Two previous posts - machine counts and open source

Written by: Warren_D Smith on Nov 1, 2005 1:06 PM EST

Dear DFAers for voting reform.
I here address two previous posts.

Q1. Is open-source voting software superior to secret code?
A. I say yes. First, open source has been done in Australia
(voting machines statendardized, all their software posted on internet).
Second, 600,000 lines of secret code (estimated for Diebold) does not inspire confidence.
A SMALL amount of public code is the only way to get any confidence the
code can actually be checked. In India they made voting machines with small
hardwired (ROMmed) un-reprogrammable code and it seems to have been highly successful.

In both these countries, the whole USA idea of competition between different voting machine
mnfctrs with a total lack of standardization - was abandoned. In the USA, all attempts
to force open code or standardization, have been blocked by voting machine manufacturers.


Q2. Is hand counting better than machine counting?
A. I say it depends on the machine. A study of New Hampshire indicated that
optical scan machines had better reproducibility than human counters. Machines
are automatically unbiased - except that in Florida 2000 Jeb Bush & Katherine Harris
arranged matters so that machine ballot-rejection standards and procedures would
differ in different areas - very black areas then experienced over 20% ballot spoilage
rates, whereas the whitest areas experienced, e.g. 1%. This was done
systematically county by county and precinct by precinct (i.e. in the black
precincts voters were not notified their ballots were invalid - they were simply silently
accepted; in white areas they were notified and asked if they wanted to try
again to construct a valid ballot...) with the result that
Bush won the election. So... I claim if the right kind of machines are used
(optical scan) with uniform standards imposed nationwide, then machines are
superior. I believe that computerized machines should be severely restricted, as
in India, because they are too easily hacked...

Warren D. Smith
PhD Mathematician and CRV founder.
Link
warren.wds /at/ gmail.com
wds /at/ math.temple.edu

Tags: