Hi, Brad - Yes, I see - it's as clear as...as clear as...mud...or a hanging chad...or not as clear as a hanging chad...oh, well...it's not too important...(is it?)...please tell me it's not...(sigh)

Am I allowed to swear here? Computer glitch my ass!!

Does anyone have a computer battery they want me to change? Of course I may have to take the whole damn machine apart and spend 3 hours doing that . .in the interim explain to people how to make the ensure that the totals on the hand and machine count match.

WTF??

From a Diebold memo:

"I'm assuming the terminals dialup a modem in central office in alameda county and establish a ppp or slip connection with their local network.
It is insane to think they would have these things connect to the internet no matter what kind of security procedurs they take people can get around them."

Just another joke directed to those who know nothing about it.

Evidently the folk doing the analysis are utterly incompetent along with the software, or they are into heavy play pretend.

No one who has relevant experience is going to buy these lame monkey-brain excuses.

Oh. Computer error. I'll stop worrying now.

NOT!!!!!!!

The "Bullshit meter" is pegging out on this one.

Let's for a moment accept their explanation. It's possible they had a data transmission problem and they're saying they were able to reproduce it.

One would think that the procedures at a voting precinct would require a counting and public posting of the vote totals BEFORE they are sent ANYWHERE.

If they had done this teeny-tiny task, any discrepancy would (should) have been picked up immediately. This is total BS.

But again, lets go back to believing their explanation. If they've got everything all figured out, has anyone asked them the obvious question: Did you go back and check all of the voting machines that used this system? If the mistake happened once, it could easily have happened all over the place. Have they proven to us that it did or did not? Have they even told us whether or not they checked? :angry:

Based on my experience, I'd say this is total garbage. There are quite a few problems with their "explanation".

One: the computer was busy with another task?? Isn't the purpose of the computer to tabulate the votes? What other task could it have been doing? And these computers can execute millions of commands per second. What could have been going on that was so resource-intensive that it couldn't finish it?

For another thing, computers multi-task. Some people do as well, to varying degrees of success, but these machines are made for it.

And there's the matter of data queues. The information sent to a computer gets in line and waits its turn. Could some of it be lost? Possibly, but only if they were doing something really intense with the computer. If that were the case, any techno worth his/her salt would just... wait. Wait until the machine finishes up what it's doing. Our democracy isn't going to crumble because we have to wait 5 more minutes for a vote tally.

Sorry folks, but I don't buy it. If the glitch were so simple, it wouldn't have taken over 3 months to explain it. If anything, it was likely caused by a conflict with the "special" vote tabulating code. But they couldn't come out and say that. So it took them this long to try to replicate the "glitch" using a different cause.

And the blanket term "computer error" has long been used to placate those who aren't techno-savvy. How many times have you had to argue with a utility company who, when proven to be in error, blame "the computer"? In reality, it is most often operator error (usually involving incorrect entering of data). But rather than admit they weren't paying attention, they pass the cyber-buck.

Next they'll be telling us how the data turned in midair, and entered Governor Connolly (sp?).

Keep up the good blogging, Brad and Co. It's people like you who made a difference in the 60's. You guys will ultimately make a difference in the 21st century. While the MSM tries to "mushroom" us, you keep the citizenry informed. Viva la revolution! VR

"Technicians concluded the laptop was busy completing another task just as numbers from that precinct were being fed into it.

"As a result, the laptop did not receive the data as fast as it was sent," said an elections board report on the probe. "Consequently, data was lost"

If data was "lost" then how did it OVERgenerate 3893 votes?

If the problem was in the computer why don't we have more problem in areas that use computers constantly every day (restaurants, atms, on-line banking, stores). I used touch screen computers 10 years ago in a restaurant, and we never had people receiving more that one steak if they didn't order it. Surely that technology wasn't better than what we are voting with.

Computers are *supposed* to multitask. That is their purpose. It is not a valid excuse.

If data was lost, how did they recreate this activity? What were they doing in the first place to muck it up (if indeed it did get mucked up), and how were they able to document this? They got exactly the same number of over votes? That's pretty impressive.

what are they hiding?

Maybe we're all looking at this the wrong way! Maybe it's a historical breakthrough! This would be first time in the history of computing that a loss of data produced an erroneous result that was greater than the actual result! Amazing!!!

