Reply #29: rocket science

[Neil B Forzod]

I'm not sure whether they all use the same mix of applications. Certainly they all have their own software products, but there may be some common 3rd-party components that are used by more than one vendor. Don't know.

As for scrupulous companies cashing in, I suspect there are lots of reasons, and I can play devil's advocate and hazard a guess as to a few:

• Reliable hardware actually isn't an easy problem -- it's got to have particular performance characteristics, it's got to survive drops from a height, bombardment with electromagnetic radiation, it's got to be tamper-proof, it's got to operate on battery power for several hours in the event of power failure while still being self-contained and portable. Software is easy, anyone can sit down and write it. Hardware design and (especially) manufacturing is a harder problem: it's expensive and lots of companies just don't have the resources.
  
• 100% open source code would be nice, but almost all software companies vigorously protect their source code since it's their intellectual property and what sets them apart from people who don't have some source code to do the same thing. Despite claims to the contrary, election management software isn't easy and most companies don't want to make it easy for potential competitors by putting all their intellectual property out there to copy. (Licensing agreements would theoretically protect them, but then they have to litigate with every competitor they think might have used their intellectual property inappropriately).
  
• Verifiable paper trails aren't a huge problem, every current vendor who was asked to provide one has either done so or has one in development. The main issue is hardware design and manufacture (see point #1 above) -- you can't just bring a bunch of inkjet printers and hook them up to some DREs. What happens in the event of failure, or a paper jam? How will the printer work in the event of power failure? The DRE has to keep working, but what do you do about an external printer that's attached? Oops... so maybe you build the printer to make use of the existing report printer instead (which is what all vendors have essentially done, incidentally) but now you're using a lot more power (because you have to print a bunch of paper for every ballot and running a mechanical device like a printer is relatively expensive compared to driving some electronics) and you find out that your DRE won't run on battery power long enough to satisfy FEC regulations because it was originally designed without planning to drive a printer. So now you need to redesign some hardware, and you can refer back to point #1 again. :) That's the main sort of issue that arises with voter-verified paper audit trails.
  
• Most major election equipment vendors (ES&S, Diebold, Sequoia and to a lesser extent Hart-Intercivic) sell not only the vote-counting equipment but a bunch of additional services and expertise for election officials that are often overwhelmed and underfunded. That includes everything from ballot printing and absentee handling, ballot layout, election setup, technical support, etc. All these companies have a significant amount of industry experience (mostly because half their respective staffs are made up of former election officials, lol) and support infrastructure. The scrupulous startup hypothesized in your post probably doesn't have any of that, probably has no connections with the industry or any election people, and would have a tough time selling its wares as a result. Take a look at AccuPoll. They have a DRE and an EMS product. They're federally certified to the latest standards. But they have less than 10 employees, total, and no customers. Suppose they bid on NC against the big vendors... the RFP asks the vendor to indicate the number of support representatives the company will provide to support the state's 100 counties, and to specify their experience. AccuPoll responds with "uh, 6?". ES&S responds with "we'll supply 5 subject matter experts, a 10-man state help desk with phone support, 20 roving support techs to address local county problems, and a two-person project management team to coordinate it all." Diebold responds with "all of that, plus we'll make available 50 local ATM technicians who can be on call to replace faulty equipment, act as runners and who can quickly be deployed wherever there's a problem." Everything else (including price, etc.) being equal, if you're a beleaguered, overworked and understaffed election official, who do you buy your equipment from? Probably not AccuPoll, if you're honest, and probably not your hypothesized scrupulous startup vendor either.
  

All of which is way off topic, lol. :)

Also, I take back part of my earlier post too: my sources in the industry tell me today that Diebold is probably not going to withdraw its bid from NC and is going to let their bid stand. I'm still waiting for an independent confirmation of that, but the source was pretty reliable so I'm inclined to believe it.

Neil