[ale] Its over. Maybe

[Jeff Hubbs]

On Wed, 2004-11-03 at 20:55, Jim Popovitch wrote:
> On Wed, 2004-11-03 at 20:31 -0500, Jeff Hubbs wrote:
> > The two concepts are not in any way, shape, or form analogous to one
> > another.  
> 
> Yes they are.  Both are machines, both count/control critical functions,
> both produce data that is analyzed by an external device.

Are you blind to the notion that on one hand you have a computer that is
used to determine how much fuel to shoot into a car engine's cylinders
and the other you have a massive system of computers whose combined
output is used to decide who the President of the United States is? 
Perhaps if the role of the computers in our cars expanded to, say, guide
the car itself and were all under central control, you'd have a better
analogy, because then you'd have a situation where the computers could
be commanded to simply make everyone stop right where they were, or to
even slam all the cars into each other.  If you do not think that a
single person or organization can build and command machines to kill
tens of thousands of people at once, I direct your attention to two
since-rebuilt cities in Japan.  

> 
> > The Diebold machines hide the voting process from their users
> > and even their operators, 
> 
> No more so than a calculator hides the process by which mathematical
> calculations are achieved.  One could argue that the Diebold box does
> even less than a calculator.

Utterly wrong.  A calculator can be subjected to arbitrary tests.  A
calculator's results can be independently verified, even by hand if
desired.  Everyone's calculator is not inscrutably connected with every
other calculator (yet).  My worst-case scenario for the Diebold machines
is that all they really do is have the outward appearance of a voting
machine but actually pass no data on for tabulation; their controllers
simply report whatever numbers they wish.  

> 
> > putting the company and parties unknown into a position to subvert
> >  the election process with little fear of detection.
> 
> Sure, if you want to believe that.  I don't. That is just too extreme
> for me to think that a company would willingly produce something that
> could ruin itself.  You are just grasping for straws by suggesting such
> a possibility.

History is full of such companies and in each case, the company hoped
that secrecy and misinformation would be the key to survival.  Such
companies depend on a never-ending supply of incurious, uncritical,
passive, and blindly trusting people.  

> 
> > OSS voting machines and the means to prove that the OSS in question is
> > really running on the machines is the only way to computerize the
> > process with integrity.
> 
> What about the hardware and the firmware that the OSS runs on?  If
> Diebold's black-box solution won't satisfy you how can some OSS app on
> some other black-box?

It wouldn't.  My hope is that any means used to count votes uses a
process that is transparent to enough people that it would be extremely
unlikely that all such people could or would conspire to subterfuge. 
Any process that is too complex or too opaque for this should never be
adopted.