Re: Voting System Standards

From: Alan Dechert <adechert_at_earthlink_dot_net>
Date: Sun Jul 27 2003 - 16:10:17 CDT

For example, strictly speaking, this voting machine is NOT a DRE. It is a
voting machine that produces paper ballots. The electronic record may
facilitate vote tabulation but the electronic ballot images will not be the
official ballot.

For example, in this document (
http://www.fec.gov/pages/vssfinal/v1/_Toc7196985 ) I read:

      DRE systems shall include an audible or visible activity indicator
providing
      the status of each voting device. This indicator shall:

           a. Indicate whether the device has been activated for voting; and
           b. Indicate whether the device is in use.

I would say that the printout would be sufficient to satisfy these required
indicators. On the other hand, since our system is not really a DRE, it may
be that nothing here is applicable. These standards were developed without
the concept of a "voter-verified" printed ballot as the authentic vote.

There are other DRE requirements that we could reasonably adopt, such as
multiple recording media. We're already planned to do this. Selections
would be written as they are made to the HD as well as floppy (or some other
media). When the polls close, the electronic ballots are sorted by ballot
number and written to CD.

I found some amusing things in this document (
http://www.fec.gov/pages/vssfinal/v1/v1s4.doc )

      4.2.1 Selection of Programming Languages
      Software associated with the logical and numerical operations of vote
data shall
      use a high-level programming language, such as: Pascal, Visual Basic,
Java, C and C++.

They lump VB in with C as a "high-level programming language." The
"Software Standards" they discusss are hardly rigorous. They seem happy
with "high-level programming languages" for everything.

Then there is this gem:
      4.1.3 Exclusions
      Some voting systems use equipment, such as personal computers, that
      may be used for other purposes and have resident on the equipment
      general purpose software such as operating systems, programming
      language compilers, database management systems, and Web browsers.
      Such software is governed by the Standards unless:
           - The software provides no support of voting system capabilities;
           - The software is removable, disconnectable, or switchable such
             that it cannot function while voting system functions are
enabled; and
           - Procedures are provided that confirm that the software has been
              removed, disconnected, or switched.

So it appears they allow most any programs to be on the voting machine PC as
long as these programs don't have any purpose there. Okay, you make a list
of all the junk that's on the voting machine and issue a statement that
says, "ignore all these programs because they don't have anything to do with
the voting application." This sounds pretty easy.

In any case, none of this matters for the demo. The real system we propose
to develop in the larger study will need to be developed with attention to
these matters.

It's hard to take this stuff seriously. It's worth reading before we finish
the demo, but the "Voting System Standards" are crap. In the larger study,
we will help them to re-write them so they are a bit more meaningful.

Alan

- ----- Original Message -----
From: "Douglas W_dot_ Jones" <jones_at_cs_dot_uiowa_dot_edu>
To: <voting-project@lists.sonic.net>
Sent: Sunday, July 27, 2003 7:46 AM
Subject: Voting System Standards
Message-ID: <5717@initial.digest>

> Everyone involved in writing code for a voting system should read
> the current 2002 voluntary Federal Voting System Standards. These
> standards are available from the Federal Election Commission web
> site, http://www.fec.gov (look under elections and then under
> voting system standards). I can't get through to them right now
> to get the detailed web address (could be the local thunderstorm).
>
> These standards contain extensive rules that any voting system must
> satisfy -- behavioral rules, handicapped accessibility rules and
> software development rules.
>
> These standards aren't fun reading, but they've been enacted into
> law by the majority of the states, and voting systems that don't
> conform are thrown out without apology.
>
> There are also big problems with these standards. The IEEE
> is currently working on better standards; if you look at the
> web pages of the IEEE technical committee P1583, you'll find
> draft standards and discussions of the drafts. These are better,
> but just as maddening to read, and far from perfect.
>
> Doug Jones
> jones@cs.uiowa.edu
>
Received on Sun, 27 Jul 2003 14:10:17 -0700

This archive was generated by hypermail 2.1.8 : Wed Aug 06 2003 - 12:50:26 CDT