Try this at home, folks. It's easy. You don't even need a computer...

[1] Pick a number, any number. Write it down.

[2] Write the same number again, but this time omit one or more digit(s).

Here's the question: Is the second number larger than the first?

No? It's smaller, isn't it? That's the way it usually works when you lose a digit. Losing a byte is the same as losing a digit. So now you can see how remarkable this event was!!

I wonder whether the Association for Computing Machinery has a 'believe-it-or-not' department. They would be very interested in the machine that did this. It's one of a kind. Take that computer to the museum!

This "explanation" is even more bogus than the one given by Edison-Mitovsky (about why the exit polls were "wrong").

Oh, well, you go into election day with the computer you have, not the one you wish you had...

Computers do not really multi-task, unless they have dual processors. Many do, but I doubt this lap top did. I'm not going to try to explain why they claim that single processor machines 'multi-task' but it's your typical advertising BS, and the more computer savvy among us could do better at explaining it. To actually perform two or more calculations or operations simultaneously would require multiple processors. I'd love to be able to afford a dual processor machine. I'm sure the Rebuplicans can afford them, but then this might not be an acceptable explanation. This is more BS from the same folks I wouldn't buy a used car from.

I am thoroughly confused with that explanation. I've read books that better explain Relativity to a Laymen like myself. Where was the source of the "error" again?

As Gilda Radner used to say, "Waaii--t a minute!"
Some guy is sitting with a laptop, tallying votes? The kind you can stick into your briefcase?
In a national election? Can this be real? Who was this clown using a laptop?

If 4,000 votes could be created from whole cloth in a small precinct like Gahanna by some whiz with a computer in his lap, couldn't 50,000 have been created in Cincinnati and another 50,000 in Cleveland?

Has anyone considered that this could have been a stunt by Republicans to show how "the system worked" because an isolated error was detected?
Think about it. Create an obviously erroneous tally. Announce right away, "We found this mistake and fixed it." Everyone applauds and says, "It wasn't nearly enough to overturn the result anyway, but it proves that Ohio runs a fair, bi-partisan system."

This draws attention away from disenfranchisement of voters, dirty tricks, unfair allocation of machines, vote flipping all day in Youngstown, the lockdown in Warren County, exit poll discrepancies and the ridiculous tallies for Judge Connolly and third-party candidates in Cuyahoga County, etc. etc. National media characterize all complaints as the product of conspiracy theories. The Democratic National Committee says, "The election is over." Clinton says, "He won fair and square."

Does anyone doubt Blackwell is capable of a scheme like this? Hopefully, now that he has stiffed both the national secretary of state convention and a Congressional hearing and ticked off even his fellow Republicans, people will begin to put two and two together and get...maybe 100,000.

Okay, here goes:

I've been an IT professional for ten years now and I've been into computers since my early teens (I started programming on a Commodore PET!) and I can say that this explanantion is (as we say over here) unadulterated shite.

A quick breakdown:
A PROCESSOR is brains of the computer, the "chip" that does the calcualtions.

A PROCESS is a task which runs on a PROCESSOR

All modern operating systems (and a few older ones - Un*x anyone?) handle mutiple PROCESSES on a single PROCESSOR - that is what they're meant to do. They do this by sharing PROCESSOR time between mutiple PROCESSES. It looks like they're all running at once, but they aren't.

If you run too many PROCESSES at once, the machine slows down as there's only so much PROCESSOR time to go around. Give it any more and the operating system will likely dump a PROCESS to prevent the whole machine going belly up.

To prevent one PROCESS from writing data into another PROCESSES data area, each PROCESS has it's memory area protected by the operating system so that if one PROCESS crashes (that is fails and starts writing garbage into memory), it can't corrupt the whole machine and cause a complete system crash.

About the only way for their expanation to make sense is that if the software is deliberately written to allow other PROCESSES to write into it's memory area - now didn't we hear about a "nice man" in Florida requesting that soemthing like this be done? It does sound strangely familiar to me...

Keep the faith brothers and sisters, their time will come.

PS Re #11 LGM, I just got a multiprocessor machine and it's lovely. We geeks are easily pleased....

