Re: recounts in Colorado (under present law) -- The definition of insanity

[Ralph Shnelvar <ralphs@xxxxxxxxx>]

Dear Bo:

On Mon, 24 Nov 2003 12:23:54 -0700, you wrote:

>Ralph.....
>
>Is there no middle ground?

Nope.

>Everything is black or white?

Well, when it comes to voting it should be as close to black and white as
possible.  Isn't that the point?

>
>I'm not saying we should have Corporate control of voting software.

Frankly, I don't care if the voting software is private or public, open or
closed.  If the result of the operation is a paper ballot that can -
eventually - be hand counted then I really don't care any more than I care
if my Visa card is manufactured by company X or Y.

What I care about is that the transactions - my vote and everyone else's
vote - is accurately counted, anonymous, and auditable.

Any electronic counting system does not guarantee that.  If it goes in as
electrical signals then it can be programmatically manipulated.



True story: A few years ago, on a lark, I took a picture of my youngest
daughter.  I electronically cut out the image of my daughter from the
background, flipped the image left-for-right, changed the color of her shirt
from blue to red, and pasted the flipped picture back into another portion
of the picture.  It looked real cool.  There was my daughter and standing
next to her was a picture of her non-existant twin.

I called her into my office and showed her the picture.  "Who is that
standing next to you?"  "I don't know," she said.  "Isn't that you?"  "No."
"Why do you say that?"  "Because I don't own a shirt that color."

Honest.

Anyway, the point is that I could make a picture of my daughter and her
nonexistent twin and create that illusion electronically and easily. 

How hard do you think it would be to create phony images of ballots?

No, Bo, I want something that can be touched and not easily altered.

That doesn't mean that they can't be counted electronically.  It just means
that the hand count trumps.

>I'm saying that it *is* possible to make technology work for us.

Of course we can make technology work for us.  I am certainly not opposed to
technology.  I make my living off of technology.

Sitting on my desk are:
	5 computers
	3 monitors
	1 hand calculator
	1 notepad and pen

I trust computers to some extent.  As a Republican president said: Trust but
verify.

Hell, I don't trust the code that I write until it goes through extensive
testing and even then bugs show up years after the code has been running on
thousands of computers.

That's me trusting my code.  How do you think I feel about other people's
code especially when it comes to something as ephemeral as an anonymous
vote?

>I agree that "secret" software is ridiculous. (I also think it's mostly
>secret because only a programmer can read it.....but that programmer should
>be one of "us" not one of "them".....;-)
>
>Having a Corporation in charge of *anything* to do with elections is like
>having the fox guard the henhouse, and that's not the issue here. This is
>about whether electronics can be trusted to count votes.

The electronics can be trusted as far as a machine is more able to repeat an
action over and over again.

But as someone else has pointed out here: the machines do what they are told
to do.  At least some people have a conscience and so if there is a massive
conspiracy ...

>(I'm having problems with the whole concept of "Corporate" rights and
>privileges....but that's another story.)

Love to talk to you about it some day.

>I think there should be a way to place trusted (not sure how you can be
>satisfied of this, tho.......;-( individuals in charge of the process.

Bo, that's the point.  No one can be trusted and so the system has to be set
up so that there are multiple redundancies.

There's an old rule in business security: the person who writes the checks
should not be the person who balances the checkbook.  To allow the same
company/person to record the vote to be the same company/person that counts
the vote to be the same company that reports the vote is an invitation to
fraud and disaster.

I wouldn't allow my bookkeeper to balance my checkbook no matter how much I
trust them.

The parents of a well-known Libertarian did exactly that: the bookkeeper
took in the money and balanced the checkbook.  The bookkeeper skimmed
$100,000 off the top.  Bad, bad business practice.

I don't want that happening to the voters of Boulder, Colorado, or the
nation.

>Like
>the Clerk having a Programmer on the staff who deals with it. Do you trust
>the Clerk and their staff?

