In a good first step to enhancing the confidence of voters in their vote counting equipment, Sequoia corporation announced that they would be revealing their source code.

From their press release, I’m unclear as to what exactly this means.  There is software to run the optical scan machines, there is software to run the DRE machines,  there is software to design the ballots, and software to tabulate the votes.  I’m assuming and hoping that all will be released.

Citizens should expect a host of computer experts who will be critical of the Sequoia software.  There is no computer code that is going to please everyone.  Here are some of my thoughts about what to look for in this move.

First, when citizens hear about so called security issues, they should be asking whether these are real issues, or if they are like some of the vulnerabilities indicated in the EVEREST report in Ohio.  One of the “security” issues identified there was based on someone pulling a memory card out of a tabulator at a polling place and manipulating it in the presence of voters and election judges, reinserting it, and then hoping that no one noticed that the number of people who voted didn’t match the number of ballots counted on the tabulator.  It is important that any security flaws identified be classified by their feasibility and by their potential impact.   Security experts will be doing a disservice to the process by hyping problems that really aren’t problems.

I’m very happy that the code is being released prior to certification.  One of the concerns that I have had with vote counting equipment is that flaws in the design can’t be changed without recertification.  That is costly, and generally isn’t going to be done by a company without a glaring need.  Instead, we are forced to work around those issues and test and audit to make sure they are addressed.  By allowing the public to see the logic in the software, these issues can be resolved early.

I believe one of the results of this will be for proponents of IRV to propose code that will allow the machines to accommodate that method of voting.  This will be one of the very positive steps as I believe IRV and similar methods of voting will be gaining momentum.

I’ll be interested in the specs put out on the new Sequoia Equipment.  Hopefully there are a number of statistical items that the new equipment can track.  Despite its overall deficiencies, the core goal of the book The Democracy Index is the aggregation of data.  This new software can hopefully help.  For example, I’d like to see an actual log of the error messages on the machine (which I believe is one of the data elements Gerken is looking for in her book).

In terms of ballot design, I’d hope that the new software will help jurisdictions meet some of the design guidelines set out by the Brennan Center.  Our ballot design software works well, but could be better.
If Sequoia is actually going to listen, and it appears they will, the publication of their source code prior to certification is going to turn their new equipment into the industry leader.

Trackback URI | Comments RSS