Rebuttal of DRE-only approaches to eVoting?

["David Webber \(XML\)" <david@drrw.info>]

Of interest and alot of good material.

http://www.openvotingconsortium.org/ad/shamos-rebuttal.pdf

I only found one flaw WRT to OASIS EML approach - page 2 -

"And cryptographic marks on electronically printed ballots can make box
 stuffing more difficult.19"

I view crypto marks / bar codes, et al on paper ballots as not
acceptable - as they open up a Pandora's Box of issues.  #1 is
compromising voter privacy - the voter has no idea what the
crypto contains - #2 - is ability to pervert scanning software -
that changes the paper vote scanning to match a changed
electronic record.  Therefore - any marks on the ballot that
cannot be human diagnosed and understood should be
avoided - everything has to be human verifiable by default.

To avoid ballot stuff is simpler - use embossed paper in
the printer - with serial #'s on it, along with a printed
ballotID - see IEEE on ballotID work for DRE with paper
ballot printing.

On to overall content - I would have made Item 2.5 Page 5
a main entry here - fundamentally this is what
makes voting different from a simple form-filling exercise,
and accounting methods needed to secure access to
the voting process (ability to cast a vote) so essential -
because once you are inside that "door" then you are
legally entitled to vote.   People need to "get it" right
away - so they see why giving vendors all the "keys" is
just craziness.  I like the analogy to gambling systems
security in that regard.

Enjoy, DW