I trust them no more nor less than I trust any random stranger.  Have they
earned my trust?  

>Never sure with humans.......think Terry Baker
>and the SoS who died with all those registrations and papers in her
>office....;-(

Yes.  Imagine if Vicki Buckley had all of those ballot initiative signatures
(it was medical marijuana) on a computer disk instead of paper.  We would
have never known.

>
>I believe firmly that electronics is the only way to count such a massive
>amount of *anything* with little error.

Bo, it is an old bumper sticker joke: To err is human.  To really foul
things up requires a computer.

>(I'm not sure if I believe this can *ever* be accomplished perfectly and
>flawlessly.......nature of the Beast.)
>But machines *are* impartial and incredibly more proficient at this kind of
>task.

THey are much more proficient at repetitive tasks.

>We need to ensure the human interface is above reproach.......and there's
>the rub.
>
>Saying *all* electronic vote counting/sorting is bad, is throwing out the
>baby with the washwater.
>HUMANS are the ones to worry about.......
>HUMANS are the ones who have a motive for Fraud......
>HUMANS are the only ones who will *commit* fraud.....
>All the machine "cares" about, is doing it's job the best it can.....an
>admirable goal.
>
>If we find a way to remove the "human" factor, once the human's job (voting)
>is done, *then* we'll be able to have a truly confident election.
>At least that's what *I* think.......

You have identified the major glitch in computer voting: the human factor
when it comes to fraud.

I like things spread out.

If I were king of the voting world, the sequence of events would be:
  1) User uses touch screen to create a paper ballot.
  2) User validates paper ballot.
  3) User deposits paper ballot in ballot box at polling place.
  4) At end of day, paper ballots are scanned at polling place.
  5) The precinct count is recorded, publicized, and transmitted to
       county clerk's office
  6) The ballots are transferred to the County Clerk's office.
  7) The ballots are scanned again and matched against the values
       recorded at each polling place.
  8) The totals are recorded and publicized.
  9) The ballots are then hand counted over the next several days/weeks to
       validate that the machines where not in error.



Today I had lunch with a new friend: Evan Ravitz.  Both Evan and Al Kolwicz
have proposed - in one way or another - that votes be centrally counted.  I
want that, but I also want the votes "disbursed counted".  The more
redundancy we have and the more people we have doing the counting and the
more ways we have of doing counting, the happier I will be.

Trust but verify, verify, verify.

>
>Good thing it's a free country, eh?
><VBG>

Freedom is not free.

