hand-count, you have to solve the "where do we get the warm bodies" problem. Hold a school holiday the day of the election and the day after, use high school and college kids is one answer.
2) If you make the electronic record as good as possible, we *might* be able to use it as a fraud-check against the paper ballots in SOME forms of "old fashioned paper fraud" of the type that date back to Tammany Hall and the like (late 19th century). BUT if there's disagreement between paper and electrons and there's no way to tell which is the more honest, the paper wins.
Why?
Paper ballot fraud isn't as dangerous as electronic fraud. Paper fraud requires a massive system of con artists all working together. It IS possible but it's got to be really systematic and obvious...think New York City circa 1900, Chicago of the early 1960s.
Electronic vote fraud allows as little as ONE fraudster to do mass hacks.
Starting with Diebold: basically there were FOUR different groups that all made mistakes with this stuff in general, but esp. where Diebold were concerned. No...wait, FIVE. In no particular order:
1) Federal Election Commission: the FEC makes the rules for voting machine certification, the so-called "1990" and "2002" standards. Problem is, they didn't codify them into regulations. They don't have the force of law...they're literally known as "voluntary guidelines". The FEC also approves the testing labs, private companies licenced by the FEC to do source code and functionality reviews paid for by the vendors. The testing labs are called "ITAs" for "Independent Testing Authorities".
2) National Association of State Elections Directors: NASED was in control over how the ITAs did business. They would check over the ITA's paperwork on any particular certification and assign a "NASED number" signifying Federal certification. They didn't happen to notice that the ITAs were acting like a pack of diseased baboons...when it was pretty damned obvious. NASED got some operational support via cash donations from the big vendors.
3) The ITAs themselves, esp. Ciber Inc and the elections division of Wyle Labs, both in Huntsville Alabama. Complete and total wastes of skin. Jam a pocket calculator halfway into a banana, they'll certify it as a voting machine for the right money.
4) The various state certification panels. Some were OK, others said "well hey, as long as it's been Federally certified, well by golly that's good enough for us!" It wasn't. (Oh, and despite NASED's name, the states were NOT able to control NASED much. NASED appears to have gone "rogue" years ago and right now their certification oversight ability is being *stripped* from them and given to the new "Election Assistance Commission"...which isn't functional yet. Shows you how hosed NASED was though.)
5) Various academics and "experts" who were supposed to be checking this stuff out. Even the best of them (Prof. Doug Jones of Iowa) didn't want to get too "vocal" about the issue, esp. early on. Others like Brit Williams and Paul Croft just actively aided and abetted the chaos. There were a small number of notable exceptions such as Dr. Rebecca Mercuri but she was a "voice in the wilderness" drowned out by the "nothing is wrong" crowd. See my other post in this article covering "test mode" for testing and ask yourself if something is wrong.
Basically, the FEC created a crappy program and let a total cheezewiz-for-brains name of R. Doug Lewis run it over at NASED. See also:
Lewis and his minions weren't watching the ITAs. The ITAs missed multiple glaring security holes. The vendors knew nobody was watching the store and Diebold in particular acted like a pack of Goths sacking Rome.
To criticize Diebold is to critique the WHOLE SORRY HOUSE OF CARDS who all generally acted like they were all members of the same big happy club...vendors and ITAs included. It gets worse: people from one part of this structure often relocated to other parts, including back and forth between vendors and government oversight. Diebold, Sequoia and ES&S *all* hired high-level staff from within the California SecState's office to go lobby their former co-workers and bosses, and that's just in California. This was and remains common nationwide.
That's why Diebold has been protected...they go down, people might look too close, the whole thing collapses in scandal.
Mind you, some people in high places are STARTING to get it.
Example: in California, Diebold tried to get approval for a new touchscreen setup in mid-2005. Somebody at the California SecState's office wisely decided to do a "volume test" and without even worrying about security, checked for basic reliability - and found a 30%+ failure rate. In the "aftermath repor
Slashdot Mirror
...early in '06. There were only three labs approved to test voting machines, Ciber, Wyle and Systest.
i fornia-ITA-hearings-27281.pdf
Ciber and Wyle are in Huntsville AL, right next to the Redstone arsenal. They mostly do military gear testing, voting systems are a sideline.
Systest is in Colorado and *might* be the most competent of the bunch. Wyle and Ciber were the two used most by Diebold.
All three labs were invited to testify in California. Ciber was a no-show.
The other two come across as complete loons, Wyle especially. There is a transcript up of the hearings here, plus a link to the original MP3:
http://www.bbvforums.org/forums/messages/2197/Cal
I have a Verizon EVDO broadband "wireless" (basically a cellphone signal) contract.
.002 cents/KB. Because that's not the TOTAL cost of the data, but a surcharge on top of the money they get off me every month regardless.
