Reply #7: How Scanners work and how they can (and did in 2004) work against voters [View All]

[Kip Humphrey]

Scanners:

INTERPRETING THE PAPER BALLOT
Most vote scanner systems (some operate somewhat differently like the Mark-Right? system) operate by overlaying a data map over a scanned in bitmap of the ballot. Every bubble on the ballot has a defined area overlay (usually 1/2" square ((X,Y)(X,Y)) coordinates each which is fully configurable and MUST be configured to each ballot). If a program discerns the number of 0-0-0 pixels (black, but typically some configurable gray-scale range) is sufficient within the defined area (square), then a positive is registered in the program for that square area of the ballot.

If the scanner gets more than one positive from configurable contest groups (president, for example), the program would trigger the scanner to reject the ballot due to over-votes (entirely dependent how the scanner was configured). If the scanner detects no mark within a contest group, the program should reject the ballot or query the user due to an under-vote (again, this is entirely configurable).

The results of the scanner's configurable interpretation are logged (usually in a text file or table) which in turn is exported and loaded into a database, deriving a database data record (picture a row in a spreadsheet). The database program then tabulates the votes based on the interpreted, derived data.

Please Notice: the words "configurable", "program", "interpret", and "derive" in the above description. They represent points of vulnerability for vote tampering or just plain programmer/configurer error.

Check EIRS (Election Incident Reporting System) for optical scan machine-related incidents - there are lots of them.

Scanners, just like DREs, share the requirement of being configured for each ballot layout. Even without intentional fraud, the chance for human error propagated thousands of times just in configuring all of the different ballots in a county or state is HUGE!

HELPING THE DISABLED VOTE:
HAVA requires one device per precinct to accommodate voting with a disability. It does not specify the type of device.

Here's a FACT: there are no DRE's specifically designed to accommodate the disabled. That is not to say voting machines can't be designed specifically for specific disabilities (voice systems, braille readers and translators, ergonomically friendly levers, buttons and controls, etc.). And there are very good companies out there that have the experience and expertise to design effective disabled-specific ballot mark up equipment. The point is, if you or your state truly want to provide the disabled with an intervention-free voting experience, then provide them with machines they can work! Otherwise, the whole disabled discussion is simply a ploy by Die-bold and others to foist their crappy equipment on American voters at exorbitant prices.

If your state or county is stuck with machine contracts and can't or won't renegotiate them, then adapt the machines to mark up paper ballots (add a printer), count the paper ballots BY HAND, and use the elaborate vote processing system to provide quick, preliminary, unofficial vote totals for the press on election night.

NO MATTER WHAT: DO NOT SETTLE FOR HIDDEN BALLOTS OR SECRET COUNTING!

And ask yourself: What could a Kenneth Blackwell-type SoS do with the system to defraud voters?

Finally, answer this: What is simpler to get across to voters, politicians, and election officials? 6 paragraphs explaining how a scanner-based vote processing system works (and we haven't even gotten to tabulators!) or: HAND COUNTED PAPER BALLOTS?

If you want a, simple, direct, easy, reliable, familiar, and cheap voting system, hand count paper ballots.