>
>Bo

Ralph

>----- Original Message -----
>From: "Ralph Shnelvar" <ralphs@xxxxxxxxx>
>To: "LPBC Chairman" <chairman@xxxxxxxxxxxxx>
>Cc: "Evan Daniel Ravitz" <evan@xxxxxxxx>; "Paul Tiger"
><Paul.Tiger@xxxxxxxxxxxx>; "BCV" <bcv@xxxxxxxxxxx>; "William Crook"
><slayer@xxxxxxxxx>
>Sent: Monday, November 24, 2003 11:49 AM
>Subject: Re: recounts in Colorado (under present law) -- The definition of
>insanity
>
>
>> Dear Bo:
>>
>> On Mon, 24 Nov 2003 00:47:44 -0700, you wrote:
>>
>> >Evan,
>> >Errors are errors.
>> >Secret ones and conspiracies and all are a bit paranoid.
>>
>> Really?
>>
>> With the results of elections controlling billions of dollars (think of
>the
>> recently defeated Proposition A), the more compelling it is to control
>those
>> results.
>>
>> >Stop being a Luddite.
>>
>> He's not the Luddite.  I am.
>>
>> Here's my background: I've been a computer programmer more-or-less
>> continuously since 1964.  For 10 years I did complex econometric modeling.
>> After that I was a consultant to companies like Exxon doing numerical
>> optimization (so that they could keep down their costs when moving oil
>> around).  Fascinating stuff.
>>
>> Since the late 1980's I've had a company that has done nothing but backup
>> and disaster recovery software.
>>
>> Now that you know how much of a Luddite I am, you can understand that I
>have
>> a technical understanding of both advanced statistics and computer
>> programming as well as a fairly good grasp of econometrics.
>>
>> And I'm terrified of these machines.
>>
>> >Technology can give us the ability to be error free as well as fraud
>proof..
>>
>> No it can't.
>>
>> By what argument can you make such an assertion?
>>
>> >We have to mold it to our purposes, not let it take over our process.
>> >This *is* possible.
>> >Just because something has been done for a long time is not a good reason
>to
>> >continue it. In fact, it's probably a good reason to *not* continue it.
>This
>> >is a dynamic world we live in and everything has to adapt. To not change
>is
>> >to stagnate and die.
>>
>> My next door neighbor has been an election judge for many, many years.
>> Elaborate mechanisms were put into place in order to minimize election
>> fraud.
>>
>> Those are good mechanisms.
>>
>> Bo, just because something is new does not make it better. Similarly, just
>> because something is new does not make it worse.
>>
>> Our freedoms hang in the balance.  To do something new with how our votes
>> are counted without proper and in-depth thought to the process can and
>will
>> lead to someone hijacking the process.
>>
>> If you want to hear about low-tech hijacking of elections just contact
>Sunny
>> Maynard of the Green Party.  She'll tell you some hair-raising stories.
>>
>>
>>
>> One of the questions people sometimes ask me is: We use credit cards and
>> electronic transactions every day to move trillions of dollars around.  Do
>> you really want to go back to barter and hand-written bank drafts?
>>
>> The answer: The basic difference between Electronic Funds Transfers (EFT)
>> and electronic voting is that voting is anonymous.  When an EFT goes
>astray,
>> someone gets hurt and yells.  When a vote goes astray then NOBODY KNOWS.
>>
>>
>>
>> With statistical sampling of the votes (a la HR2239) you can dramatically
>> reduce the probability of fraud.  But there are still problems.
>>
>> Consider a close election.  Let's say that polling data shows that your
>side
>> is losing by 6%.  Further, let us say that the stakes are really very
>high:
>> billions of dollars.  If you get caught then you go to jail.  If you don't
>> get caught then you end up being millions of dollars richer.
>>
>> There are plenty of people willing to make a much less lucrative a deal
>with
>> the devil.  Just ask your local pusher.
>>
>> So let's say that there are 300 polling places representing 30,000 votes
>and
>> that Company X controls the software.  If "your side" is low by 6% then
>the
>> vote would be 14100/15900.  To win you've got to skew 2000 votes.
>>
>> Company X knows that 1.5% of the polling places (5 polling places out of
