Re: New AVTS units - Card reader summary of discussion

If we modify the new Accu-Vote TS to accept a locking card reader we are faced with the following issues:

1. There must be a major design change made, therefore setting back the final design acceptance, and moving back the date when we can begin production.

This appears to be the crux of the argument. I understand the time constraints and have wanted the new product as much as anyone, however, we must get appropriate market input before proceeding. A software solution for not having the time to find the right piece of hardware might not be best for our future.

2. A decision must be made on the type of locking device, i.e. does it completely take the card inside of the reader so that no one could “accidentally” remove the card while voting as happens now, or do we continue down the path of using a card reader like we currently have.

3. The more complex we make the card reader and with the more moving parts, the increase in failure and the replacement of readers must be anticipated.

Perhaps some of you saw Phil Foster, Sequoia Pacific, demonstrating their touchscreen device Thursday on the Bryant Gumbal morning show. Foster got about 8 minutes or so to show S/P's punchcard, op scan, DRE and TS equipment. The on air reporter just gushed over the TS technology. She called it the system of the future... They showed the smartcard, and made a point on security as their unit nicely (and securely) sucked the smartcard completely into the device. After the voting demo was over, Gumbal raised the obligatory security/anonymity of the voter issues. Foster held up the smartcard and gave the appropriate answers. They asked him about the Internet and Foster blasted it based upon security and voter reg issues. It was a pretty good segment.

I'm not suggesting we must suck the card completely into the TS unit just because S/P does. However, I am reminding this forum that S/P is our main competitor in TS voting. They have the national track record with their DRE Advantage system and have brought a TS unit to market that is self contained in a booth, with smartcard technology and the same size screen we utilize. Additionally, as Foster pointed out, they have the huge Riverside reference. They secure their smartcard via the reader, not via software. They seem to have found (over a year ago) a reader that is slim, secure, and appears to work. I don't know if they have "moving parts" problems. I do know their unit works.

Installing a locking card reader would require redesign of the plastics in the area of the card reader mounting. For a card reader to lock, the card must be inserted into a reader that encompasses the end of the card to allowing a locking mechanism to block the end of card from traveling back. As you know, our current design allows the card to stick out approx. an inch from the front face of the unit.

Sticking out an inch is really not the issue. I'm not suggesting we must completely swallow the card, or even 2/3s of the card. I am suggesting we must "hold" on to the card in our system.

If the card reader can't be moved back due to interference with internals mounted behind it, then a protrusion would have to extend from the front of the unit to allow a locking mechanism to engage the end of the card. This could compromise the robustness of the unit as a protrusion on the front of the unit could easily be broken off.

Let's design it so the booth protects it from being broken off.

The best solution to this problem is by addressing the problem with software. We need to take a look at solving the problem by setting the expectations level in the proper fashion and improving on our existing process. Allow me to ask if anyone has had experience in the real world (outside of voting) with locking card readers. I am a believer in using plastic for most of my transactions. The only locking card readers I have encountered are those that completely “suck in” my card so that I do not even have access to removing them. I have on occasion not had my card returned, meaning that it was stuck in the reader and therefore that reader was deemed inoperable. All of the other are simply a swipe method to gather information.

I would like to replace the old thinking with the new concept. I look at the voter card as a transaction device. On the new AVTS unit, I insert my card into the AVTS; it reads the information to provide me with a ballot. The software then is sensing if the card is removed. If during the voting process a voter removes the card, we immediately flash a screen that provides the voter the option to “CANCEL the VOTE” or “REINSERT THE CARD TO COMPLETE VOTING”. By providing this it removes the need for poll worker interaction to press the secret portion of the screen to cancel the ballot. If the voter does not elect to cancel the ballot, but reinserts the card, then the ballot will reappear so the voter can continue voting. If the voter presses, “Cancel the Vote” the ballot is cancelled, but the card is still active. The voter can then reinsert the card and begin the process again, since the voting transaction was never completed. Once the voter reaches the final screen, a message could be displayed only as a reminder that the voter should leave the card in the reader until prompted to remove after the vote has been cast and the card disabled. This should remove the paranoia about locking card readers and the need for them. Reports generated will identify the number of voters that pressed, “Cancel the Vote” so that voter counts can be balanced.

This will allow us to stay with the existing design, and we are not as limited on the card readers that we could use. By handling this problem with software, we have much more control over the process. The only reason that we believe that we need to lock the card is to keep a voter from removing it during the voting process, only so we can disable it at the end. That requirement goes away once we allow the software to properly sense the need for a voter card to be in the card reader.

Much of this makes a compelling argument. However, there are other issues that must be considered and reviewed.

I don't believe we want voters to "swipe" the card. We have to be cognizant of our competitors "locking" card unit. We can certainly set voting expectations after the sale, yet we must have a unit that demonstrates the appropriate card security, or we might never get to the implementation level.

Let's look at a typical early voting scenario:

Perhaps a voter gets part of the way through a 6 page ballot, removes the card and walks out. He/she doesn't see, or care, about the software message. The next voter walks up and inserts his card. This might require some pollworker interaction. We now have to vote for the absent voter or cancel the ballot. We have to deal with marking the voter in the VR system. (Have they voted or not?) And of course we have perceived security issues in the eyes of the pollworkers, voters and election administrators. Sure, this could happen if the card was still in the reader, but the next voter couldn't insert the card. At least this gives us a "perceived" security advantage. The card hasn't left the polling place.

The point is we can "sell" through this in our presentations.

I'm not suggesting the card must "lock" or we are doomed to failure. We've demonstrated the ability to sell products with some weakness. I am strongly suggesting we carefully review our ENTIRE new product (Booth, battery life, VIBS, smartcard burning, smartcard reader, VR interfaces, etc.) BEFORE we rush a product to market. We've seen the problems when we push our products, ie blue/white unit, accufeed. Perhaps a manufacturing engineer could solve our card reader concerns. My point is we have a significant amount of market experience that resides in the sales team. Perhaps Barry could organize appropriate input from the sales teams that have been installing and selling the TS product line (Steve, Deborah, Robert, Mickey come to mind.) and quickly get some ideas into the process. I don't feel qualified to design anything, yet I do feel strongly about the future of our company. Why not invest in a couple of conference calls before we finalize designs. We must have a unit that appropriately competes in the marketplace. Hart and S/P have brought units to market with some very nice features. We can build from their experiences and problems.

It seems to me a couple of conference calls in the next week or so might be helpful.