portland independent media center  
images audio video
newswire article reposts united states

election fraud | political theory | technology

Fed wants e-voting source code disclosed

DeForest Soaries, chairman of the Election Assistance Commission, or EAC, said disclosing the source code--the line-by-line instructions that make up an electronic voting machine's software--would help to restore public trust in the elections process. Vendors should not "have the right to keep this source code a secret," Soaries told a dinner gathering of Maryland election officials.
Feds want e-voting source code disclosed

By Declan McCullagh
Staff Writer, CNET News.com
 http://news.com.com/2100-1028-5229162.html

Story last modified June 8, 2004, 10:15 PM PDT

HUNT VALLEY, Md.--Electronic voting machine vendors should make their source code available for scrutiny by state elections officials, the head of a federal voting commission said Tuesday.

DeForest Soaries, chairman of the Election Assistance Commission, or EAC, said disclosing the source code--the line-by-line instructions that make up an electronic voting machine's software--would help to restore public trust in the elections process. Vendors should not "have the right to keep this source code a secret," Soaries told a dinner gathering of Maryland election officials.

Soaries' suggestions, which also include standardized security checks and better record-keeping of problems, stop short of calling for paper receipts from electronic voting machines. Some advocacy groups are lobbying for "voter-verified paper ballots" that would create a physical audit trail to flag what could be a buggy computerized election machine.

"I find myself at the middle of a national debate that will quickly go global," Soaries said. "How do we secure electronic voting devices for the 28 (percent) or 29 percent of the population that will use them?"

Some 50 million Americans are expected to use e-voting machines in the November election.

It's unclear, however, what effect these recommendations will have. Soaries readily acknowledges that the commission he chairs has no authority to impose its views on state election officials, and he said he had not yet approached the other three members of the commission to seek their endorsement.

"Now is the time for computer scientists and election officials to get together and solve the problem," Soaries said.

His recommendations include:

The EAC should ask voting machine vendors to release the source code to states under nondisclosure agreements. Computer scientists in each state should be asked to sign the agreement and review the code.

An existing National Software Reference Library, operated by the Department of Commerce, should be expanded to include source code for voting machines. Using a technique such as a checksum, state officials would be able to verify that their machines are running the same code as the version in the library.

States should undertake "enhanced security measures" in November. One suggested option is cryptography, which is receiving favorable reviews from the computer science community.

Problems with electronic voting machines should be compiled and analyzed. No federal database of such glitches currently exists.

Linda Lamone, Maryland's election administrator, called the recommendations "terrific" in an interview after the dinner speech.

Maryland, which uses Diebold e-voting machines everywhere but Baltimore, already has access to the source code to the devices under a nondisclosure agreement, Lamone said. "We already follow some of the recommendations," she said.

Congress created the EAC as part of reforms enacted after the November 2000 Florida election debacle. The group is charged with certifying election hardware, doling out billions of dollars in grants to states, and "conducting studies and other activities to promote the effective administration of federal elections."

homepage: homepage: http://news.com.com/2100-1028-5229162.html

Thank Fucking God 10.Jun.2004 22:45

pix

And I mean it.....thanx.

useless! destroying the e-voting machines one of last points 11.Jun.2004 03:15

another view

Bullshit. This is hardly going to help. Sure their going to 'look at it' then they will simply rubberstamp the corporate criminals like they ahve been doing since this whole nonsense about e-voting began.

This is merely going to make more state actors complicit in the e-voting FRAUD!

And no one wants to talk about the whole issue is database fraud as 'instant updating' (instant vote fraud) without records. Who cares if the state wants to get in on the vote fraud as well? That's all this means.

I would destroy all e-voting machines on days of elections in the name of CIVIL DISOBEDIENCE and take the whole thing through the courts! The U.S. is a criminal state, period. Letting state criminals guard the corporate criminals is hardly a workable soluttion.

Come November, you can bet that THE WHOLE COUNTRY WILL BE ONE NATION UNDER INSTITUTIONALIZED VOTE FRAUD, THE VOTE WILL BE MEANINGLESS, AND THEN THEY WILL DAMN CLOSE IN ON EVERYONE.

Civil disobedience destroying the e-voting machines, which are illegal and criminal anyway, is well within a justifiable reaction, should anyone take up this suggestion.

It's a useless safeguard, here's why 11.Jun.2004 14:43

xyzzy

Electronic voting machines, like all computers, don't execute source code.
They execute machine code, either as a result of compiling (translating)
source code into machine code ahead of time or interpreting the source
code into machine instructions as it is run.

Unless one can verify the machine code -- either that of the compiler or
the interpreter, achieve is the case -- is valid, looking at the source
doesn't mean squat. Diebold could release one set of source code to
the states and install machine code built from another into its machines.

The only way to be sure voting machine vendors aren't playing hanky-panky
is COMPLETE oversight over ALL stages of hardware and software production
involved in these machines.

And that just solves the problem of willfully dishonest voting machines.
It's still possible to write and release code which contains accidental
errors. More than possible, in fact: it's certain. All non-trivial
software systems have bugs. Proving that a given non-trivial program is
"valid", i.e. produces correct output for all possible sets of input,
is one of computer science's unsolved (and, most likely, unsolvable)
problems.

There simply is no substitute for a voter-verified paper trail.

Portland