Miami Herald now reporting that 70,000 votes were “discovered” last night in Broward County, FL for a gambling initiative that reversed the previously announced outcome. [emphasis mine, and mostly notable to those who have been following along with previously reported items]:
The bug, discovered two years ago but never fixed, began subtracting votes after the absentee tally hit 32,500 — a ceiling put in place by the software makers.
…
The problem was quickly traced to software in what is known as the central tabulation machine, a computer that collects data from optical scanners that read the individual mail-in ballots.
…
County officials blamed Election Systems & Software, the company that sold the machines and counting software to Broward.
County officials say they think ES&S failed to follow through on a problem that was brought to their attention two years ago, during the 2002 general election.
ES&S spokeswoman Becky Vollmer said the glitch — which limits the number of votes that can be counted in each precinct to safeguard against ballot stuffing — will be fixed in software updates they are submitting to the Division of Elections next year.
…
Alia Faraj, [FL Secretary of State, Glenda] Hood’s spokeswoman, said ES&S had not previously submitted any information about the counting cap in its tabulation software, which is supposed to be certified by the Secretary of State’s Office.
…
The tabulation software was set to reverse the vote count at 32,500. It was triggered when Broward counted all 97,535 absentee ballots in one mega-precinct Tuesday night and early Wednesday.
Hello? Connect. The. Dots.
(For more on the Optical Scan oddities in Florida, and info on who ES&S is, see this previous item.)
Brad,
Are you sure that the 32,500 isn’t actually 32,768? When I heard that number I immediately thought that the error is actually a very common computer glitch. If the program uses the wrong kind of data storage type, then it can only hold values from negative 32,767 to positive 32,768. What happens is that the counter will get up to 32, 768 and then it will "roll over" and start counting negative numbers. (Technically, they are storing the number in an eight bit, signed number data structure. However, the final bit is considered a flag to denote whether the number is negative or positive).
This is a very common, but very boneheaded, error that inexperienced programmers make. I wonder if that was the problem.
Mike
Mike, I believe you are correct about the 32,768 number. My sense is that the reporters there rounded it off.
You explanation is one that I agree with and have confirmed with a programming expert friend of mine. I was hoping to getting around to explaining it more, so thank you for doing me the favor 🙂
(DATAHEAD-SPEAK ON)The sense that I get, and which has been confirmed for me, is that the part of the code that caused the error may have been created in a 16-bit environment some years ago, and for whatever reason that INTEGER value with the 32,768 top value would have been updated to a 32-bit INTEGER — allowing for more than enough whole numbers for the code — if it had been re-compiled with a 32-bit compiler. Either that, or it was sloppy programming as you suggest. Though sloppier still is the fact that there was no Error Trapping for such a possibility in the code at all apparently!(DATAHEAD-SPEAK OFF)
All of which points up the fact that it is simply both stunning and appalling that lawmakers have allowed these companies to privatize our election to the point that the code is proprietary and the authors (Diebold, ES&S, etc.) will not allow Independent 3rd Parties to evaluate their code against precisely such possible problems!
For all we know, there is a trigger set to add 1 Republican vote for every 5 *real* Democrat votes that get cast in the code!
Do I think that’s the case? Not necessarily, but because REPUBLICANS have supported these company’s desire to keep the code proprietary, secret, and therefore, unvalidated, anything could happen!
But why not? Diebold & ES&S have paid Republicans enough for that privilege so I guess they’re entitled to get what they paid for.
Ah, and P.S. the other thing I forgot to mention, and the most troubling in my opinion.
If it occured on this one machine, and was theoretically caught, how many other machines did it occur on and *not* get noticed??
They don’t program each of these machines seperately ya know. So whatever code this machine was using, it’s probably in use in hundreds (if not thousands) of other precincts around the state and country.
Which again begs the question: Where the fuck are all the "Investigative Journalists" that we’re supposed to have in our Mainstream Media?!
Just an interesting thought, there were literally thousands of mailed absentee ballots in fla that "disappeared" after they were "supposedly" mailed, I believe it was in Broward county. There was a news story that one of the wire services had up about the disappearing ballots, in the news story it mentioned that officials were investigating what happened to the ballots after they were dropped off at the post office, I noticed the story because I work for the post office and found it curious that literally tons of mail could all of a sudden not be found, believe me that would be one very large pile of paper, seems like it would be quite difficult to lose something that massive with no trace.
http://www.reuters.com/newsArti...8;section=news
Let me guess. The slot machine inititiative which has now passed is tied to one of Diebold’s many subsidiary companies? Someone should check into that whole issue, as I am aware that in Ohio, one of four companies providing electronic voting machines was decertified which cleared up a conflict of interest problem for Diebold there. I have some preliminary discrepancies and irregularities worked up on Ohio, but I have seen much more advanced statistical analysis elsewhere on the web.
However, working up Florida 2000 elections versus 2004 elections, not only did "new" votes for Bush outpace "new" votes for Kerry, they did so on a nearly precise 2:1 basis, and this following the GOTV efforts, and the determination on the parts of Democratic voters who were robbed in 2000, and again in 2002. Doesn’t make much sense to me
In the practice of the law, there is something called the "tainted fruit" theory. It is a very simple principle and it goes something like this:
ANY evidence that is obtained through any illegal or unConstitutional means is inadmissable in court.
The question is; why isn’t this standard being applied to our voting in the United States of America? To extend the comparison, if a single voting machine is found to have a programming or tabulation error, its results are tainted and inadmissable. Thus, ALL votes on ALL machines that are of the same make are inadmissable and must be thrown out and a REVOTE done on new, flawless, machines or by paper ballot hand counted.
It is time to revolt for this!
-W.L. Swarts
angry: Do you know how hard it is do think of uniting with our "re-elected president" if you don’t think he won? Any Kerry supporter who reads the information coming out of the "ustogether.com" site will come to one solution: when will the rest of the nation find out what has taken place in Florida. I know this because I am one of them. Lately, there has not been a minute that has passed by without me thinking is some form or another when will the rest of the nation learn of the tragedy that has taken place in Florida’s polling precincts. In my eyes, the irregularities are on a massive scale and I wonder why hasn’t news media responded with a large reaction bigger than any story they are covering at present. I hope they are not of the idea we need to "move on" and not let this voting irregularity issue "polarize" our nation. Right down to the last fiber of my believe system I feel this election cannot stand until I have prove that all those democrats really did vote for Bush in Florida.
Joseph