>> 300) will be hand counted.
>>
>> If Company X knows which 5 places then manipulating the election is
>trivial.
>> I leave this as an exercise for the reader.
>>
>> If Company X does not know which 5 polling places will be hand counted,
>then
>> things are a bit more difficult but not impossible.  All that need be done
>> is that 30 polling places have their results heavily skewed.  Instead of
>> each polling place having a (average) of 47 yes v 53 no, you shift it to
>80
>> yes v 20 no for the 30 polling places that Company X is manipulating.
>>
>> Now the total vote is 15090.  Voila!  You've won and the chances of being
>> caught are less than 1 in 5.
>>
>>
>>
>> No, Bo, hand counting is necessary to preserve the honesty of the system.
>>
>> >
>> >Bo
>>
>> Ralph
>>
>>
>> >
>> >----- Original Message -----
>> >From: "Evan Daniel Ravitz" <evan@xxxxxxxx>
>> >To: "LPBC Chairman" <chairman@xxxxxxxxxxxxx>
>> >Cc: "Paul Tiger" <Paul.Tiger@xxxxxxxxxxxx>; "BCV" <bcv@xxxxxxxxxxx>;
>> >"William Crook" <slayer@xxxxxxxxx>
>> >Sent: Sunday, November 23, 2003 9:25 AM
>> >Subject: Re: recounts in Colorado (under present law) -- The definition
>of
>> >insanity
>> >
>> >
>> >>
>> >> Bo, Canada (20-something million people) and Britain both hand-count
>> >> ballots. It's been done for hundreds of years. The journey of 1000
>> >> miles begins with a single step, counting a million starts with one.
>> >> It's "scalable". I'm sure it's more error prone than computer
>> >> counting, but with the proper measures, those are honest errors
>> >> instead of secret ones which can never be discovered now.
>> >>
>> >> Evan
>> >>
>> >> On Sat, 22 Nov 2003, LPBC Chairman wrote:
>> >>
>> >> > Evan....
>> >> >
>> >> > Hand counting???
>> >> > a hundred thousand ballots???
>> >> > (in Boulder County alone....)
>> >> > And you think *that* would be more accurate???
>> >> > Maybe someday you should volunteer at the Clerks office and see what
>it
>> >> > takes to just open and sort ballots by hand......let alone count
>them.
>> >You
>> >> > would be amazed at the errors people make.
>> >> >
>> >> > We need to come up with a way to make technology work.
>> >> >
>> >> > Bo
>> >> >
>> >> > ----- Original Message -----
>> >> > From: "Paul Tiger" <tigerp@xxxxxxxxxxxx>
>> >> > To: "Evan Daniel Ravitz" <evan@xxxxxxxx>; <paul.tiger@xxxxxxxxxxxx>
>> >> > Cc: "BCV" <bcv@xxxxxxxxxxx>; "William Crook" <slayer@xxxxxxxxx>; "Bo
>> >> > Shaffer" <bo@xxxxxxxxxxxxx>
>> >> > Sent: Friday, November 21, 2003 6:46 PM
>> >> > Subject: RE: recounts in Colorado (under present law) -- The
>definition
>> >of
>> >> > insanity
>> >> >
>> >> >
>> >> > > No, Ben Franklin actually. Go do your homework.
>> >> > >
>> >> > > -----Original Message-----
>> >> > > From: Evan Daniel Ravitz [mailto:evan@xxxxxxxx]
>> >> > > Sent: Friday, November 21, 2003 1:18 PM
>> >> > > To: paul.tiger@xxxxxxxxxxxx
>> >> > > Cc: BCV; William Crook; Bo Shaffer
>> >> > > Subject: Re: recounts in Colorado (under present law) -- The
>> >definition of
>> >> > > insanity
>> >> > >
>> >> > >
>> >> > > On Fri, 21 Nov 2003, Paul Tiger wrote:
>> >> > >
>> >> > > > "The definition of insanity is doing the same thing over and over
>> >and
>> >> > > > expecting different results." -- Benjamin Franklin
>> >> > >
>> >> > > Einstein, actually. State law mandates recounts by the same method
>as
>> >> > > the original count. That's why I support hand-counting until the
>State
>> >> > > law is changed; public-source software isn't as good as
>hand-counting,
>> >> > > but much better than secret software used over and over.
>> >> > >
>> >> > > Evan
>> >> > >
>> >> > > ----------------------------------------------
>> >> > > Evan Ravitz     303 440 6838     evan@xxxxxxxx
>> >> > > Vote for the National Initiative! www.vote.org
>> >> > > Photo Adventures:          www.vote.org/photos
>> >> > >
>> >> > > Kucinich: the ONLY candidate to vote against the
>> >> > > "Patriot" Act and the Iraq war:  www.kucinich.us
>> >> > > ------------------------------------------------
>> >> > >
>> >> > >
>> >> > >
>> >> >
>> >> >
>> >>
>> >>
>> >
>> >
>>
>>
>>
>
>
>