Re: 12/12 Daily: County to vote on voting system finalist

[Paul Walmsley <paul@xxxxxxxxxxx>]

Hello Richard and fellow listmembers,

It's worth commenting on some of Mr. McClure's assertions before
the Citizens Review Committee meets tomorrow.  The quotes below are from 
Richard Valenty's story in the Colorado Daily tomorrow, kindly forwarded 
by Doug, two hours before print publication :-)

On Thu, 11 Dec 2003, Doug Grinbergs wrote:

> "Since when did paper become a tamper-proof media? I could print ballots
> on my $175 home printer that would look like real ballots," said
> McClure. "If a group says that tampering and election fraud has been a
> way of life in this country, you've just put a media that everyone is
> accustomed to right back into their hands."

It's not clear why Neal McClure believes that his ability to print out
fraudulent paper ballots at home would negatively affect the security of a
paper-ballot election.

Perhaps he is asserting that he could print out hundreds of fraudulent
ballots at home and then substitute them for genuine ballots at some point
in the voting process, or stuff the ballot box at the polling place.  
Since this problem is well-known, as a society, we've developed some good
safeguards.  We staff polling places with sworn election judges to prevent
one person from stuffing many ballots inside the ballot box when they come
to vote.  We seal ballot boxes after the polls close so the boxes cannot
be tampered with enroute to the vote counting center.  Citizens and
election judges are able to monitor the vote counting process, and verify
by sight, if they so choose, that the count recorded by an optical scanner
matches what is recorded on the paper ballots.  

These security measures are fairly easy to understand -- unlike the 
security measures required to ensure that Mr. McClure's DRE machines are 
working correctly.  For malicious or buggy computer programs running in 
his DRE machines could change citizens' votes at any point in the voting 
process.

> "We store the votes in three physically separate memory locations.
> There's a primary path from what we call a 'mobile ballot box,' which is
> just a flash memory card. We provide a software application that will
> collect and store the alternative storage of those electronic ballots,
> and allow you to compare that with the primary data path," said McClure.

Mr. McClure's "three separate copies" do no good if his software records
the wrong vote immediately before those three separate copies are stored.  
In that case, the wrong vote would be stored in three separate places.  
Triple redundancy ... for the wrong information!  Any recount would merely 
result in the same incorrect vote tally.

With DREs such as Mr. McClure's, the voter has absolutely no way of
ensuring that the machines did not corrupt their vote.

On the other hand, in an election using voter-verifiable paper ballots,
the voter can carefully review their ballot before depositing it in the
ballot box.  And the paper is always there for a recount.

> CVV is currently attempting to get Boulder County to delay purchasing a
> new system. According to a position statement from the CVV Web site,
> http://coloradovoter.net, technical standards for voting systems "have
> not yet been developed by National Institute for Standards and
> Technology (NIST) as required under the Help America Vote Act (HAVA)."
> 
> McClure called the Daily from Washington, D.C. on Thursday, where he was
> attending a NIST conference titled "Building Trust and Confidence in
> Voting Systems." According to McClure, standards do in fact exist
> despite the claims of CVV and other nationwide groups.
> 
> "There are standards in existence, the Federal Election Commission (FEC)
> standards of 2002," said McClure. "In the HAVA, it's required NIST to
> manage the voting system certification process. The NIST people have
> been involved with the National Association of State Election Directors
> (NASED).
> 
> "Voting systems are now being built to a set of standards. To think that
> NIST is going to come up with a set of standards that are completely and
> absolutely different is inaccurate. What NIST is looking at doing is
> taking FEC standards and doing some incremental improvements on them,"
> said McClure.

The issue is not that "some" standards do not exist -- it is that the NIST 
standards do not exist.  Why is this so important?  One reason is because 
the Colorado Secretary of State's election rules specifically require DREs 
purchased or leased in Colorado to conform with NIST's standards.

No one, not I nor Mr. McClure, know what these standards will be, since
they do not yet exist.  So CVV's position in this regard -- that it is
unwise to purchase any systems before NIST releases their standards -- is
actually an election rule, promulgated by the Colorado Secretary of State.

(Colorado's election rules are at 
<http://www.sos.state.co.us/pubs/rule_making/current_rules/electionrules.pdf> 
- in particular, see rule 34.2.)   

(As an aside, CVV has never claimed that standards do not exist.  In fact,
CVV's website has had a link to the FEC standards mentioned by Mr. 
McClure for several weeks - see 
<http://www.coloradovoter.net/moin.cgi/ElectionLaws>.)

> "I'm the project manager for the IEEE voting systems standards
> committee," said McClure. "There's a federal law that says if there's a
> commercial standard, a government entity must adopt it. I will tell you
> that the IEEE standards have taken FEC standards as their starting
> point."

I challenge Mr. McClure to back up his claim that "there's a federal law
that says if there's a commercial standard, a government entity must adopt
it."

> "What happened then is that companies and people got organized and
> addressed the issue. At the stroke of midnight, planes didn't fall out
> of the sky and financial markets didn't crash. As engineers, we have to
> say 'Sure, problems are possible, but are they likely?' I'm confident in
> the accuracy and integrity of an election conducted on eSlate," said
> McClure.

As many news stories have demonstrated, problems with DRE machines are 
widespread.  We should not rely on a vendor's word that their machines are 
trustworthy.  Rather, we must require that any machine that a voter uses 
must not record the vote directly - rather, it must generate a full-text, 
voter-verifiable, paper ballot.  Anything less is disrespectful to all 
those who have fought and died to give us the democracy we have today.  


- Paul