I pay a flat rate of $60/month, probably the same as in his case. (He might be on an $80/mo plan but that's rare.)
I doubt VERY much that data transmission costs are any higher in Canada than they are in the US.
So having already paid the flat rate, it would make perfect sense to me that the add-on extra cost for data transmission in Canada would be as low as
What they're trying to do is bill him another $76 - making him pay over double his normal costs when in reality, THEIR costs for that month is only fractionally higher...probably pretty damn close to 76 cents max.
So yeah, I'd buy into that original quote no problem - and it wouldn't seem like a mistake at all!
Which brings us to the next point: if he had known it was that much higher he could have throttled back his usage and probably gotten this bill down near $5/$10 or something...probably by switching to WiFi wherever possible.
Which is why the setup that does the scanning is all open-source, with the standardized code set published with hashes so that observers can confirm the code is the "legit stuff" that 10,000+ geek eyeballs have looked at.
i nst_votersupervised_elections_attorney_0906.html
OK. Lemme 'splain something here. Speed MATTERS when we're talking about congressional races.
There have been two recent court decisions in California (the recent CD50 case in San Diego) and Nevada in which state court judges have ruled that each house of congress has the right to approve new members based on EARLY, uncertified, unofficial election results. And in the CD50 case that I know of, this "stamp of approval" on a GOP candidate by the GOP house leadership happened before the final canvass and before the 1% hand recount spotcheck mandated by state law.
That's right, folks. According to these courts, somehody like Hastert can deep-six any state election protection law that state passes and "annoint" candidates.
If this demented thinking holds, then we don't just need to find fraud or screwups, we've got to do so FAST. Hand recounting of all the paper (or even a random selection!) will take too long. We need to throw known-good code at trustworthy data in a hurry.
For those not aware of the San Diego fiasco:
http://www.bradblog.com/?p=3353
http://www.rawstory.com/news/2006/Court_rules_aga
Jim March
The idea behind the graphic scans (about 200dpi mono will do) is to be able to throw a COMPLETELY different code set at the ballot stack quickly and cheaply. Or more than one additional code set.
:). She runs a Diebold setup doing optical scan ballots. Her plan is to spend about $15k on a new scanner, a big commercial monster with it's own integral disk burner, fast, double-sided and with at least a 500-page intake hopper. That's doable.
h .cgi?file=/1954/44242.html
The registrar of voters of Humboldt County California came up with a good idea. Please, no pot jokes
She wants to take every paper ballot and just feed it through that monster and produce a set of CDs with the output, and put them on the county website as ZIPs or something, or hand out CD copies to anybody.
It's the same idea: let's throw a whole 'nuther code set at the stuff, independent of every line of code to ever come out of Diebold.
Problem is, it doesn't include hashing...but hey, it's a first step.
If you go further and build the whole setup around graphic scans from the get-go, you CAN hash them. You can also burn a serial number onto the CD (or coming soon, HD-DVD, whatever, so long as it's "-R" and not "-RW").
You're envisioning CDs (or other media) that are specially marked *externally*. By marking them "internally" (via data) you can cut as many disks as you want on election night and if two of the same thing happen to get fed into the county's central tabulator software (or anybody else's!) it's easy to write code that says "if you eat serial number "x" once, and then see it again, go ahead and check that it's the same - if it is, report the dupe and don't eat it. If it does NOT match then somebody is up to no good, scream bloody murder and halt while humans sort it out".
And that's another thing: whenever the software encounters a glitch, it should say so AND record it to unalterable media - an audit trail log from hell. Upon each halt, observers representing the public, parties and/or candidates should be given the opportunity to at least view what's up, photograph or otherwise record the errors, get an explanation of what was going on then and what the plan is to recover.
We've seen counties deliberately cover up errors during elections or pre/post election "Logic and Accuracy" tests. In some cases observers were reading bluescreen text when the election officials literally yanked power plugs to blank the screen.
For a real freakshow example of election officials behaving badly (including loading PC Anywhere on the central tabulator and hooking it up to the county intranet with no firewall!) see also my report on Memphis TN:
http://www.bbvforums.org/cgi-bin/forums/board-aut
Jim March
First, the "ES&S" machine they're talking about as a base is the Automark, which ES&S bought and is now downplaying over their less secure setup.
Second: the optical scan half of this equation should scan GRAPHICS of each ballot, store them for later review, hash them to prevent later tampering and make them available by the DVD load (or HD-DVD or whatever) as a public record. Remember, the voter name is already stripped out at this point. And if you're burning data to "-R" media of some sort, that eliminates tampering with the data once it's out of the initial box. (Right now most vendors distribute the "electronic ballot box" data on either PCMCIA or USB flash read/write media, which is insane...only Avante burns end-of-day tally data from touchscreens onto CD-R.)
