[E-voting] Open source voting system

Kommer Kleijn kommer at vooreva.be
Mon Apr 8 14:38:11 IST 2013


My opinion is that any computer system that counts votes can be 
tampered with. Even systems like TEVS don't solve this unless a very 
thorough legal system is in place that determines how a correct count 
is acknowledged by all. The problem with systems like TEVS that allow 
access to images of all ballots is that although it theoretically 
allows a recount by citizens (problem 1 -->) there is no guarantee 
that any such recount will indeed actually happen every time, but 
even more important (problem 2 -->) if a citizen recount would yield 
a different result then what legal means are available to challenge 
the official result.

Experience shows that here is the culprit of 'verifiable systems' and 
that even if verification is possible it is either not done or has no 
strong legal value.

The point of manual counting is that the main count (and not a 
secondary check afterwards) is done by citizens. This is what 
democracy is all about.

I actually distinguish three levels of trust in a voting system: 
1)not-verifiable, 2)verifiable, 3)verified. Only the third is 
acceptable for democratic elections. "It can eventually be verified" 
is not good enough. The count needs to be actually verified by human 
citizens. Up to today only forms of manual counting provide that feature.

The only possible aids I see that machines can eventually provide are:
  - assist disabled persons in printing a human readable paper ballot
  - sorting of paper ballots into piles (explanation follows)

A machine could be used to help sort paper ballots into piles per 
party. Then once in piles, the ballots could then very easily be 
counted manually. The important difference is that the machine will 
not count and will not (officially) record any count data from the 
sorting jobs in memory. It would then be very simple to verify the 
accuracy of the sorting job. Any polling station worker can grab a 
pile at will and check the ballots are indeed all for the same party. 
The machine could also be very simple to operate (read-on).

For example:

The ballot sorting machine would be a scanner with a two output trays 
and only two buttons. The two buttons would be labelled "Learn" and 
"Sort". You would operate it like this:
- Take the first ballot from the big pile. Feed it to the machines 
input tray and press "Learn". The machine will scan this first ballot 
and record the party info then output the ballot to tray one. Now 
fetch the entire pile (or as much as the machine can deal with in its 
input tray), and press "Sort". The machine would then recognize the 
ballots for the previously recorded party and output those ballots to 
tray one. All other ballots would be output to tray two. Now put 
aside the ballots from tray one (all sorted for one single party) and 
have someone check the reliability of the sort and then eventually 
already starts counting them. In the mean time you can repeat the 
exact same operation with the output of tray two. Take the first 
ballot of the pile, scan with "Learn", then scan all the rest with 
"Sort". You will obtain a pile sorted for the second party. After all 
parties have past, a manual count of the piles is a peace of cake and 
can be done manually in a short time.

It is the only form of 'computer aided vote counting' that I have 
been able to come up with that can be followed and monitored by 
anyone who can only read and count, which is, IMHO, a condition 
needed for any democratic voting system. (And was confirmed the 
German constitutional court ....)

The key, IMHO, is the fact that the machine will not count! Nor store 
results, as computer memories can be tampered with and can not be 
verified by 'normal' people.

I think it would be great to develop a sorting machine as described 
above. It could run on open source software.  In the hope I did not 
oversee a major culprit it would provide a democratic solution to the 
desire for the use of machines during democratic voting processes.

Any comments?

Best regards!,


  On 20:17 22/03/2013, J.Pelan (Gatsby) wrote:
>Hi Margaret,
> > I'm in the very early stages of trying to set up a project to develop
> > a Free and Open Source optical scan voting system. If you have any
> > interest in getting involved (in whatever capacity), please contact me
> > either on-list or directly.
>We discussed this concept on the list many moons ago.
>Have you seen TEVS ?
>   http://code.google.com/p/tevs/
>   http://www.tevsystems.com/
>All the best,
>John P.
>E-voting mailing list
>E-voting at lists.stdlib.net

More information about the E-voting mailing list