Fwd: ES & S Vote tabulation software

From: Edward Cherlin <edward_dot_cherlin_at_etssg_dot_com>
Date: Mon May 17 2004 - 09:42:11 CDT

Worse than I thought.

---------- Forwarded Message ----------

Subject: ES & S Vote tabulation software
Date: Saturday 15 May 2004 03:11 pm
From: Joseph Holder <conquip@ix.netcom.com>
To: undisclosed-recipients:;

    I am attaching two jpeg images that originally came from
 Miami-Dade County's Elections Division. It appears that an IT
 person on their staff was examining the ES & S vote tabulation
 software that Miami-Dade uses. He wanted to establish that the
 "reports" generated by it could be used for 1. auditing an
 election; 2. recount an election; 3. if necessary use the
 reports to certify an election. His conclusion was they could
 not. I would urge people to print out the images and read the
 memo for yourself. It is obvious this person is very
 conscientious in doing his job. He was so disturbed by his
 first test that he ran it a second time on a different
 computer. He picked by random selection a precinct (115) from
 the May 20, 2003 election. At that precinct they had printed
 out a open of polling "zero tape report", and a close of
 polling "closing report". Both reports listed 9 machines with
 the identical serial numbers. This report also listed the
 number of votes cast on each machine. Bear in mind these
 reports are generated at the polling location, prior to the
 upload of the data to the ES & S election management software.
    Now starts the problem area. The central tabulation software
 after upload of the precinct data and tabulation issues an
 "Event Log". Two of the nine machines are missing from the
 Event Log. Now there appears a machine that never was used at
 that precinct. The number of votes cast on it is equal to the
 number of votes cast on the missing two machines. Thus the
 total of votes cast is the same. What we do not know is the
 apportionment of those votes.
    Then he compared the original precinct reports to the Vote
 Image report. Now three of the original voting machines are
 missing. Instead there is now the same serial numbered machine
 as showed up in the Event Log appearing in the Vote Image
 report. The total of votes remained the same.

    He wants to confirm that he did not do the test wrong or
 there was a "computer glitch" with the computer he was using.
 He and his staff ran the reports again on a different computer.
 This time, rather than relying upon the information coming from
 the central tabulation database, they recreate the election by
 processing the original PCMCIA card, the PEB, and the flash
 cards. The result is:

The same two serial numbered machines are missing again from the
 Event Log.
More ominously, The 38 total votes that had been cast on those
 two machines were also missing. Also this time there was no
 creation of another machine serial number with the "missing"
 votes. This time the Vote Image Report omits the same three
 machines as the first time. But different this time is the
 addition (substitution?) of two machines that were not used in
 the precinct. One of the serial numbers is the same as the
 created machine in the first report. That one lists 6 votes on
 it, the same number as on one of the "missing" machines. The
 entirely new serial numbered machine lists 32 votes on it, the
 same as another one of the missing machines.

    This raises some very serious questions.
Since the "Event Log" is supposed to be the primary proof of
 what occurred at the polling place, and the means to track the
 vote from the voter to the final "ballot image", how could two
 machines out of nine disappear?
How could the vote totals for those missing machines be
 transferred to a completely new machine that was not ever used
 in that precinct? How could the Event Log and the Vote Image
 Log both create the same new serial numbered machine and assign
 the same missing votes to it? When they ran the "raw" data back
 through the system the Event Log lost the same two machines
 again and this time the 38 votes that they had recorded were
 also lost. Why did the Event Log not create a new serial
 numbered machine and assign the votes to it?
Why did the Vote Image Report lose the same machines again, but
 this time, ON ITS OWN, created two new machines (one with the
 same serial number as the first time), and assign the missing
 number of votes to each of the new machines.

    This means that the vote tabulating software has
 instructions within itself to move votes around between
 machines. What would be very interesting is to examine those
 particular machines and see if there was any change in the
 apportionment of those 38 votes between candidates.

    At first glance the first time they checked this out they
 could say that possibly a human discovered the two missing
 machines and substituted another serial numbered machine so the
 votes would balance. Not an acceptable manipulation of data or
 audit logs, but understandable given what we now know about
 election officials "playing" around with reports.
    The second time they did not take any direct human
 intervention. That means the software program itself is capable
 of acting on its own. That is totally unacceptable.
    Today the Oakland Tribune is reporting that it is very
 likely the TS & S software used in tabulating the votes for
 both the touch-screen machines and the optical scan systems may
 never have been federally tested or examined. We already knew
 from the Compuware Report that the database on the central
 tabulating software program could be accessed and manipulated.
 Now it is appearing that the program itself is capable of
 acting without human intervention too.

The evidence keeps mounting that the major vendors might not
 only build unreliable and error prone equipment, but also have
 software that may be designed to rob us of accurate election
 results. It is no wonder they are fighting us so hard against
 any paper trail.


Edward Cherlin, Simputer Evangelist
Encore Technologies (S) Pte. Ltd.
No aid, no charity, and no poor

= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain

ESSmemo2.jpg ESSmemo1.jpg
Received on Mon May 31 23:17:49 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:16 CDT