Fourth: once "we the people" have the graphic scans in hand, we can tabulate them with OUR software tools versus trusting the county's tools, providing a "software check and balance". The open source "Votoscope" program written by Harri Hursti was the first attempt at this. Now that we have a decent open-source OCR package (the one HP/Google just released), we have a foundation for building more. Imagine a world in which each news agency brought their best "gamer class" powerhouse monster PC to the elections office to get the scans and do a tabulation before anybody else can, with various non-profits like BlackBoxVoting, VoteTrustUSA or local equivelents chugging along on slower gear but still able to churn out accurate numbers on election night. If the various copies don't match, OK, let's figure out why, by eyeballs on the original paper if necessary.
A good electronic voting setup can backstop against paper fraud just as much as the paper backstops electronic fraud.
Jim March
Member of the Board of Directors, Black Box Voting Inc.
http://blackboxvoting.org/
The memos themselves show that the lawyers KNEW Diebold was performing "unnatural acts" with America's elections and hence hose *every* voter. That part is plain as day. Jeez, they describe budgets for contingency planning for criminal charges probably coming...this when no such charges were filed (and thanks to Bill Lockyer never were...).
So deciding that this should be run through the typing pool can be described as Pretty Damn Stupid[tm].
Let's see, it's kinda like sketching out a memo saying "to reward our new clients running a gay S&M club, we'll invite them over to anally rape the whole typing pool" THROUGH THE TYPING POOL and not expect all hell to break loose.
Diebold fired 'em. Gee, I wonder why?
Jim March
Staffer - BBV
http://blackboxvoting.org/
Quoting: "No, again, his interest was only that of any other citizen."
Exactly. We are ALL being screwed sans lube. Keep reading.
Quoting: "Only if we can accept your overwrought premise."
Heh. "Overwrought?"
Where do we even start?
Diebold's central vote tabulator was written in MS-Access. But wait, it gets worse!
One of the key developers of this "crapware" was a gent name of Jeffrey Dean.
Dean...oh my God, go grab something to munch on and follow this tale:
1986 - 1988 range, Dean worked as an accountant running the computerized bookkeeping for a Seattle law firm name of Culp, Guterson & Grader. Dean was convicted of 23 counts of computer aided embezzlement of that firm, totalling over $400k. Odd thing: he said years later in testimony in another court case that "yeah, I took the money, but it didn't go into my pocket" (paraphrase). On being questioned further, he claimed that a large group of lawyers within the firm were diverting money out, investing it, keeping the profits and returning the original cash. As the accountant, he was a key "bit player" in the scam and the only one to go down for it, four years state pen when it all blew up. This was in Seattle, all handled out of the King County (WA) prosecutor's office.
So he gets out of the pokey and a year later he owns a printing company (under his wife's name to hide his assets from the court-ordered restitution) called "Spectrum Print and Mail". He lands a printing contract with the King County elections office - to print ballots. By around 1996 - 1998 timeframe he's developing software for the county under contract, including "VoteRemote" for automating the absentee ballot checking process and the rest of the software used to sort and process absentee ballots for the county at the Spectrum printing plant in Everett, WA. (VoteRemote is now a Diebold product and Diebold now owns and runs that plant with it's procedures and software written by Dean.)
Towards the latter half of the period in which he was working literally at the King County elections office, the prosecutor's office on a different floor of the same building was allegedly looking for him due to the restitution failure. Yet despite riding up and down the elevator with the SOB daily they did nothing.
By 2000 he sold Spectrum to Global Elections Systems, the voting machine manufacturer later bought by Diebold in 2002 and now known as Diebold Election Systems. Immediately upon joining Global he was put in charge of development for that crappy MS-Access-based central tabulator, GEMS...whose Windows program icon is a fist holding a globe I $hit you not. Dean was writing add-ons to GEMS while still at Spectrum. Diebold Corporate kept almost the entire Global management structure in place in the new elections division.
One of Spectrum's top employees was John Elder, somebody Dean met in the pokey. A coke dealer and no, I don't mean the kind you drink.
OK, back to King County. Who was still a major customer of Global after their buyout of Spectrum and where Dean frequently supported the elections division until in 2002 elections supervisor Julie Ann Kemf (may have that spelling wrong) got one too many "bad vibes" off of Dean and did some checking, found out about his massive felony embezzlement conviction and booted him out of the building for good, and started the process of ripping out every voting machine the funky SOB had ever touched. At which point Kemf was arrested for DUI and assault on a local cop, a lurid story made the local paper and she was fired...after which all charges were dropped. Her punishment for trying to clean up that county's election process.
Right, let's switch gears to Global, the chaps.