Let's assume for a moment we are forced to agree that this "glitch" is indeed the explanation, isn't that then a great reason to correct the system? If all these idiots are going to try to shove these explanations down our throats, then maybe we need to accept them for what they are and begin to use them against the people trying to make the hard sell.

I spent 34 years as a programmer and computer systems engineer, and I concur that the explanation offered in the AP story is inadequate.

Comment #14 makes a valid observation that, if there's any truth at all in the explanation, then all the other results from all the other counties, perhaps nationwide, would possibly (maybe even likely) be invalid as well. In fact, the reported lack of rigorous configuration control on the software and firmware used in the election systems is a horrible scandal in and of itself, regardless of any effect it may have had on the election results. Responsible software engineers just don't behave that way anymore.

Comment #15 has it right too, except to say that the Windows operating system, which may or may not have been involved, (Perhaps depending on the version too--I've been retired awhile.) is not a true multiprogrammed system, but relies upon all the concurrently running programs to be written in such a way that they cooperate with each other, and there ARE ways for one process to peek and poke into another. It's quite a feat, but doable: You'll need a Windows specialist to explain exactly how it's done. In a Unix-based system I'd have more confidence.

Cheers for Brad!

The only question I want answered is, why do all the discrepancies and mistakes go in **sh's favor?

And the results are in! 8.5 million Iraqi's votes - 58%, which is just what they expected, even though Anbar province only accounted for 2% of their voters and Ninevah province 17% of theirs. With no international monitors or UN involvement and with the *delay* in announcing the results of the vote - SORRY! Don't believe it. Another fix has just come down the pike.

The following link describes what News Junkie and Da Wookie are talking about which is heap corruption.

http://www.windowsitpro....ticleID/22275/22275.html

Cheryl #19
Heard on Face the Nation this mornng that the Shia didn't get "quite" enough votes needed to have a true majority. Ya think they needed the extra tabulation time to make this result come about?
Nah...............

The "reasoning" in this AP explanation is just as sterile as the operator of that laptop is going to be.

Gee, it sure is nice to have that all straightened out - for a while there I was afraid someone might have cheated, but now that it's been so beautifully explained how losing data can result in gaining votes I can rest easy. How stupid do they think we are? As incredibly stupid as they are? Sorry.....

Die bold, die snivelling - they don't care as long as we die!

Good one, Czaragorn.

Hi Peg C! I just had this wonderful thought: what we are going to look like in history, what this struggle is going to look like: I think it's going to be an ugly struggle, but I also think it's going to be surprisingly brief, thanks to the blogosphere (soon to be renamed the bradosphere, BTW). This is our chance, nay, our imperative: we must rise above the adversary and save our good old way of life. Norman Rockwell, like that, Sid and Imogene, West and Fields... Where have you gone, Joe Dimaggio? Well guess what? We're up at bat, and we're behind, but I have a feeling that true Bradvillers everywhere were in some way behind the Red Sox last fall, and pleased with the result (we can be sure it wasn't fixed, because if it had been the Yanks would surely have won), and even if we're down 3-0 in games, having just dropped No. 3 by some obscene score in the jewelbox, we'll never give up. It's our turn now, it's up to us. We must now be the greatest generation. With friends like you guys, I know we can do it!

Czaragorn#26
As a fan of the so called "Evil Empire", I want to thank you for reminding me of their humiliating and complete collapse to the Red Sox. I always pull for underdogs, (like us!), but will never, never, never, root for the Sox.

Thanks again, no really.....thanks.

Dang

Perhaps the voting machines are Quantum computers? This would explain it being able to calculate ALL the votes at once and know Bush was the winner. THe 4000 some odd votes were just it getting bored since it did everything already. I do know that if I have my computer running one process to count something and that process crashes it does not factor by one 10 and start counting the next job.

I'm not computer saavy, but this sounds like just another stonewall to me. For starters, I'm not impressed when the people under the gun for a mistake triumphantly present an explanation. I'm not dismissing it out of hand, but I'm not going to automatically buy it either.

Gotta keep pushing. Ohio is where the Juggernaut of lies meets the iceberg of truth.

Hi. This is my first time posting because this is simply outrageous.

