[EM] Re: [UWSA] Voting debate - missing the target

[Dave Ketchum]

Let me expand a bit on what I wrote Mon, 28 Jul 2003 11:40:16 -0400

Jerry has talked of part of it.

BIG DEAL is to REQUIRE that potential vendors provide open source code to 
both encourage them to provide code that does what it promises for the 
public, AND that poll watchers can get validated by whoever they trust as 
willing and able to do this.

Machine should:
      Probably have a Z80 chip - old and stable, yet hopefully not too 
easy to make hacked imitations without getting caught.
      Battery backup, inside the machine.
      Have a clock, to record time whenever appropriate.
      Able to print results both at beginning of a mode to show zeros, and 
at end to show voting totals.  This printer only needs dependability, not 
speed, for it does little printing.  If ballots are to be printed for some 
reason, that gets done by some other device.
      Able to read and write CDRs
      Have random access memory (likely a hard disk) to stow ballots as 
voted, but in random order for secrecy.
      Probably some firmware, but the voting machine program comes from 
the CDR - thus versions tailored to the election methods being used can be 
loaded.
      Have visible modes of operation:
           Setup before polls open.
           Checking - operate EXACTLY as with polls open, except that 
indicators visible to everyone, ESPECIALLY "voters", show the mode. 
Includes printing results when machine switches to another mode.
           Polls open - exactly same as checking, except mode is 
displayed.
           Leaving checking or open mode - need, when polls close, to 
automatically print results on the same paper as has been in the printer 
since zeros were printed at time polls were opened.
      Any access doors, etc., that were open during setup must be locked 
in other modes (such as the CDR drive).  If machine is able to access the 
internet, this prohibition applies to that ability.

CDR content:
      Voting machine program - provided by vendor, validated by poll 
watchers and others, and copied onto a CDR for each machine before 
election day.
      Ballot definition - provided for each precinct.  Probably done by 
election workers before election day.  Vendor provides PC program for 
doing this and I see no need for validating this program - sensitive part 
is what format the voting machine program is willing to accept, and that 
election workers validate this on a voting machine in checking mode at 
preparation time and poll watchers validate during checking mode.
      Diary - voting machine program records everything that might be of 
interest as it happens on this single voting day:
           Probably omit this during checking mode, or have two checking 
modes, one including diary to be like open mode.
           Must NOT record votes when voted (must keep secrecy, thus 
record these only in random order, thus I believe the CDR cannot be used 
for this.
      Record of all ballots for history, but arranged in random order for 
secrecy before writing to the CDR.

Note that I put no special effort into seeing to it that the right program 
and ballot definition was on the CDR.  Instead, the CDR gets identified by 
the diary that is recorded on it while polls are open, and election 
workers are going to try to make sure that the program and definition on 
that CDR are the correct ones.

Dave Ketchum

On Mon, 28 Jul 2003 10:30:30 -0700 Audrey wrote:

> 
> At 09:05 AM 07/28/2003 -0500, you wrote:
> 
>> You seem to focus on the SOFTware part of the problem, Jack. And I don't
>> like your solution. All that does is to provide the fire hose through 
>> which
>> will pass a flood of complaints about the dishonesty of the system, 
>> and that
>> flood will be dealt with in the same way that Wolfowitz and the rest 
>> of the
>> PNAC cabal are dealing with the truth about Bush's war.
>>
>> What we need is an integrated solution that includes hardware as well as
>> software. If voting machines were tamper-proof in a hardware sense and the
>> software were embedded as firmware and thoroughly documented and published
>> to the world, and if there were no processing of data outside of each
>> precinct's machine then the output would not be questionable. There 
>> would be
>> required hard copy outputs in addition to the vote count that would be
>> checkable against data provided by the human voters outside the 
>> clutches of
>> the black box voting machine, and partisan watchers at each precinct.
>>
>> This could work. It is when code is loaded into the box and there are
>> interfaces for additional input of code before results are certified and
>> reported that trouble occurs.
>>
>> Jerry
> 
> 
> 
> REPLY:   OK, Jerry.  I will say the same thing that I wrote to Jack.   
> We need your knowledge and insight in cleaning up the system.  I don't 
> care if you fix the hardware or the software.  I just want to see 
> election results that truly reflect the will of the majority.  Election 
> results in recent years have been highly suspect.  I think the majority 
> of Americans are not happy with the government that we have, but somehow 
> have not been able to express their views at the polling places.  If we 
> do not do something then maybe soon we will have Hillary Clinton for 
> president!
> 
> Help us.
> 
> Audrey of AUDREYS MISSILESS

-- 
davek at clarityconnect.com  http://www.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.