Re: [tlug] Somewhat OT- open source software for US voting machines

[Edward Middleton <emiddleton@example.com>]

Curt Sampson wrote:
> On 2008-09-30 17:54 +0900 (Tue), Edward Middleton wrote:
>
> > To prove this he would have to show that it is not possible to
> > implement an e-voting systems that is more effective then the current
> > manual methods.
>
> I'm dubious about arguments along the lines of, "it must be possible
> unless you can prove that it's not."

My point was that the current system is imperfect so arguments of the 
sort that e-voting isn't perfect are meaningless.  The issue is not 
whether it is possible to build a perfect e-voting system but whether it 
possible to produce a better system then what we currently have.  All 
software development faces problems, to prove his argument he needs show 
that the challenges in building a better e-voting system are 
insurmountably not just beyond his ability to solve.

> And in this particular case, I'm more inclined to accept the conclusions
> of the guy who spent three years of his life implementing e-voting
> systems.

Well if I were going to create a successful e-voting system I would not 
have approached it in the way this guy did in GNU.FREE[1].  To be honest 
I can't see any way an internet based voting systems can be anywhere 
near as effective as a paper ballot.  The fact that this wasn't 
immediately apparent to this guy is something that makes me question his 
conclusions.

In order to run free and fair elections at a minimum

1. you need to have universal access to voting.

2. you need to be able to insure that voters can cast their ballots freely.

How can you possibly do this as effectively with a web based system?

If I were going to build an e-voting system I would have approached it 
more like this[2].  I don't think this process was perfect and there are 
obviously vectors for attack but this is a more realistic starting point 
to look at the viability of e-voting systems.

> > The problem is that his arguments are based solely on the premise that
> > all problems relating to the voting process can/should be solved by
> > technical means. This is flawed whether you used electronic or manual
> > voting procedures.
>
> If you consider things such as how you use a ballot and how you count
> the vote as "technical means," which if you're using that phrase in the
> context of "manual voting procedures" I would guess you do, I'm not
> clear on how there would really be any other way of solving problems.
>
> In other words, what are the "non-technical means" that are being
> applied to the non-electronic systems that could be, but are not being,
> applied to the electronic systems?

Well the non-technical issues I was referring to were organizational 
issues, the processes and procedures, the legal and political issues.

Hardware and software were generally what I would refer to as the 
technical side.  You could also say that a voting slip was the technical 
means in the case of manual voting.

If you take the example of "universal access to voting", making voting 
compulsory (a non-technical means) is an extremely effective way of 
insuring this.

Edward

1. http://www.free-project.org/download/
2. http://www.wired.com/techbiz/media/news/2003/11/61045