New York State Voting Machine Certification

Hurry up and Wait

With the submission of voting machines to the New York State Board Of Elections last week the certification process for new voting systems has officially begun. Certified systems will be first used in September 2007, when New York State is scheduled to replace it’s current lever voting machines (for details on New York’s plans see my earlier post “New York State’s HAVA Plan”).

But much about the certification process is still undefined, and an exact date when new voting systems will be certified for use in New York is still hard to determine. It could be as early as this October, or as late as January or February 2007! And in their haste to move the ordering process along, the State Board of Elections may ask county election commissioners to indicate their choice of voting machine before any machines have completed certification testing, a move guaranteed to make New York voting rights activists see red.

Three Dates and a Schedule

At this point there are only three firm dates, two of which have already passed:

July 5, 2006 - Voting machine vendors submitted applications and $5,000 application fees to the State Board for each voting system they want certified. A total of 11 systems, both DRE and precinct ballot scanners, have been submitted by six vendors(details here).

July 20, 2006 - Machines and technical data packages (specifications, analyses, and reports required by state regulations) were delivered to the State Board.

August 15, 2006 – The State Board of Elections must submit its detailed 2007 schedule to the Department of Justice. This schedule, a requirement of the now settled DOJ lawsuit should lay out the important milestone dates along the way to September 2007, presumably including dates when counties must place orders for new systems.

Ciber Certification

Ciber Inc., one of the big three “Independent Testing Authorities”, or ITAs, has been contracted by New York to oversee the certification process. This includes creating test plans for each system, and conducting and managing the actual testing. History indicates that New Yorkers should have some concerns about this vendor choice.

The so called “Independent Testing Authorities” (see “The Dirty Little Secrets of Voting System Testing Labs”) haven’t done a very good job when it comes to finding problems with electronic voting systems. Ciber is one of the companies which tested and approved the flawed Diebold Accuvote TSX machines which have been revealed to have enormous security holes.

This gaping hole has been described as “the most severe security flaw ever discovered in a voting system”. But yet, the flawed DRE was given a clean bill of health by Ciber. How rigorous a test can we expect in New York?

Next Step – Test Plans

A test plan specifies the full suite and scope of testing for each submitted system. It is the critical blueprint for the certification tests the system will undergo, describing which tests will be performed and which will not.

Ciber anticipates it will take approximately 30 days to create each test plan. Given that there are 11 separate systems that require plans, an important question is how many can be developed simultaneously? Statements made at a State Board of Elections meeting indicate that Ciber may only be able to complete 3 test plans at a time. And that begs the question, which three systems will be “blessed” to be the first three to have their test plans created?

Why is this important? Because once the test plan for a system is complete, the actual testing will presumably begin promptly. If there is also a limit to how many systems can be tested simultaneously, the first machines to be certified will be the ones which have the test plans completed first, giving an unfair competitive advantage to these first three systems!

Timing Studies

It's vital to know how many voters can be served on Election Day by a single voting machine. This figure is crucial because it determines how many machines must be available in each polling place, and consequently, how many new systems must be purchased for each lever machine being replaced. In New York, a single lever machine currently serves up to 800 voters, and no more.

But there is increasing awareness that touch screen voting machines may only be able to serve 400 voters or less! If it takes substantially longer to vote on a DRE than it does on a lever machine, as is likely to be the case, then the replacement ratio of DREs to lever machines may be 2 to 1, or higher. That is, for each lever machine in a polling place, two or more DREs must be purchased to replace it. Given that full face DREs offered in New York cost $8,000 and up, and there are over 19,000 lever machines to replace, the cost of acquiring DRE technology in New York, already huge, could go through the ceiling.

If more than one DRE is needed to replace each lever, the extremely high costs of acquiring and maintaining touch screen voting machines will far exceed available HAVA funds and force cash strapped counties to come up with the additional funds out of their own budgets!