If we are to believe what we've been told, then lost or corrupt data resulted in more votes going to Bush. Contrary to the view that some other posters have expressed, a loss of data CAN conceivably create more votes - if this were 1984 and they were transfering the data over a 9600 baud serial connection. Depending on how the bytes of data are formatted and parsed, a shift in a single bit could create a completely different outcome, but would be just as likely to output garbage or something unreadable.

So what do we know about the handling of the votes? Damschroder said the error occured when cartridges were inserted into a card reader on a lap top. Just as an aside, when I hear the word cartridge I can't help but think of an Atari 2600 and its big clunky carts. I doubt the memory cards are some proprietary format, considering that the voting machines were running Windows and the tabulators were 10 year old PCs, so I'll assume they're compact flash or something similar, and that the card reader is attached to a USB port. If all the above are true, then the scenario presented is bogus. I've transfered thousands of gigs of error-sensitive data from memory cards to a PC under heavy load, and I never had a single problem. I can't imagine any practical situation where memory space or process sharing would result in the writing of wrong data. Just about any device maintains proper control over the flow of data, and a serial buffer would make sure data is written in correct order.

We should demand more information about this, including: 1) The hardware and software specifications of the memory card, card reader, and laptop. 2) The process used to transfer data from the cards to the laptop, and from the laptop to the central server. 3) Logs of what processes were running on the machine at the time of the initial card reading. 4) Exactly what data was "fed" to correct the results.

Additionally, we should be asking why this problem occured in this instance only. Was the laptop used a standard set for all counties and precincts? Was the error in the data on the card, in the transfer of the data, or the writing of the data? Didn't the data have a CRC check on it? Is this another example of faith-based voting?

Anyway, this is all small potatoes in light of the bigger picture. The issue, as it relates to voting machine fraud, is that changing a SINGLE BIT of data can result in THOUSANDS of extra votes, swapped votes, or destroyed votes. This shift of bits can occur at numerous stages, through numerous means - either incidental or intentional. THAT is why we need full disclosure, and THAT is why the theory of voting machine fraud remains a very viable one.

Now, how can we work on bringing this to Conyers, and getting the information we need?

Ah, Supersoling - I wish I could have the pleasure of taking you to a game in Fenway Park, some game that really didn't affect your interests at all, just to absorb the atmosphere. Christ on crutches, Ruth played there before he went off to build that Dark Star... Fenway's special, Boston's special, sure, but let's get real - we Bradvillers are special, and we're gonna be the ones that do the thing!

Yo BVAC! Welcome aboard! We're going to need your help. Let's go get 'em!

"As a result, the laptop did not receive the data as fast as it was sent," said an elections board report on the probe. "Consequently, data was lost."

In other words, the software driver used to send data from the external hardware device to the host laptop/OS did not utilize buffers correctly or at all. (Since the machine in question is a laptop, this data was most likely transferred via a PCMCIA, USB or Firewire bus/interface.)

Since this problem happened no where else, it seems reasonable to conclude that the software driver in use on this laptop was not the same as the one used elsewhere. [Why? And isn't this a violation of state or federal law?]

For comparison's sake, let's look at similar data transfers that happen all the time on our PC's that effectively deal with "maxed out" CPU & bus saturation:

If your computer is at 100% CPU usage (maybe you're running Seti@Home or something similar) & you refresh a web page, does the page refresh incorrectly just because your CPU is maxed out? No, it just happens a little more slowly than normal. Your network card driver, the bus/interface this network card uses to deliver data to/from the CPU & the host operating system all work to throttle the data transfer as needed.

If you have an IDE CDROM drive and an IDE HD hooked up to the same channel - and you copy data from the CDROM to the HD, that IDE channel/interface is maxed out (or close to being maxed out) while the data is flying over the IDE bus. If you try to load some other data off the HD while the copy process is still happening --- for instance, a web browser --- does your web browser program load incorrectly just because the IDE bus is maxed out at this point? No, it just happens a little more slowly than normal. The drivers, bus & OS involved all throttle the data transmissions as needed.

Similar case scenarios can be described for PCMCIA, USB and Firewire buses/interfaces.

"Danaher went further, identifying what computer bytes in the data disappeared. By feeding in data missing those bytes, technicians produced the same wrong vote total."

