"... In PAGE 5: ... However, this search could have wasted a lot of time, when maybe all we need is a relatively minor modification to option A. hard headed business approach, may not be the to be effective are As is shown in Table2 , electronic voting on options can highlight divergent opinions in a useful fashion and enable any important differences t fully explored. What we see in Table 1 is the results of a choose which of Options A, B or C is to be implemented to solve our problem.... ..."

"... In PAGE 6: ...able 1:Electronic and paper ordinary votes issued at pre-poll centres ................................ 38 Table2 : Summary of first preference electronic votes by electorate/ACT total .... In PAGE 17: ... In the event, there was a noticeable difference in the voting pattern between those who voted electronically and those who voted on paper ballots. Table2 shows the numbers of electronic votes cast in the 3 electorates and in total. Table 3 shows the number of paper ballot votes cast by all electors.... ..."

"... In PAGE 6: ...able 3: Summary of first preference paper ballots by electorate/ACT total........................ 40 Table4 : Summary of all first preference votes by electorate/ACT total.... In PAGE 17: ... Table 3 shows the number of paper ballot votes cast by all electors. Table4 shows the number of votes cast by all electors. It can be seen that the party in column A in Ginninderra and Molonglo, the Australian Democrats, received a higher percentage of electronic votes compared to the paper ballots: 13.... ..."

"... In PAGE 6: ...able 2: Summary of first preference electronic votes by electorate/ACT total ................... 39 Table3 : Summary of first preference paper ballots by electorate/ACT total.... In PAGE 17: ... Table 2 shows the numbers of electronic votes cast in the 3 electorates and in total. Table3 shows the number of paper ballot votes cast by all electors. Table 4 shows the number of votes cast by all electors.... ..."

"... In PAGE 9: ...3 Using Margins for Regression Within the context of regression, once a case is classi ed, the a priori mean or median value associated with the class can be used as the predicted value. Table2 gives a hypothetical example of how 100 votes are distributed among 4 classes. Class 2 has the most votes;; the output prediction would be 2.... ..."

"... In PAGE 2: ................. Table4 : Comparison and contrast of major voting machines 22 Shoup .... ..."

"... In PAGE 5: ... 3#28a#29. Table1 #28a#29 lists all the y i apos;s that have implication value 0 for each wire. We explain the interpretation of Table 1#28a#29 by examples.... In PAGE 5: ...lications on the example circuit shown in Fig. 3#28a#29. Table 1#28a#29 lists all the y i apos;s that have implication value 0 for each wire. We explain the interpretation of Table1 #28a#29 by examples. The meaning of the second row is that we expect wire b ! g 1 to be removed if we choose y 1 + y 2 + y 4 + y 5 as the core divisor.... In PAGE 5: ...Table 1: Vote table choose, and hence has no candidate core divisor. The remaining entries of Table1 #28a#29 can be interpreted in a similar way. The abovevoting scheme demonstrates our criteria for choosing a good core divisor.... In PAGE 5: ... This is done bychecking if the candidate core divisor voted by a wire w is a SOS of the cube that is connected to wire w. For example, from the #0Crst entry in Table1 #28a#29, the candidate core divisor of wire a!g 1 is y 1 + y 2 = ab + c. The cube that is connected to wire a!g 1 is x 1 = abd.... In PAGE 5: ... Since the can- didate core divisor ab + c is a SOS of cube abd, we knoweventually if we add core divisor ab + c into the circuit, the added wire will be a redundant wire and therefore the circuit functionality would not change. In Table1 #28a#29, the only candidate core divisors that do not hold for this condition are wires d ! g 3 and a!g 4 . The candidate core divisor for wire d!g 3 is y 4 +y 5 = bd+e, which is not a SOS of the correspond- ing cube x 3 = cd.... In PAGE 5: ... On the case of wire a!g 4 , candidate core divisor y 1 = ab is not a SOS of the corresponding cube x 4 = ae. We therefore need to delete these two entries in Table1 #28a#29 and wehave our #0Cnal vote table, shown in Table 1#28b#29. To #0Cnalize the choice of the core divisor, various heuristics can be used.... In PAGE 5: ... On the case of wire a!g 4 , candidate core divisor y 1 = ab is not a SOS of the corresponding cube x 4 = ae. We therefore need to delete these two entries in Table 1#28a#29 and wehave our #0Cnal vote table, shown in Table1 #28b#29. To #0Cnalize the choice of the core divisor, various heuristics can be used.... In PAGE 6: ...8#25 10.5#25 Table 2: Experimental results table as shown in Table1 #28b#29. The only slight modi#0Cca- tion we need is in the #0Cnal maximal clique formulation, where we need to model the fact that some cubes in the second column of Table 1#28b#29 originally come from a di#0Berent node.... In PAGE 6: ...5#25 Table 2: Experimental results table as shown in Table 1#28b#29. The only slight modi#0Cca- tion we need is in the #0Cnal maximal clique formulation, where we need to model the fact that some cubes in the second column of Table1 #28b#29 originally come from a di#0Berent node. Note that, as is also the case for basic division, we can perform extended division in terms of sum-of-product form as well as product-of-sum form.... ..."