Founded in 1988 in Vancouver BC by:
Norton Cooper - jail for a year mid-1980s for fraud against the Canada government; ordered out of stock pitch schemes and was part of the collapse of the Vancouver stock exchange - ordered by decree not to pitch stock after 1992 o
Would it change your opinion to know that large portions of the file show Jones Day actively helping Diebold cover up ongoing violations of election law?
Understand something: as a voter, Heller was a direct *victim* of the misconduct being jointly committed by Diebold and Jones Day. So he had more than a passing interest in the matter, he was directly being screwed over by these actions and through these documents.
That's part of what sets this completely apart from most other breaches of this nature. It's as if a rape victim working for a lawyer found out that the lawyer was concealing the identity of her assailant.
Screaming bloody murder about it is a natural result.
Jim March
Staff, Black Box Voting
http://blackboxvoting.org/
...I can tell y'all that the most important thing in it is what's NOT in it: any indication that Diebold was fixing the actual problems, or for that matter even interested in doing so.
There are also several places where Jones Day was actively involved in covering up ongoing illegalities by Diebold, which is probably what got Heller as upset as he was.
Jim March
Staffer, Black Box Voting
http://blackboxvoting.org/
Lead plaintiff, March v. Diebold, the suit that got California that $2.6mil refund...which believe me, had a TON more support than just the Heller material and would have come out the same way regardless, basically quashed by California Attorney General Lockyer before we could even get to discovery. It should have been a refund of all Diebold contracts in California, somewhere over $100mil.
Sigh.
The pattern of date discrepancies does NOT look like "pure machine glitch" (hardware issues like a CMOS battery failure or corruption) and also doesn't look like the possible result of an OS bug.
The way they're mostly "clustered" in a limited date period of Oct. 13th - 20th of the correct year says to me "human intervension". It's not "randomized" the way most computer glitches are.
Next: by way of Jeremiah Akin, Riverside County elections staff have said that the PS/2 keyboard port on the back of each touchscreen terminal is used for, among other things, "to change the date and time".
We know from the logs on the serial numbers of the machines affected that the dates were accurate during the "logic and accuracy test" typically performed up to a month before the election.
OK, let's assume the Riverside folks are right about the keyboard being required for manual date/time changes.
Standard practice in the elections biz is to do the L&A then shut the machine down and DON'T mess with it until election morning. This is basic across all voting machines and has been since the lever days going back to the 19th century.
If the date was messed with by a human with a keyboard between the time of the L&A and the time of the election, well...what the holy hell were they doing? Once the keyboard is in you can tweak the boot order in ROM, loading new code off of new media, or maybe individual programs. (We know little about the OS on these but the boot ROM system is basically same as any laptop.)
In other words, it's not that radical a guess to say that somebody was up to something no good and the date weirdness was just a side effect.
If they were doing a very serious hack involving loading new code, it's possible that what they did hosed the date and they needed to reset it by hand...and in 40 or so cases they forgot that part?
Under this hypothesis the range of dates from the 13th to the 20th is maybe the time the "midnight black hat crew" spent touching each machine. The number of days involved is about right.
Again, this is speculation. We need the manuals on these things to understand the date function in detail. And the process by which new code or data is loaded, probably via PCMCIA card.
We need to replicate ALL these various errors and figure out how they happened, what could cause them and whether or not they're "pointers" to deeper problems, whether that's just "bad gear" or somebody actually loading a vote-shaving routine of some sort.
Jim March
Black Box Voting staff
http://blackboxvoting.org/
Yo Mark Radke: the register and login buttons are up near the top left corner.
:). Slightly. I'm sure the location is different on the diebold.com home page but you'll adapt.
You're slightly saner in person
Sorry I can't respond to posts scattered throughout, I'm kinda busy right now :).
:).)
But here's some general info not found in the story:
1) Glade County Florida gladly handed us at Black Box Voting a copy of their GEMS data file (the MS-Access abortion). Diebold didn't do squat to 'em. So the people saying Alaska's elections office is to blame are dead on. What are they hiding? They're among dozens of other jurisdictions also refusing these data files across the nation. Diebold has been distributing a memo asking them not to but legally it isn't worth it's weight in broken video card parts.
2) If y'all want to see the cease'n'desist from Diebold to me asking me to take my site down (containing these same types of files) in 2003, it's still online at:
http://www.equalccw.com/liebold.html
(If you see a black Buell S3 motorcycle running around the Seattle area with the words "LIEBOLD" on both sides of the gas tank, wave, that's me
The point is, they've known the files are out there, I dared them to sue me via a DMCA counter-notification including giving them my home address for process service and they backed down. There's no more "secret sauce" here as the trademark lawyers put it.
(The files on my site are being moved this week over to http://www.blackboxvoting.org/ with pointers to the new locations as I'm now paid staff at BBV. That's a fairly recent development but immaterial to this situation.)