So, they can reproduce the problem. If you can reproduce it, you were also able to test this type of issue before using the code on election day.

In short, if this explanation holds up, it only proves that the programmers working for this e-vote company are incompetent and/or didn't test their software drivers adequetly. And as I said before, it also raises the question of whether or not the software in use on this laptop was different than the software in use elsewhere. And if so, is it a violation of law?

This is yet another reason why the computer systems used in our democratic elections *need* to be opend up for inspection by IT professionals. If you want a contract that's this important to our democratic process, you've got to be willing to allow peer code review by non-partisan IT professionals. If these e-vote companies are concerned about their intellectual property rights, all they have to do is ask the peer code reviewers to sign an NDA. The NDA should also allow exception for reporting the nature of discovered coding flaws & the progress made to resolve them to the general public.

I'll say it again - JUST LOSE THE FRICKIN' MACHINES.

I'm not a luddite, I work with them all day long, I used to build web servers for a living, I can calculate an IP subnet in my head, I have programmed in C, C++, basic, pascal and numerous assemblers and they make this sort of shenenigans wayyyy too easy. Whether you massage the votes on the voting machines themselves or further up the chain, it's too simple to do and too difficult to track down. Even open sourcing it won't make it safe IMO.

Paper and pencil votes and hand counting is the way to go. As we've seen with the optical counting machines, even getting them to just count is open to violation.

Next election, let's KISS.

Keep the faith and damn the machines.

Look, everyone... not one but TWO beautiful new programming experts have joined us! BVAC and A SYSADMIN, welcome!!! Please stick around - we need all the help we can get.

Each seemingly bogus explanation that comes out gives us more to chip at.

I agree that ditching the e-vote machines entirely would be the most simple & verifiable solution. Ideally, that is what I'd like to see.

However, in principle, secure e-voting can be implemented & it can work. And with Bush & the GOP in control, we're going to continue to see privitization of previously-government run programs/processes. Over the next 4 years, I doubt we'll see any of these decisions reversed (unless a huge amount of corruption is exposed). Personally, I think we're more likely to be listened to if we try to find a solution that fixes what we've already got. Hopefully, peer code reviews & the new Voting Itegrity and Verification Act introduced last week (by a Republican, no less) which will require a paper receipt for each vote cast, will restore some confidence in our democratic voting process.

Click here for VIVA info.

Also, ITAA's objection to VIVA is complete bull. Do ATM receipt printers ever jam up causing long lines at the ATM? If so, how is this any more of a problem than the e-vote machine software or OS potentially crashing (like when all those unpatched ATM's got hit with the Slammer worm last year)? Or a potential power outage at the polling place? Or somebody potentially tripping over an e-vote machine's power cord? If ITAA can make an objection based on potential printer jams, we can make equally reasonable objections to the entire idea of e-voting...

As for the added cost of a printer on each e-vote machine... First, I'd sarcastically say "kudos" to the e-vote companies for trying to squeeze as much money as possible out of the state. Second, adding a stupid little dot matrix printer to each e-vote machine won't cost that much money. Hell, the government could start a bidding process for e-vote dot matrix printer add-ons independent of the e-vote companies themselves, and require that the e-vote companies use those printers through a RFC'd interface w/ peer-reviewed drivers. That would keep the cost under control. (These printers would also need to be adequetly tested under high usage to verify that they don't jam up --- and that would not be hard to do).

It was pre-arranged to fiddle the votes for each county - the person on the laptop was given the fiddled votes/software for another LARGER county to input by mistake - it's simple - really.

Yet another excuse out of the "Random Excuse Generator".
I hope Bush is wearing a condom because you people are being fucked every day in every way .........

Please educate me. Ohio brags about its bi-partisan system. One Republican and one Democrat oversee everything, right? And we're told there can't have been massive fraud because both parties would have had to be in on it. Right?

So who was this clown with the laptop, creating 4,000 votes for Bush? What party was he working for? We all know the answer, but where was the Democrat who supposedly was watching him while he was doing this?

Sure, the error was caught...but if 4,000 votes could be flipped in the first place, using a laptop computer, it could have happened 100 times elsewhere in the state without being detected. That's the point. Can anyone assure us it didn't?

> Can anyone assure us it didn't?