"... In PAGE 5: ... 3#28a#29. Table1 #28a#29 lists all the y i apos;s that have implication value 0 for each wire. We explain the interpretation of Table 1#28a#29 by examples.... In PAGE 5: ...lications on the example circuit shown in Fig. 3#28a#29. Table 1#28a#29 lists all the y i apos;s that have implication value 0 for each wire. We explain the interpretation of Table1 #28a#29 by examples. The meaning of the second row is that we expect wire b ! g 1 to be removed if we choose y 1 + y 2 + y 4 + y 5 as the core divisor.... In PAGE 5: ...Table 1: Vote table choose, and hence has no candidate core divisor. The remaining entries of Table1 #28a#29 can be interpreted in a similar way. The abovevoting scheme demonstrates our criteria for choosing a good core divisor.... In PAGE 5: ... This is done bychecking if the candidate core divisor voted by a wire w is a SOS of the cube that is connected to wire w. For example, from the #0Crst entry in Table1 #28a#29, the candidate core divisor of wire a!g 1 is y 1 + y 2 = ab + c. The cube that is connected to wire a!g 1 is x 1 = abd.... In PAGE 5: ... Since the can- didate core divisor ab + c is a SOS of cube abd, we knoweventually if we add core divisor ab + c into the circuit, the added wire will be a redundant wire and therefore the circuit functionality would not change. In Table1 #28a#29, the only candidate core divisors that do not hold for this condition are wires d ! g 3 and a!g 4 . The candidate core divisor for wire d!g 3 is y 4 +y 5 = bd+e, which is not a SOS of the correspond- ing cube x 3 = cd.... In PAGE 5: ... On the case of wire a!g 4 , candidate core divisor y 1 = ab is not a SOS of the corresponding cube x 4 = ae. We therefore need to delete these two entries in Table1 #28a#29 and wehave our #0Cnal vote table, shown in Table 1#28b#29. To #0Cnalize the choice of the core divisor, various heuristics can be used.... In PAGE 5: ... On the case of wire a!g 4 , candidate core divisor y 1 = ab is not a SOS of the corresponding cube x 4 = ae. We therefore need to delete these two entries in Table 1#28a#29 and wehave our #0Cnal vote table, shown in Table1 #28b#29. To #0Cnalize the choice of the core divisor, various heuristics can be used.... In PAGE 6: ...8#25 10.5#25 Table 2: Experimental results table as shown in Table1 #28b#29. The only slight modi#0Cca- tion we need is in the #0Cnal maximal clique formulation, where we need to model the fact that some cubes in the second column of Table 1#28b#29 originally come from a di#0Berent node.... In PAGE 6: ...5#25 Table 2: Experimental results table as shown in Table 1#28b#29. The only slight modi#0Cca- tion we need is in the #0Cnal maximal clique formulation, where we need to model the fact that some cubes in the second column of Table1 #28b#29 originally come from a di#0Berent node. Note that, as is also the case for basic division, we can perform extended division in terms of sum-of-product form as well as product-of-sum form.... ..."