3) The MS-Access data files contain a "double set of books"...all of the vote data is duplicated in two tables. If you query the Diebold-written front end ("GEMS") for data on the whole county's election totals, those numbers come out of one table. If you query for any one precinct or a group of same, the numbers come out of the other table. By default they match. To hack an election, you rig the numbers that provide the whole county totals via MS-Access itself or VB scripts or Java or whatever tweaking on the Jet database engine. That way, the hapless clueless honest GEMS user at the county elections office who can't tell there's two tables is hosed. IF they suspect trouble at all, they spot-check individual precincts, hand-counting the totals and matching them to the individual precinct totals in GEMS. Do that a few times, they'll think it's all cool. They have no way of knowing there's "two sets of books" in the damned thing unless they print out EVERY precinct and add them up on a hand calculator.
4) If Diebold concerts tables to Excel, y'all REALLY think they'll export both if somebody hacked one? Riiiight. Hence the need for the raw file. (Oh yeah. There's a THIRD table. We don't know what it's for. Yet.)
Now look, it's not certain this was done in Alaska, OK? Actually, this whole thing in Alaska doesn't really look like a deliberate vote hack. We've seen some already, they're slicker than this...like James Bond (well except for that idiocy in Volusia County 2000 but nevermind). Whatever happened in Alaska was more "Inspector Clouseau". Probably just a dumb screwup on the part of elections officials.
But "we the people" (or at least the geekier among us) damned well have the right to sort it out, and that's why this is going to get pushed to a lawsuit, if not in Alaska, somewhere else. There are other states like Washington and Colorado where there are cash penalties for wrongfully denying public records so they're reaaaally tempting targets if the Alaska Democrats drop this ball. But...having talked to them, I don't think they will, I think they're going to follow this all the way to court and win.
One way or another, we're going to get access to these data files, it's a no-brainer.
Then...let's talk source code.
Jim March
Staffer/investigator
Black Box Voting Inc.
http://www.blackboxvoting.org/
PS: Alaska
To quote the latest article on the Black Box Voting site (and then some background below that):
---
http://www.bbvforums.org/forums/messages/1954/1559 5.html?1134523376
Due to security design issues and contractual non-performance, Leon County (Florida) supervisor of elections Ion Sancho told Black Box Voting that he will never use Diebold in an election again. He has requested funds to replace the Diebold system from the county. He will issue a formal announcement to this effect shortly.
Finnish security expert Harri Hursti proved that Diebold lied to Secretaries of State across the nation when Diebold claimed votes could not be changed on the memory card.
A test election was run in Leon County today with a total of eight ballots - six ballots voted "no" on a ballot question as to whether Diebold voting machines can be hacked or not. Two ballots, cast by Dr. Herbert Thomson and by Harri Hursti voted "yes" indicating a belief that the Diebold machines could be hacked.
At the beginning of the test election the memory card programmed by Harri Hursti was inserted into an Optical Scan Diebold voting machine. A "zero report" was run indicating zero votes on the memory card. In fact, however, Hursti had pre-loaded the memory card with plus and minus votes.
The eight ballots were run through the optical scan machine. The standard Diebold-supplied "ender card" was run through as is normal procedure ending the election. A results tape was run from the voting machine.
Correct results should have been:
Yes:2 No:6
However the results tape read:
Yes:7 No:1
The results were then uploaded from the optical scan voting machine into the GEMS central tabulator. The central tabulator is the "mothership" that pulls in all votes from voting machines. The results in the central tabulator read:
Yes:7 No:1
This proves that the votes themselves were changed in a one-step process that would not be detected in any normal canvassing procedure - using only a credit-card sized memory card.
Diebold Elections Systems head of research and development Pat Green specifically told the Cuyahoga County board of elections that votes could not be changed on the memory card.
According to Public Records responses obtained by Black Box Voting in response to our requests shows that Diebold promulgated this misrepresentation to as many as 800 state and local elections officials.
In other news, according to Bradblog a stockholder suit was filed today against Diebold by the law offices of Scott and Scott:
http://www.bradblog.com/archives/00002153.htm
Permission to reprint granted with link to http://blackboxvoting.org/
---
Jim again. Let me fill you in on the background.
Six months ago Leon County elections administrator Ion Sancho asked us (Black Box Voting) to "test hack" his Diebold optical scan system. We brought Finnish security expert Harri Hursti and Dr. Hugh Thomson from Florida along.
Dr. Thomson proved that the central tabulator's database (in MS-Access of all things) can be hacked without a retail copy of MS-Access present. He used Visual Basic to control the MS Jet database engine directly, using very small script files...small enough to be typed in via MS-Windows Notepad at the tabulator console. We already knew the MS-Access database was tamper-friendly but this was real-world proof that you didn't need to bring in and load a copy of Access to tamper. The same things can almost certainly be done in Java and probably other ways as well.
