mechanicals for an open voting system

From: Tom McCarty <tom_at_tom-mccarty_dot_com>
Date: Thu Jun 22 2006 - 21:44:50 CDT

Hello all,

I was added to this list just recently, but have been a monthly
supporter of OVC (albeit a small one) since Nov. 2004, and finally have
some time to donate to the cause.

What I'm interested in doing is designing and prototyping the
mechanicals to support the EVM2003 software so that real demonstrations
of the system could be done. This would include a secure case for the
computer, as well as fabrication of the voting booth and any other
associated mechanicals parts. I have a small machine shop, and can
design, machine, and fabricate anything needed. All work done would be
"open design", covered under one of the free licenses so anyone could
make their own voting system by buying the parts on the BOM, and going
to a machine shop with the freely available drawings.

I have so many questions with branching possibilities, I'm not sure
where to start. I've read everything on the OVC website and the EVM2003
sourceforge site, yet I still feel like I'm coming out of left field in
trying to actually build a system.

First off, has anyone else worked on the mechanicals of an open voting
system? Is there already a demo system somewhere? If not, what do
people do for demonstrations now? Is EVM2003 far enough along that a
demo can even be shown, or has it been abandoned in favor of making
existing vendors improve their offerings? I don't see much recent
discussion of EVM, and the lack of pictures on the website and the use
of "an open voting system will..." rather than "our open voting is..."
is not encouraging.

The OVC website says that refurbished PCs can be used, but I really
think an industrial single-board computer (SBC) is the way to go.
Rolling out a voting system across a sprawling state like California
using thousands upon thousands of refurbished PCs sounds like a
nightmare, even assuming you could find enough machines with all the I/O
and the video resolution you require. Since most systems work best with
different device drivers, you can't just use one disk image for all
machines, either, and having hundreds of different kinds of PCs will
make for longer service times whenever something does go wrong.

I propose using an industrial SBC in the EPIC format (which has
multi-vendor support) like the EPX-GX from WinSystems:
http://sbc.winsystems.com/products/sbcs/epxgx2.html
<http://sbc.winsystems.com/products/sbcs/epxc3.html>

The biggest advantage of industrial SBCs is that identical boards, down
to individual components, are produced for years. These WinSystems
boards are $470 each in quantity 100 - a bit more than usual - but these
boards will be available for another 7-10 years, significantly reducing
maintenance headaches over this time. Because of their long lifetime,
industrial PCs are well-designed and actually tested, unlike commodity
PCs, which reduce failure over time. There are many more issues about
SBC requirements, like storage (disk-on-chip for booting, and voting
data on CD-ROM?) and interconnection (ethernet? serial? USB?) to be
figured out, but that's enough for now.

Next would be a secure case for the SBC to control physical access.
Rather than having a black or beige box for a voting machine, clear
lexan could be used to allow an unobstructed view at the interior
components. I think it would be great for demonstrations, esp. for
people that have difficulty visualizing computer security concepts (like
reporters). A clear enclosure would give a strong visual reminder that
this is an "open" system, and would make for some good pictures on the
website, too. It could use a real lock (like Medeco) for I/O access,
and yet another lock for technician-level board access.

The final obvious part is a stand to hold the touch-screen, computer,
UPS, and printer, as well as blinders to keep others from seeing the
screen. I don't even want to speculate on the design for this until
figuring out the PC & case issues, but it should be a simple human
factors analysis and lots of sheet metal.

Obviously, my proposed design is not the cheapest possible design, but
given that there's no profit margin or R&D costs to recover, it should
still be 25%-50% cheaper than Sequoia or Diebold equipment, and offer
unparallel security, reliability, and maintainability. My thought it
that the best way to compete with the big commercial vendors is not on
price alone, but rather, with a system that is ultra-reliable. When it
comes down to it, election workers just want elections to happen without
any problems, and they care about reliability and ease of use more than
they care about security, so I think starting the design with a
high-reliability industrial PC is the best way to insure this, even if
it does make the system more expensive than absolutely needed.

So, what do you all think? Is this a good thing to work on, or would it
be a waste of time? Am I missing some huge piece of the puzzle? I have
no idea where this might lead to (design docs on OVC website? a single
demo system? or even a company that manufactures OV hardware?), I just
see a necessary piece that hasn't been done, and I am trying to do it.

    Tom

ps: to see some other projects I've worked on, check out
http://www.aipengineering.com.

_______________________________________________
OVC-discuss mailing list
OVC-discuss@listman.sonic.net
http://lists.sonic.net/mailman/listinfo/ovc-discuss

==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Fri Jun 30 23:17:10 2006

This archive was generated by hypermail 2.1.8 : Fri Jun 30 2006 - 23:17:12 CDT