You're a wild-eyed conspiracy theorist. You're trying to steal another election. There is absolutely no proof of fraud or disenfranchisement. Sour grapes. You lost. Get over it.

These are the things we are conditioned to believe. Everyone here can see beyond that, but that's not enough. The general public needs to be made aware, not just that computer failure changed the vote, but that anyone capable of changing a few BITS of data could have easily manipulated the vote in a big way.

Full disclosure of what hardware and software was used needs to be made available, and every single sector of hard disk space analyzed. Laptops, machines, and memory cards from different precincts must also be analyzed and compared to the one in Franklin County. If all machines were equally as vulnerable to failure and/or manipulation, this will provide a good springboard for reform.

For us to bring any of our issues to the surface, we're going to need bi-partisan support and independent analysis. Public awareness is growing but needs to continue to grow. In addition to Chris Hitchen's article about the problems in Ohio, some Republicans and conservatives are beginning to question the results there as well.

Welcome to conspiracy reality.

Oh yes, BVAC.

"Nothing to fear but farce itself"

This may seem like a radical suggestion, but why doesn't someone with credentials contact AP and point out that one cannot float a
news story in a collander? That they can't plug their holes fast enough to keep the "explanation" from being and appearing exactly what it is - all wet? It's one thing to see it. It's quite another to cry "foul" to the source itself.

Speaking of which, you'd think that amonst a;; of Associated Press's vast resources of information, there'd be a software engineer lurking somewhere in the wings.

Sorry, typing in the pich dark. Make a;; "all" and make more sense out of the above!

I offer an amendment to Hannah's #10--
"You can have the best computer in the world and still 'their' vote can be blown up!"

FYI, I have alerted the folks at Conyers office both to this story, and to the excellent comments posted above.

Thanks for coming out of the woodwork BVAC and SYSADMIN. Your thoughts are appreciated, and will hopefully get to the folks who need to read them.

The contributions of the regulars (and irregulars) are much appreciated as well! Glad to hear it wasn't just me who finds that explanation wanting...to say the least.

We'll get to the bottom of it yet...

Doesn't this mean that computerized voting machines do not pass voting security tests, if this can happen? Well, now that this is explained (even though it shouldn't be able to happen), next let's work on the exit poll discrepancies, the FBI terror alert lockdown during counting in that one Ohio precinct (I mean, in Kentucky North), the 250-million-to-1 odds that all the errors on computerized voting machines and optical scanners favored George Bush, the 10 hour lines in Democratic precints only, (I can go on and on)...Note to Kentucky North: The vote can be hacked only if it is close, and you made it too close. You got what you deserve!!! Bush is ruining your state and jobs there, and you idiots voted for him at close to a 50% clip. Thanks! And have a great 4 years! Was Ohio in the Confederacy???

Thank you, Brad. Notify also the news service - that you have notified Conyers. These people have GOT to be put on alert that their BS is NOT passing the credibility test, even with their "listening audience."

Be aggressive. We've all got to be.

I think the people are starting to become disenchanted with their great president. If this continues to happen, the crack in the shell just might split. If it really gains momentum, stories like this might be the food they've been waiting for.
Their cognitive dissonance needs to break, and then maybe all this effort will pay off BIG TIME.
His Social Security scam could be his biggest mistake. His popularity is plummeting, and they are just getting the sense that they are being ripped off.
I think relentless, behind the scenes pursuit now, total focus, and persistent collective effort are imperative.

Welcome to ALL new posters! I see there are more and more everyday. AWESOME!!!

Just wanted to say that all the technical comments since my last post ring true to me as well.

A key point is that, unlike it seems to say in the article, the card plugged into the laptop would not initiate the transfer of data except to tell the laptop that it was there and ready to communicate. Then it would just wait. When time was available, the laptop would supervise the downloading of the data according to its own schedule and thus would not be too busy to receive the data. If data were received incorrectly, that would be known by means of several error detecting and correcting techniques that have been in use since the 1960s. The action in that case would be to redo the data transfer until it came through correctly, or, if not, to cease the transfer attempts and issue an error message to the operator. We've all seen this in action many times, I'm sure.

The idea that something might be wrong with the I/O drivers that handle all of the above is extremely unlikely, since they would be the same ones that handle very widely used and tested PCMCIA devices, and all that stuff has been running smoothly for many years.

