BBV: Alameda County vote devices tested

[Bushfire]

Posted on Tue, Jul. 13, 2004

Alameda County vote devices tested

By Lisa Vorderbrueggen

CONTRA COSTA TIMES


Alameda County's touch-screen voting machines performed flawlessly during recent security tests, which should clear the path for their use in November, says county election chief Brad Clark.

<snip>

But in the meantime, he allowed counties to use the machines until the deadline if the vendors can demonstrate compliance with 23 security requirements.

<snip>

While Clark says he is pleased that the state appears likely to endorse (Diebold) Accuvote in time for the general election, it's only a short reprieve.

<snip>

"We're not sure what we're going do," Clark said. "We may have to buy new equipment or we may be able to modify what we have. We'll have to wait and see what the vendors come up with."

more...reg. req.

http://www.contracostatimes.com/mld/cctimes/news/local/states/california/counties/alameda_county/9141925.htm?1c

Does Clark know something Shelley will do before the rest of us?

[Trajan]

1. So: they pass a little memory march test ...

a nice checksum validation test ... and BANG ! .. EVERY thing is A-OK ..... RIGHT ? ...

WRONG ! ....

They can make any silly assed test that sits back and files its nails until it flashes a light and rings a bell ....

main()
{
int a;
char win = 'TEST PASSED > Bush Wins !';

for( a = 0 ; printf("*") ; a=+1 )

if( a > 10000 ) break;

printf(win);

)

[Bushfire]

2. Are you sure it was a check sum test, or one of those Logic(?) tests?

Nice that the author of the article wrote a complete puff piece without getting an opposing view, and didn't leave an email address for those who don't believe her or Brad Clark's drivel to follow up with questions.

[Trajan]

3. Oh hell ... I dont know WHICH 'tests' they did in fact use ...

but those particular tests are universal puter tests: a memory march test is a VERY basic memory validation test, performed at the machine level .... whereas checksum verification simply verifies that a specific program file loaded into a storage device matches the expected and known checksum of that specific program ...

As you can see: my knowledge of C is not only rusty, but the obsolete K&R version ....

My point is: developers of voting systems can develop WHATEVER logic verification processes they wish, and declare that their systems 'pass' those 'tests' ... It doesnt mean that each system is free of backdoor hooks or that fraudulent results cannot be engendered though 'untested' software functions that are physically unobservable without actually scanning a system's firmware and comparing it's structure with the expected source code structure ...

Furthermore: the use of modular storage devices which can be physically moved by individuals between systems, leaves open the opportunity to surreptitiously switch devices, much like ballot boxes being manipulated .... NO 'logic test' can preclude such an occurrence ...

The whole system, as implemented, STINKS .....