[E-voting] intriguing possibilities using digial scanners

Marian Beddill beddill at nas.com
Sun Sep 11 02:43:34 IST 2005


Catherine et al;

We discussed digital images at a recent forum in my city, at which Bev 
Harris (of BBV) was one of the panelists.

In our state (Washington {Seattle}) somebody wrote into law many years ago 
that a computer "record" has the value of a "cast ballot".  I was astounded 
when I saw that, obviously the work of someone setting the stage for 
internet and DRE voting.

Thus such a digital image, being by any standard a "computer record", could 
be considered "the ballot".  We just barely escaped in recent years 
language which would have made that able to trump even the presence of a 
voter-verifiable paper ballot, but it still MAY BE used, and that 
permission is what allows paperless DRE's to be used.

That leads me to the review of the ability to tamper with such a digital 
image file.  Simple software may edit an image, such as the common TIFF or 
JPEG files, thus such a record could easily have the voter intent subverted 
by changing the file after the voter cast their ballot.  Such a subversion 
could happen at any place in the chain-of-custody of the image, and I 
believe that it would be very difficult to prevent or even detect such edits.

It is true that files may have "signatures" or tags like the older CRC 
number either internally or externally to the image file. But that itself 
only constitutes yet another "computer record" which itself could be 
subverted.  And the additional security management to verify such coding 
would be a burden on the election staff - subject to human failure allowing 
the images to be changed.

Marian Beddill


At 9/10/2005  04:14 PM, Catherine Ansbro wrote:

>Have a look at www.blackboxvoting.org - new items re: optical scanning.
>
>Finnish computer expert Harri Hursti has apparently developed some new 
>methodology using existing optical scanners which produce a digital 
>image.  (This doesn't relate to older scanners or ones that use infrared 
>technology.)  Apparently he's developed open source software.  In the USA 
>a scanned image of a ballot is considered to be a public record so at 
>least in theory people should be able to get a copy of all the ballot 
>images, and count them for themselves, with or without software.
>
>Could this be of interest here?  Snag--I don't think we have any similar 
>rules here relating to /images/ of ballots as public records.
>Catherine



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.stdlib.net/pipermail/e-voting/attachments/20050910/9bba4879/attachment.html


More information about the E-voting mailing list