As an EE specializing in computers, I HAVE seen smaller numbers suddenly grow.

Consider... if you send the number "10" in decimal, you generate the binary pattern "1010". Rarely, however is data so small, so let's pad the upper bits with zeros - the number is still decimal "10", but now the binary pattern looks like "00001010".

Now for the smoke and mirrors: transmit this string, but leave out one of the leading zeros and back fill with a zero (what a computer might do if a bit is lost in transmission and it is expecting 8 bits).

The binary representation now reads "00010100" - this is a decimal "20". Ooops! a smaller number now reads larger!

This said, regardless of the method of moving sensitive data from one place to another, this should have generated a checksum or CRC error.

Whoever coded this/designed this system really had to WANT to allow errors to occur in this system.

As far as the machine doing too much, what else SHOULD it have been doing? I've seen plenty of machines, both DOS, Windows and UN*X thrash (and often crash), but NEVER generate bad data.

I'm concerned about bad machines, but I'm more concerned about lines of authority in Ohio. To say "a machine malfunctioned" is to say metaphorically, "a light bulb blew"; nobody takes responsibility, because it "just happened." The same thing occurs in everyone's house maybe once a week.

Problem is...the light bulbs that blew went out in Kerry's house, while Bush's lights stayed on. That means the problem couldn't have been the light bulbs themselves. It had to be the person or persons who ordered them, or screwed them in to the lamps, or both.

Questions: Who gave this clown with the laptop his job? Assuming he was a Republican appointee, where was the Democrat supposedly working with him? Was he alone when this so-called error occurred? If so, that blows the "bi-partisan safety" argument out of the water. If not, was the other person watching at that moment, and is he or she available to testify as to how this happened?

The point remains...until we know how and why 4,000 votes were created for Bush in Gahanna, we can't be sure it didn't happen all over the state. The fact that the error was detected has no bearing on the question, "Did it happen elsewhere?"

Well, as a Software Engineer, I must say I can see (as pointed out a few times) how a bit-shift (loss of bytes) would have caused this kind of problem. What I can NOT see is, how the hell the bit-shift could have happened?

It's been shown above that 'standard' communication paths in computers are (for all intent and purposes) 100% accurate. That is, given the nature of computers and how important it is for ALL of the bits to be 'correct', reading from things like USB ports, IDE channles, etc etc, has been 'perfected'.

Now we are left with 2 concernes (for me). One is that they seemingly aren't using "standard" hardware (someone above pointed out flash cards in standard readers hooked up to USB ports, for example). I would think we can support that [my claim] even further because you are talking about something that -also- has to be -inside- the actual voting machine.. that is, the cartridge must go from collecting info in the voting machine to putting it into a lap-top. To presume it's a falsh type device is to presume that the manufacturer of the e-voting machines incorperated existing tech into their device instead of making proprietary tech.. and since you are dealing with security issues and the like, I'd guess the readers/writers/cartridges are proprietary. If the 'reader' hooked up to the lap-top on a USB port started pushing data before the receiving driver fully acknowled the cartridge was inserted, you could lose data. That is, part of the transmission buffer -could- be lost. However, why aren't there checks to ensure the integrity of what -was- passed?

The second large concern I have is, why was the data being moved off of the voting machines at all? Why weren't the voting machines set up to -directly- submit the votes to the main system (modems installed, etc)? How is it reasonable to ADD a possible curruption step into the process by moving the data from the voting machines to a lap-top (a perfect place to have an interim program that lets you tweak the votes any way you want, which can then be deleted and made untraceable)?

I agree completely that we need to have full disclosure of all the info (what hardware, software, etc etc. that was involved). Given the 'possibility' of their claim under the proper circumstances, -every- system that uses that hardware/software setup needs to be re-evaluated. If it "happened" where it was blatently obvious, nothing stops it from having possibly happened in most places this setup was used.

The Electronic and Software Engineering needs to be fully reviewed by peers to verify there are no ways in (that is, no ways for silly bugs like they claim happened, and no way to run external programs that might access memory and tweak counts). Current technology (not proprietary) wouldn't allow for "missing bits", and 'solid' code would have caught them on proprietary systems.