I posted earlier this morning, the thread
The Case of the Diebold FTP Site - UIOWA Professor Detailed Research
(
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=51975&mesg_id=51975).
Which was a reference to a very detailed and thorough description of the Diebold GEMS
application, and the various finds over the past year. I wrote an email to the
professor at UIOWA, and he sent back a very quick and interesting reply. I thought
I would share...
<<EMAIL
> I was writing you, in curiosity to what your thoughts are with
> the current election results. Do you find it surprising that
> issues and irregularities have been found and reported?
With the controversy over Florida 2000, and the public revelation
of security problems in Diebold's code a year ago, election 2004
attracted massive scrutiny. Both parties had lots of lawyers
at the polls and at county courthouses observing the process.
Groups like Techwatch had computer professionals mobilized. For
the first time ever, we had a national election incident reporting
system in place (run by Verified Voting, with help from many
places).
In this context, many election officials were nervous and took
unusual amounts of care to get things right.
Others were in a
state of denial and did everything they could to ignore the
observers; some even impeded them. Many in places like Ohio said
"we're not Florida".The net result? I think the average election official nationwide
did a better job than usual. I think that the problems that have
come to light in this election are therefore very significant
because they show how badly run elections generally are.
We have shoddy software, shoddy voting systems, and shoddy election
administration, so that even when things are done better than
usual, you find reports of lost votes by the thousands.
> Do you think Diebold took the necessary steps to do a complete
> audit and restructuring of their application after all the
> discoveries made by Bev Harris and Dr. Rubin?
No, I believe they changed some of the window dressing and made some
improvements. The fundamental flaws we found were the kinds of things
they can't change in just a year with a few patches. They really
need to rethink the overall design of their software.
> In my opinion, I truly think some of the numbers aren't making sense.
I don't see anything pointing at massive fraud at this point. I'll
never say there was no fraud, though, because I suspect that there
are still many counties (a small minority) with varying degrees of
corruption in their election administrations. But I see nothing
that suggests that this residue of corruption has had a national
effect.
I was talking to a professional pollster yesterday, and he had an
interesting comment on the exit polls: The morning exit polls
vastly oversampled women and undersampled men. Drawing conclusions
about the results from these exit polls would be wrong because of
the partisan differences between men and women. Furthermore,
most of those involved in exit polling haven't designed their
exit polls to predict the outcome, they've designed them to study
what the issues were that decided the election. Only as the polls
start to close do they fire up some very powerful mathematical models
to correct their exit polls using real vote totals from real ballots
in order to derive projections for the precincts that have yet to
produce ballot counts.
Doug Jones
cs.uiowa.edu
END EMAIL>>