Conformity Magazine - June 2008 - (Page 30)

gas tank but forgetting to connect the gas line to the engine. A full gas tank unconnected to an engine does little good. One way to analyze the election system is that you traditionally have people, paper and processes. That combination has a long history and documented problems with accuracy and fraud. To improve on that situation, computers, software and automation were introduced. However, computers and software have their own vulnerabilities. The right solution is to find the optimum combination that simultaneously: 1) reduces human error; 2) protects against all forms of election fraud; and 3) increases the total system accuracy and security. Setting up an either/or scenario simply offers a choice between two sets of problems. A hybrid system could potentially provide safeguards against the problems of each approach, and result in a much better total outcome. It is now becoming generally recognized that, as originally proposed, the voter verified paper trail was deeply flawed, and that its implementation, unguided by carefully specified requirements, has created even more problems. The concept of voter verification is sound. What was missing in the original concept was a strategy for assuring that the ballot the voter verified actually made it into the final tally. The proponents also failed to develop a rigorous set of specifications to guide a coherent system design. It was also not generally recognized that the paper ballot was really a metaphor for a durable, tamper-proof record of the voter’s ballot. A serious debate was never held asking simple questions, like “Is paper the best medium for a durable record?”, and “What is the most tamper-proof type of record we can create?” Despite great sincerity and enthusiasm, the verified paper ballot systems currently in use are widely being rejected as an approach that no one wants. Impact of Recent Changes Improvements The changes made over the last eight years have been fundamental and far-ranging. For the first time, the federal government has taken responsibility for setting national standards and certifying voting systems. NIST NVLAP has responsibility for accrediting the test laboratories. State and local election officials have introduced a variety of significant improvements. The system for certifying voting systems has been profoundly improved. Partial Completions A problem that the system currently faces is determining what further changes might still be needed. It was widely recognized, while the 1990 standards were pivotal as the first national standards for voting equipment, by 2000 they had become dated and were in need of revision. The 2002 revision of the VSS added approximately 40% more requirements and strengthened many existing requirements. The 2005 VVSG added over 100 new requirements, taking the total number of testable requirements to nearly 1000. 30 Conformity JUnE 2008 Today, equipment qualified to the 2002 standard has only been used in a few election cycles, and no equipment has been certified to the 2005 standard. So while there are many opinions on how the current 2005 VVSG might be improved, there is no field experience to guide further changes. Repeatedly investigations of field problems with voting systems have shown that requirements were in place in the 2002 standard as well as in the 2005 standards. A common root cause when investigating various field incidents is that flaws get through the testing process, since there are no standard test methods. Each VSTL is responsible for developing their own test methods, and that variability allows for some differences in how systems are tested. The common perception that we need tougher standards actually draws our attention away from what would really do the most good, that is, more effective test methods for the requirements that already exist. Getting the VSTLs to the place where there is good lab-to-lab repeatability has yet to be seriously addressed. NIST NVLAP is just beginning its second visit to the accredited laboratories. Just how much of the variation in test methods between labs will be improved over time is not yet clear. It is extraordinarily difficult to determine how much improvement will occur naturally over time and where further work will be needed. Unintended Consequences One issue that is becoming very clear is that there are unintended consequences from so many changes being made by so many relatively independent agents. The cost of getting a voting system certified at the national level has risen dramatically. In 2000, qualifying a system to the 1990 VSS cost a vendor typically $100-250K in laboratory fees. With the 2002 VSS, that cost rose to between $500K and $1M. No system has yet successfully been qualified under the EAC program or to VVSG 2005, but the final cost for systems in process appears to be $1-2.5M. It is estimated that testing to the draft revision proposed by NIST will require somewhere between $4-6M. With certification costs that high, a number of consequences roll out. First, new companies, especially smaller companies are locked out. To get into the business today, a company would need to fund the development of a system, pay to have it nationally certified, then pay for state certification in the states where it intends to market. Then and only then could a company market their system and have the chance to generate revenue, making the risk of entering the business enormous. Another consequence is that established vendors are taking through fewer versions for certification. Changes are being grouped, and going for national certification is a very carefully considered event. This means that some modifications that are much needed by election officials are being held off because of the cost and time it takes to certify a system.

Table of Contents Feed for the Digital Edition of Conformity - June 2008

Conformity - June 2008 Contents Editor's Note 700 MHz Auction Raises a Record $20 Billion FCC Modifies PCS and AWS Power Limit Rules for Broadband Wireless Commission Designates Spectrum Test Bed FCC Releases Report on Wireless Competition Commission Proposes $5 Million Fine for Slamming Using EDX for Non-Destructive Detection of Lead in Consumer Products ESD Open Forum Challenges in Testing - Improving Election Security and Accuracy Part 1 Challenges in Testing - Human Body Model: The Hidden Challenges High-Speed Signal Integrity Considerations for ESD Components Focus On...EMC Components Buyer's Guide FCC Levies $2.6 Million Fine for Junk Faxes FDA Updates Guidance on its Product Review Process EU Issues Updated Energy Star Regulations EU Repeals Directive on Veterinary Electro-Medical Equipment New Product Announcements Updated Standards List for the EU's PPE Directive CPSC Announces New Effort to Keep Out Hazardous Products CPSC Actions in the News IEC Standards Update UL Standards Update Product Reviews Telcordia Standards Update From Our "You Can't Make This Stuff Up" Department Looking Back: Items from Past Issues of Conformity Advertisers

Conformity - June 2008

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.