[Cialug] voting think tank comment

[Tom Poe]

Jeff Davis offered the following on 09/02/2005 08:39 AM:
> It depends on how accurate and tamper-proof you need it to be.
> 
> A while back there was thread on one of
> securityfocus mailing lists on this topic.
> 
> Some of the problems/issues were:
>  One of the biggest issues was that if
>  you have 1 machine running a program that
>  just displays the ballot and keeps a count
>  of the votes, there is no paper trail to
>  verify the results.  There are many scenarios
>  where a verifiable paper trail is needed.
> 
>  The software needs to be open source for
>  several reasons, including ensuring that
>  the software developers didn't add in any
>  mechanisms for altering the vote count.
> 
>  There needs to be a process in place to verify just prior
>  to voting that neither the hardware nor software has
>  been compromised.
> 
>  You'll need at least a UPS and possibly a generator.
> 
>  How will you handle a hardware failure?
> 
>  Physical security:
>   If you have more than one machine in the set up, then
>   you have introduced a network connection which brings
>   the possibility of the vote being compromised.
> 
>   How is the actual vote recorded? Via touch screen?
>   You don't want a voter alone in a booth with your
>   voting machine, a mouse, and keyboard.
> 
> 
> -Jeff
> 
> 
> 
> 
> Tom Poe wrote:
> 
>> Hi, All:  Over at:
>> http://www.bbvforums.org/cgi-bin/forums/discus.cgi
>> if you click on the link for Think Tank, there's a huge international 
>> discussion going on about voting, electronic and paper, stuff.
>>
>> The question is, can there be a simple computer program that displays 
>> the ballot, records the vote, and prints out the raw data for up to 
>> 1000 votes in a precinct.  The other requirement is it prints out a 
>> ballot to serve as a paper ballot record.  That's the scope.  The idea 
>> is to put a computer in a precinct, and conduct the vote for up to 
>> 1000 people.  No network connection, internet connection, just the 
>> computer performing those tasks.  Maybe there has to be two 
>> computers.  One for generating a ballot.  And, one to receive the vote 
>> count.  Seems like it should only be one computer to me.
>>
>> Anyone have a quick fix on what would be needed?  We don't want to go 
>> beyond what it takes to do more.  Oh, and it has to be GPL, not 
>> proprietary.  For all I know, there's a calculator in Engineering that 
>> already does that.  Is it GPL'd?  :)
>> Tom
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Cialug mailing list
>> Cialug at cialug.org
>> http://cialug.org/mailman/listinfo/cialug
> 
> _______________________________________________
> Cialug mailing list
> Cialug at cialug.org
> http://cialug.org/mailman/listinfo/cialug
> 
Hi, Jeff:  I'll go look for the thread.  Tamper-proof will be difficult, 
at best.
1]  Rely heavily on chain of custody, and oversight, with the object of 
not preventing, so much as detecting tampering
2]  machine displays ballot, prints ballot, counts vote, but paper 
ballot is official count
3]  Software is open source, providing public scrutiny
4]  The election is not carried out behind closed doors at any point
5]  UPS and generator are indispensible, aren't they?  Absolutely, a 
necessity, as we're going to wonder how the South will hold their 
upcoming elections as a dramatic example
6]  How many computers would be needed to ensure we get through all 500 
or 1000 voters?
7]  See #2 above, and wonder if we can find problems with that
8]  I am thinking we are raising the cost to those who want to tamper, 
by shifting the official count back to paper.  We should be able to 
control tampering with security measures as mentioned in #4.

Notice that this project is focused on just a single precinct.  We set 
up for a single precinct, rather than try to have one that is statewide. 
  Once we do it for one precinct, then each precinct can make choices 
that fit their needs.  Presently, to keep costs down, the state has us 
using a single vendor, and some truly interesting logic to move to 
electronic voting.  Maybe that will be rethought, if there's a precinct 
level option.

I really like your quick response.  Bet you have some better ideas, as 
we share more details.  My best shot at this issue, so far, has been to 
take the rather complex, but open source software from Australia, and 
modify it to meet the project goals.  Not sure if that's doable, or not.
Tom


-- 
Open Studios, Charles City, Iowa, USA
www.ibiblio.org/studioforrecording/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: tompoe.vcf
Type: text/x-vcard
Size: 293 bytes
Desc: not available
Url : http://cialug.org/pipermail/cialug/attachments/20050902/635932e6/tompoe.vcf