emailaddr.jpg










About

This page contains a single entry from the blog posted on December 14, 2006 10:28 PM.

The previous post in this blog was Le weekend.

The next post in this blog is Out of the Mouths of Babes.

Many more can be found on the main index page or by looking through the archives.

Blogs We Read

Creative Commons License
This weblog is licensed under a Creative Commons License.
Powered by
Movable Type 3.33

« Le weekend | Main | Out of the Mouths of Babes »

Trusting Software

Software has its limitations:

A technical committee that advises the U.S. Election Assistance Commission agreed unanimously that all new voting machines should be able to be audited either via paper trails or some new technology.

It’s a decision I agree with. It reminded me of working an aircraft simulator issue with Military Airlift Command (no longer in existence). Their requirements officer wanted to order a motion simulator for a cargo aircraft that had a range of freedom exceeding the space available in their facilities.

“Just software limit the simulator’s range of motion until we get the buildings modified” he said.

No can do. If the mechanical arm can reach through the ceiling we weren’t going to rely on software to keep it from doing so. The solution was to make sure it was physically impossible to hit the building.

Paper back-up for voting machines is the equivalent concept; we shouldn't have to rely solely on software when we don't have to.

TrackBack

TrackBack URL for this entry:
http://www.uncorrelated.com/cgi-bin/mt-tb.cgi/1275

Comments (2)

Mick Stockinger:

Good point.

Though having done a lot of machine design in my time, there is another aspect of this problem to consider--safety and functionality are two design goals that often butt heads.

It was generally true that the safety aspects of machine design would take a multiple of time dedicated to actually designing the machine to perform its function properly.

Voting machines have as their design objective the creation of a less ambiguous voting process and more flexibility for voter diversity (language, disability, etc...). Making it perfectly secure doesn't appear to have been part of the original design specification.

Diebold gets all the blame, but the "customer" wrote the spec.

I've been musing about the problem for the last year or so, and thought of a different approach--give a copy of the ballot back to the voter.

The reality of the electronic voting machine is that its insecurity isn't really a software or hardware problem--its a logistics problem. In a world where everyone is perfectly honest, the machine would only present the kind of glitches we saw in the past election--problems with the configuration user-interface. In other words--idiots.

That's not terribly hard to work out, and with new equipment design, experience is how you get to version 2.0

No--the real problem is that we don't live in a world of honest people and so the logistics of transporting machines, configuring machines and evaluating output constitute the real problem.

So how does a paper printout help?

Its still in the possession of the perceived corruptible election officials. If they can game the machines, they can figure out how to game the ballots as well.

What is required is a double-blind system. Just as you don't have the same person conduct purchasing and invoicing functions in a company, you don't have election officials count both the ballots and the backups. Give a record of the vote to the voter in the form of a tamper-proof record, like a CD-ROM (hopefully a lot smaller). In the event of a controversy, the public has an accurate record of how they voted and can re-record their votes in the event of a dispute.

The paper ballot isn't going to help in the case of a corrupted election official, but it will help with the corrupt programer. I liked seeing my ballot backed up when I voted. If there was a question about the vote that tape will come in handy. Our country has had to deal with corrupt election officials from time to time for the past couple of hundred years. On the whole, we've dealt with it mostly O.K. (as Hugh Hewitt says "if it's not close, they can't cheat). Electronic voting, however, is a newer animal. Done right, I expect more accurate results but people must have confidence in the voting equipment.

Ultimately we need mostly honest people for our system to work and if we don't, inaccurate voting machines may be the least of our problems.

Post a comment

(This site no longer requires authentication for unmoderated comments to be posted immediately. Simply enter your comment with a valid email address and type the challenge word into the field below before posting. UNCoRRELATED accepts no editorial responsibility for the comments posted here, but will by discretion, remove vulgar, abusive or commercially-motivated comments. You may receive email notification of follow-up comment by clicking on the Subscribe to this entry checkbox.)





Tom-Mannis.jpg thinkingblogger.jpg









Google PageRank 
Checker - Page Rank Calculator

Blogroll Me!

Powered by FeedBurner

Add to Google Reader or Homepage

Subscribe in NewsGator Online

Subscribe in Rojo

Add UNCoRRELATED to Newsburst from CNET News.com

Add to My AOL

Subscribe in FeedLounge

Add to netvibes

Subscribe in Bloglines

Add to The Free Dictionary

Add to The Free Dictionary

Add to Plusmo

Subscribe in NewsAlloy

Add to Excite MIX

Add to netomat Hub

Add to Webwag

Add UNCoRRELATED to ODEO

Subscribe in podnova

Add to Pageflakes