Proprietary Software Voting Systems?
                            by Bradley M. Kuhn

Regardless of the conflicting opinions on the voting situation in this past
election, most voters in this country now agree that voting systems could
use an upgrade.  Recently, three American companies, Unisys, Microsoft, and
Dell, teamed together to build electronic voting machines which they hope to
sell to various localities in the United States.

Accepting the systems from these companies would be a good idea if the
various election commissions and independent investigators could check the
machines for flaws.  Unfortunately, this will not be possible.  The
potential buyers will have to trust that the machines work as intended; yet,
they will not.  Election commissioners will be like a person who is
forbidden to test drive a car before buying it.

This is because these electronic voting systems will likely be built with
proprietary software.  The computer source code---the set of directions
given to the computer to carry out its task---for the voting systems will
not be available to the public, nor the government, nor the election
commissioners.

If we are lucky, it may be that the software will work without error; but no
one will know.  The few people in the three companies permitted and
encouraged to examine the code will think they know the code works, but they
will not.  Long experience has shown us that bugs escape the eyes of even
the most experienced computer programmers and quality testers.

Our recent election has showed us that allowing the public to review the
vote collection and counting process can increase voter confidence, and
ensure that votes are counted fairly and accurately.  However, a voting
system built around proprietary software will be less open to public
scrutiny than our existing voting systems.

For example, if you use a traditional lever-based voting machine, you can
take off the back and inspect the gears to ensure it functions correctly.
If you use the punch-card system, even if you use a hardware machine to ease
the counting, it is still possible to inspect the physical voter card to
check that the machine is functioning properly.  Discussions of dimpled and
hanging chads aside, at least the system leaves the possibility of direct
scrutiny by the public.

Contrast this to an electronic voting system.  Such a system has no "gears",
per se, nor any tangible representation that records the voter's intent.
The software, instead, acts as the gears, the ballot, and the vote counter.
The outcome hinges upon the function of the software.

Do we, the voters, want the outcome of our elections determined by software
that has been examined only by a small group of individuals at a few
companies?  As a computer programmer, I certainly know that writing software
without any bugs is extremely difficult, and also that it is not a difficult
task to add code into a program that would tip the election results.

However, that's not say it is a bad choice to use computer software to help
us handle our elections.  Yet, for the public to have confidence in the
system, elections must be seen as fair and honest.  The only way to ensure
such a perception is to enable every concerned party to watch every part of
the process (except an individual's actual vote, of course).

This means that all the machinery must be visible for inspection and review.
In the case of electronic voting, it means that the software source code
must be open to inspection and freely available (in both a practical and a
legal sense) to everyone who wishes to copy and study it.  This way, members
of the public and election commissioners can scrutinize it, and independent
agencies will have the freedom to audit it for bugs and discrepancies.
 
If voting systems are built around proprietary software, whose to say we can
trust that small group of programmers to resist the temptation to falsify
election results?  And, do we really want the entire outcome of elections to
rest on the programming skills of such a small group?  I hope voters will
begin to ask these questions long before companies like Dell, Unisys and
Microsoft have completed their design.

But, it is up to us, the free software community, to begin a dialogue about
this issue.  We already know the ethical and technological benefits of free
software.  Likely, those who will make the decision about accepting a given
electronic voting system do not yet understand the importance of being able
to scrutinize the source code.  We have to educate them, and soon.  Please,
if you live in the USA, contact your local election board and your state and
federal congressional representatives about this issue.  We must educate
people who make the decisions before it is too late.

Copyright (C) 2001, Bradley M. Kuhn

Verbatim copying and distribution of this entire article is permitted in any
medium, provided this notice is preserved.