In order to determine this critical replacement ratio, the State Board plans to do what they're calling a "usability" study on each system. The results of this study will determine the ratio of new systems to lever machines. But as of this writing, no organization has been contracted to do this study, it has not yet been put out to bid, and it’s anybody’s guess when it will be completed.

Since both election officials and concerned citizens consider this to be an essential ratio to have defined prior to ordering new systems, this timing study is what project managers call “critical path”. In other words, it must be completed BEFORE machine orders can be placed.

Certification Testing

When the test plans are completed, testing can begin, at least in theory. But still to be determined is the location of the testing, how thorough the testing will actually be, and how long it will take. None of these are yet known, and citizens are left guessing about when and how this all plays out.

And in another wrinkle, the State Board of Elections has been testing the waters with a strange and disturbing idea – New York counties will be asked to indicate their machine choice BEFORE the certification process is complete. Yes, you read that right. The Board has proposed that counties state which machine they intend to purchase prior to any machines actually being certified. This would circumvent much of the good that derives from a thorough, open, and independent certification process. It gives the appearance that the testing process is just for show, but irrelevant to the actual voting system choice. For voting rights activists, this is simply unacceptable.

And, as noted above, a key component of ensuring an uncompromised voting systems selection process is that all systems be certified simultaneously, and that no systems are certified before others. If some systems are “blessed” in this way, this would have the disastrous effect of favoring both the vendors and the system types that are certified first. This would bias the selection process. Again, simply unacceptable. The replacement of the state’s lever machines should be based on the merits of each system, not the order of their availability.

Touch Screen Voting and the Space Shuttle

High-tech promise, high-tech failure

After a long hiatus the Space Shuttle is flying again. But it’s now clear that the Shuttle program has been too costly, hasn’t lived up to expectations, and is headed for the junk pile - an ultra-expensive, high-tech solution which has failed to deliver on its promise. Election officials rushing to purchase touch screen voting machines should consider how the promise of new high-tech solutions often results in an expensive high-tech failure.

Most people assume that when it comes to technology, new is always better than old. But an "old" technology can also be a mature, proven, and reliable technology. The Shuttle was designed to replace the “old” technology of expendable rockets. Promoters of touch screen voting machines intend to replace the “old” technology of paper ballots.

But an “old" technology is not necessarily a bad technology. Just because a given technology is newer (and more complex) doesn't necessarily make it better, more reliable, or cost effective. In fact, when a new technology is more complicated than necessary for a given task, its use can be a step backwards in terms of reliability and cost.

We see this with the Space Shuttle. The Shuttle is a newer technology capable of doing things that can't be done using expendable rockets (e.g., in-orbit repair missions), but that does not necessarily make it the appropriate technology for all missions. If all you need to do is to put a satellite into orbit, an "old technology" rocket is a much more reliable, safe, and cost effective solution. And while a touch screen with an audio interface may be needed to enable persons with specific disabilities to vote (e.g., the blind and those with certain mobility impairments), it is technological overkill for the vast majority of voters. And, just like the Shuttle, touch screen voting machines will prove to be a step backwards in safety, reliability, and cost.

When the Shuttle program was announced, its goal was to make "expensive" expendable rockets obsolete for Earth orbital missions. It was obvious, we were told, that the Shuttle was going to be cheaper because it was reusable and minimized the number of disposable parts. The argument was "we won't be dealing with all those expensive expendable rockets”. Today we hear an eerie echo in arguments for touch screen voting machines – “we won't be dealing with all those expensive printed paper ballots".

But the Shuttle has not proven to be cheaper or more reliable than "old” rocket technology. In fact, just the opposite has occurred. The Shuttle is a much more complex system than required for most missions, and the ongoing costs have been much higher than anyone anticipated. The Shuttle technology was oversold, just as touch screen voting technology is being oversold.

Complex, high-tech solutions like the Shuttle and touch screen voting machines should be reserved for those tasks where additional complexity is necessary. For routine jobs, like putting a satellite into orbit, or filling out a ballot, tried and true "old technologies" are more appropriate, reliable, and cost effective.