A verifiable secret shuffle and its application to e-voting
C. Andrew Neff,
Proceedings of the 8th ACM conference on Computer and Communications
Security Philadelphia, PA 2001

Summary: 
	[Prepared by David Warden]

	This paper presents an algorithm for a random, verifiable
permutation of an input sequence. The probability of cheating is close
to the number of elements to be shuffled k divided by the size of the
subgroup of the multiplicative group of units modulo a large prime in
which the elements are encrypted q. The algorithm requires 8k+5
exponentiations, which is a significant improvement over prior
algorithms. This algorithm provides utility for solving the problem of
holding a universally verifiable secure election protocol. In
particular, it provides assurance that the voters came from a set of
authorized principals, while providing computational privacy with
respect to how any particular individual voted.

Pros:
	[Prepared by Christina Abad and Cigdem Sengul]

	The algorithm presented in the paper protects individual privacy
with a higher degree of assurance than previously offered. It also is
more efficient in terms of operations performed and proof size then
other algorithms it is compared with.
	The paper provides detailed proofs resulting in a convincing
argument for the claims made.

Cons:
	[Prepared by Joel Hegg]

	The level of detail in this paper is below the threshold which
many computer scientists typically want to consider ideas. As discussed
in prior sessions, when selecting and using an algorithm more
abstraction is desirable.
	The paper requires a great deal of mathematical background which
makes this paper inappropriate for the conference in which it was
accepted. 
	Costs of modifying election systems on a national basis are
prohibitive and face political challenges. There are concerns about the
reliability and security of the system. The nature of electronic votes
alters the nature of the auditing process in that there is no physical
record of the votes. 

Discussion:
	The discussion began with a discussion of the technical issues.
The group did not have sufficient mathematical background to fully
consider the technical merits of the specific proofs; however the
general claims were seen as important to secure election schemes. The
nature of zero knowledge proofs was discussed with respect to the roles
of the verifier and the prover, including the problems that can be
introduced with dishonest verifiers. 
	The group proceeded to discuss the general feasibility of
electronic and internet voting. There were several advantages to
electronic voting when compared with the accuracy of manual voting
machines. As illustrated in the most recent presidential election,
voting equipment is dependent on physical environmental factors such as
humidity, so the results are not consistent. The physical counting and
auditing processes are expensive. For this reason any electronic system
could be beneficial. There are several options on how to implement such
as system; besides the proposed algorithm smart cards could be used in
conjunction with a physical polling place. 
	In the case of Internet voting it would be unfeasible to protect
the system from all forms of attack. Any such system could be subjected
to distributed denial of service attacks. The public confidence may be
insufficient for a national election scheme, although internet elections
have been used for corporate stockholder and other elections. 
	One major concern with any change to the election system is that
it might affect the results with respect to the proportion of votes
going to the largest political parties. For that reason at least one
party would oppose the change and be able to prevent it from being
instituted. 

Conclusions:
	The group weakly rejected the paper as it was not appropriate
for the conference in which it was presented. Some members felt
unqualified to judge the merits of the paper and it was resolved to
discuss zero knowledge proofs at a future meeting.