Harri Hursti pulled off something new.
The report co-written with Bev Harris proved it's possible to doctor the poll tapes. These are the end-of-day printouts showing the number of votes for each candidate or issue taken in on that machine. It's basically
Two things:
1) Ethics matter when we're dealing with our democracy. If you can't understand that, well...
2) Diebold has a specific history of withholding modified code from the test labs and lying to the labs. There is every reason to believe they'll do the same to the government. See also these files for documented case histories of such fraud against the testing labs:
http://www.equalccw.com/sscomments1.pdf
http://www.equalccw.com/sscomments2.pdf
True 'nuff.
But our anon buddy here has a good point. Not only is it pretty odd to go hire these former cyber-crooks, it kinda turns them into "superstars" who all the little idiot script-kiddies want to grow up to be.
In other words, if the normal path to "superstardom in computer security" with a big fat paycheck is seen to run through the world of black-hat stuff, that's a bad thing for society.
Countering that though is how Godawful annoying modern "black hat hacking" is compared to the "hackerz" of 1970s through mid-90s. Nobody is going to hire phishers, bot-net artists and the like due to public outcry, we're WAY more pissed off about them than we ever were at Mitnick.
Little known fact: the source code for WinCE is fully known to the hardware vendors.
.NET stuff they stopped doing that but I could be wrong.
:).
It's unique among Windows versions in that it's not a finished product - each hardware vendor has to finish it for their own weird gear. WinCE was made to run on hardware that is NOT industry standard, everything from PDAs to TV set-top boxes.
Up through CE 3.0 you could download the entire source code from Microsoft's website. I think once they included the
At the central vote tally box, the Diebold GEMS central tabulator runs on top of WinNT/2000 series so they can't put THAT source in escrow.
Fun fact about GEMS: not only was convicted embezzler and admitted murderer Jeffrey Dean in charge of development for at least a couple of years, the program icon is a hoot. It's a fist holding a globe, basically a day-glow-colors version of the corporate logo for Dr. Evil in the Austin Powers movies
We should prowl around Diebold HQ looking for midgets, bald cats and sharks with unusual head prosthetics...
Jim March
Black Box Voting (staff)
http://www.bbvdocs.org/dean.pdf
...and you'll get about 350 hits, so this is real well known among people paying attention to this stuff.
http://www.bbvdocs.org/elder.pdf
There's their criminal records.
Mention of both are extensive in the various online databases of Global/Diebold's internal memos between 1998 and early 2003. Go google:
"Jeffrey dean" diebold
To be fair, at the time Diebold bought Global Dean was moved to consultant status, possibly to avoid the Diebold corporate background check. They damned well know about him NOW of course ever since Bev Harris broke the news.
Look, Global was based out of Vancouver BC. Bev and others have gone up there to talk to current and former employees...a LOT appeared to be "coked up" or talked about rampant drug abuse up there. If what we're hearing is anywhere close to accurate, Global acted like the set of a John Belushi movie or something.
Trust me on this: ain't no WAY Diebold will want to publish lists of programmers.
Notice how Diebold talks about source code escrow as the issue in NC? It's a red herring. Diebold does source code escrow in California no problem.
The issue is the programmer names. Major-grade doom involved.
It's true that getting a total list of programmers in an open-source system would be impossible.
But as a practical matter it's impossible to name all of the Windows programmers either. The court wouldn't expect that of Diebold any more than they'd require a total list of Linux programmers from an open-source voting project.
What Diebold could easily do is name their own programmers.
Except there's no way in hell they'd want to do that.
In 2002 Diebold bought Global Election Systems, which became the Diebold Election Systems unit. Global was founded under another name in 1988 by Norton Cooper, Michael K. Graye and Charles Hong Lee...all with damned interesting resumes (footnote 1):
Norton Cooper - jail for a year mid-1980s for fraud against the Canada government; ordered out of stock pitch schemes and was part of the collapse of the Vancouver stock exchange - ordered by decree not to pitch stock after 1992 or so because he caused havoc every time. Written up by Barron's and Forbes as a "hazard to avoid at the golf course". First convicted of political corruption in 1974 - look up a Canadian case titled "The Queen v. Norton Cooper" 1977 Canadian Supreme Court.
Charles Hong Lee - stock schemes; Cooper's partner pitching deals. Defrauded Chinese immigrants, $600,000(Can) court-ordered restitution mid-90s. Sold "real estate" which was actually the bail for the third partner below to the tune of about $300,000(can) circa 1995ish.
Michael K. Graye - nailed for stealing $18mil from three companies in the '88-'89 era, caught in '94, jailed in the US for stock fraud around '94 re: Vinex wines, released around 2000 - 2002(3?) in the US, brought back to Canada, still in jail there. Arrested for tax evasion and money laundering circa '94.
