[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: AccuVote-OS 1.96.2 corrupt count

The corrupt count is caused by a blank selection in a recall race, irrespective of the selection made in the dependent recalled race.  Any ballot or control card fed after a ballot with a blank recall race is fed gives rise to the corrupt count message.
----- Original Message -----
From: Ken Clark
Sent: Thursday, August 23, 2001 5:00 PM
Subject: RE: AccuVote-OS 1.96.2 corrupt count

Okay, this is pretty potentially serious.  Nel (and anyone else), can you please try a couple of other databases to narrow it down a little.  Does this happen with any run-of-the-mill database, or is there something special about this one that causes corruption?  Since it got through Wyle (correct?) I wonder whether there isn't something quirky about the database.  Or is it possible we went through the process without actually testing reject?  Doubtfull.
I will track this down, but it is going to take a bit to get a AVOS development environment up.  I suggest people avoid 1.96.2 in the mean time.  Are any accounts actually running 1.96.2? 

-----Original Message-----
From: owner-bugtrack@dieboldes.com [mailto:owner-bugtrack@dieboldes.com]On Behalf Of Nel Finberg
Sent: Thursday, August 23, 2001 4:57 PM
To: Bug Track
Subject: AccuVote-OS 1.96.2 corrupt count

Memory cards programmed to 1.96.2 firmware from the attached database consistently generate corrupt count messages when fed ballots marked in violation of ballot return conditions defined in the database.  These include blank voted race, blank voted card and overvote.  These corruption messages do not occur when ballots are fed that have been properly marked.
This behavior does not occur with 1.94x and 1.95e firmware.