Those three in turn hired even more "colorful" staff:
John Elder was a cocaine trafficker, in a WA prison early/mid 1990s...fellow inmate was Jeffrey Dean (see next entry). Handled ballot printing for Global late 1990s. Seems to have been the one to bring Dean into Global.
Jeffrey Dean was convicted early '90s of 23 counts of computer-aided embezzlement. He was a computer consultant for a large Seattle law firm and defrauded them of about $450,000 in what US courts called a "sophisticated computer-aided scheme". In a statement to Seattle PD, he claimed he needed the money because Canadians were blackmailing him; in that country, he'd gotten into a fistfight and the other guy had died. (Yes, I've seen the police report.) He joined Elder in the Global ballot printing business late '90s, and with Global's introduction was doing computer consulting with the King County WA elections division - they had no idea of his criminal record. By 2000 he was doing programming for Global and by early Oct. of 2000 he was a full employee and lead programmer for the GEMS vote-tally product still in use. By late Oct. 2000 and shipping in time for the November election, GEMS ver.1.17.5 contains the first "double set of books" problem where all votes are recorded twice internally and don't need to match...long story but it apparantly hides some forms of vote fraud. At the time Diebold bought Global in 2002, Dean quit and was immediately hired back as a consultant via management decision made within the division. This appears to be an attempt to keep Dean's criminal past out of Diebold corporate head office's scrutiny.
At the time Diebold bought Global, Dean owned 10% of Global's stock.
We don't know how many other lower-level progammers within Global/Diebold have criminal records. It's rather obvious that Diebold sure as hell doesn't want us finding out.
Footnote 1 - see also "Black Box Voting: Ballot Tampering In The 21st Century" by Bev Harris, esp. the "Diebold" section at the end of Chapter 8. Free PDF downloads can be found at: http://blackboxvoting.org/
Look, speaking personally as Republican with strong Libertarian leanings (a "Ron Paul Republican"), I voted Bush over Kerry in '04. I'm not all that enthralled with Dubya, far from it, but I hate Kerry's guts.
So I'm not saying Kerry probably should have won Ohio because I enjoy saying it. Far from it, the words stick in my throat. (It looks to me like it was a combination of electronic vote fraud and "disenfranchisement fraud", messing with voter registration rolls and not putting enough voting stations in college and minority areas with high Democratic turnouts.)
The fact is, we had more election-related violence before and during the 2004 election than any other that I can recall (almost age 40). If public confidence in the vote collapses, it'll be civil war within 10 or 20 years no matter WHO is running things.
We have to have fair elections. Period.
Jim March
Black Box Voting
Do we need to boot off the test CD?
Rather than booting off it, run the checksum script on the CD which in turn pulls the MD5 numbers off the CD.
Hrrrrmmmmm.
Yeah, I know what you're thinking, I'm thinking the same thing. "Rootkit".
Sigh.
Crap, we might have to boot off CD after all. Have the CD boot process put up some big splash banner?
Jim
I mentioned a bit ago a link to an article I wrote citing and debunking R. Doug Lewis' dismissal of voter verified paper trails.
:). We're geeks. We gotta have at least some place where we can tell it like it is, right? :)
I didn't know it at the time, but Dr. David Jefferson had already seen that same article by Lewis and did his own debunking of it.
Jefferson is a very capable computer security expert and one of the better academics trying to do watchdogging on all this. He's actually gotten better of late at being willing to blow the whistle on various election systems fouls although he could have done better early on.
In any case, here's what Jefferson thought of Lewis:
http://verifiedvoting.org/article.php?id=68
So yeah, Lewis is one person I have just about zero respect for. All the worst stuff happened on his watch. The entire process he screwed up has been taken away from him very publicly and is being given to the new EAC.
The other thing is, hey, this is Slashdot
Jim March
Black Box Voting
Just about.
A lot of this is about borrowing the "technologies" to do proper accounting from the world of CPAs, banks and financial accounting systems. It's necessary to track "who did what" in great detail. When massive amounts of money flow through a bank, better believe they know every human being inside and out of the bank who had a hand in the transfer...and they keep a very non-erasable copy of that data.
Many of those ideas can be transferred to tracking the processing of votes. True, the name of the voter gets stripped from the vote real early in the process, and that's one difference from accounting practice, but from that point on "bank grade tracking" is not only possible it's damned necessary.
This isn't all about electronic controls, either. Wells Fargo Bank had proper transaction processing as far back as the 1860s...Lloyd's of London had it right going back to the 1600s.
One KEY element: say you take something away from a tally. You don't erase and toss the data! Instead you record that a deduction was made from the total, who did it, when and a note on why. Then if it turns out the deduction was in error it's fixable. NONE of the major-vendor voting systems act like this. Need a record gone? Erase it. Ghaaa.
Avanti and OVC both have an interesting take on the audit records at the voting terminal: record everything to CD-R as a series of sessions. You end up with a fixed non-erasable record of votes. THAT media is what gets tallied back at county elections HQ. Diebold and the rest use PCMCIA cards or similar read-WRITE (and erasable) media.
Sigh.
Jim March
Black Box Voting
It is NOT easy to arrange a "test hack" (red team attack) on real live voting machines.
Ion Sancho in Leon County FL took *massive* political flak for allowing us to do one there. One of my posts below I describe why Diebold has been protected for so long ("Re:Just wondering...some partial answers."). This is NOT like the general PC biz where you can buy systems or components to test...you've got to be allowed access to systems that are under lock and key...
Jim March
Black Box Voting
Let me note two things:
1) To hand-count, you have to solve the "where do we get the warm bodies" problem. Hold a school holiday the day of the election and the day after, use high school and college kids is one answer.
2) If you make the electronic record as good as possible, we *might* be able to use it as a fraud-check against the paper ballots in SOME forms of "old fashioned paper fraud" of the type that date back to Tammany Hall and the like (late 19th century). BUT if there's disagreement between paper and electrons and there's no way to tell which is the more honest, the paper wins.
Why?
Paper ballot fraud isn't as dangerous as electronic fraud. Paper fraud requires a massive system of con artists all working together. It IS possible but it's got to be really systematic and obvious...think New York City circa 1900, Chicago of the early 1960s.
Electronic vote fraud allows as little as ONE fraudster to do mass hacks.
Jim March
Black Box Voting
You've asked two questions :).
Starting with Diebold: basically there were FOUR different groups that all made mistakes with this stuff in general, but esp. where Diebold were concerned. No...wait, FIVE. In no particular order:
1) Federal Election Commission: the FEC makes the rules for voting machine certification, the so-called "1990" and "2002" standards. Problem is, they didn't codify them into regulations. They don't have the force of law...they're literally known as "voluntary guidelines". The FEC also approves the testing labs, private companies licenced by the FEC to do source code and functionality reviews paid for by the vendors. The testing labs are called "ITAs" for "Independent Testing Authorities".
2) National Association of State Elections Directors: NASED was in control over how the ITAs did business. They would check over the ITA's paperwork on any particular certification and assign a "NASED number" signifying Federal certification. They didn't happen to notice that the ITAs were acting like a pack of diseased baboons...when it was pretty damned obvious. NASED got some operational support via cash donations from the big vendors.
3) The ITAs themselves, esp. Ciber Inc and the elections division of Wyle Labs, both in Huntsville Alabama. Complete and total wastes of skin. Jam a pocket calculator halfway into a banana, they'll certify it as a voting machine for the right money.
4) The various state certification panels. Some were OK, others said "well hey, as long as it's been Federally certified, well by golly that's good enough for us!" It wasn't. (Oh, and despite NASED's name, the states were NOT able to control NASED much. NASED appears to have gone "rogue" years ago and right now their certification oversight ability is being *stripped* from them and given to the new "Election Assistance Commission"...which isn't functional yet. Shows you how hosed NASED was though.)
5) Various academics and "experts" who were supposed to be checking this stuff out. Even the best of them (Prof. Doug Jones of Iowa) didn't want to get too "vocal" about the issue, esp. early on. Others like Brit Williams and Paul Croft just actively aided and abetted the chaos. There were a small number of notable exceptions such as Dr. Rebecca Mercuri but she was a "voice in the wilderness" drowned out by the "nothing is wrong" crowd. See my other post in this article covering "test mode" for testing and ask yourself if something is wrong.
Basically, the FEC created a crappy program and let a total cheezewiz-for-brains name of R. Doug Lewis run it over at NASED. See also:
http://www.equalccw.com/lewisdeconstructed.pdf
Lewis and his minions weren't watching the ITAs. The ITAs missed multiple glaring security holes. The vendors knew nobody was watching the store and Diebold in particular acted like a pack of Goths sacking Rome.
To criticize Diebold is to critique the WHOLE SORRY HOUSE OF CARDS who all generally acted like they were all members of the same big happy club...vendors and ITAs included. It gets worse: people from one part of this structure often relocated to other parts, including back and forth between vendors and government oversight. Diebold, Sequoia and ES&S *all* hired high-level staff from within the California SecState's office to go lobby their former co-workers and bosses, and that's just in California. This was and remains common nationwide.
That's why Diebold has been protected...they go down, people might look too close, the whole thing collapses in scandal.
Mind you, some people in high places are STARTING to get it.
Example: in California, Diebold tried to get approval for a new touchscreen setup in mid-2005. Somebody at the California SecState's office wisely decided to do a "volume test" and without even worrying about security, checked for basic reliability - and found a 30%+ failure rate. In the "aftermath repor