[E-voting] Open source voting system

Margaret McGaley mmcgaley at gmail.com
Sat Mar 23 12:41:42 GMT 2013


Hi John,

I'd appreciate if you could set up a mailing list. We may end up
making a project under OSDV (http://www.osdv.org/) but I'd like to get
the discussion going sooner rather than later. I want the mailing list
to run on the same terms as this one: open membership and public
archives.


> How likely do you think it is that a country would use such an open source
> system?
>

I've decided not to concern myself too much with whether any
particular country will take it up, though I got the impression at the
seminar that the idea of open source for e-voting is becoming fairly
mainstream. I'm focussing on creating a really useful system that
contains the hard work of detailed specification and security
measures, that could be the basis for individual national
implementations. I'm a big fan of countries producing their own voting
systems with local expertise but if we could provide a real
implementation of all the high-level ideas I've been talking about for
years, it would give them a much better chance of creating something
decent.

Let me put it another way: I want to create a system that I would be
happy to vote on. I want any country that *does* decide that e-voting
is worth implementing in their context to get a lot for free. That way
they're more likely to have the security measures that I consider
important, and they're likely to get a much cheaper solution.


>
> Below is a very unfinished initial list of questions/comments:
>
>
> Some initial questions:
> - Are you thinking of a system that scans ballots at polling stations, as
> voters vote, or at a central location?

I think probably Polling-Station scanning, but my mind is still open on that.


>   (The former provides greater security against tampering. Are there
> suitable scanners readily available?)

That's a good question. OSDV have been working on the assumption that
off-the-shelf scanners are the best option. My original idea was to
create an open design for a scanner from easily available parts.
Again, I'm not wedded to either approach.


> - Would it support ranked voting systems (PR-STV, De Borda, etc.)?

That's more complicated, obviously, and I think we should keep it in
mind, but it's not an early priority for me. I don't think there's a
good business case for introducing e-voting in Ireland for example, so
I don't feel very urgent about making the system work here.


>   - If so, do we want to be able to read handwriting, or would we use
> something like a lotto ticket (if there were 14 candidates, each one would
> 14 boxes, and the voter would tick one to indicate their ranking)?

I know they're much easier to scan, but I believe that people find
those interfaces surprisingly hard. We'd have to be sure we weren't
reducing accessibility.


> Some decisions to be made:
>   Features of the system
>   Technical
>     Software
>     - Programming languages to be used in development.
>     - Operating Systems to be supported.
> - Whether the management and counting software is web or desktop based.
>     - Database Server/Software (if any).
>         Preferably one with a compatible open source licence?
>     - Software interface to the scanners (Twain / WIA).
>     Hardware
>     - Supported/recommended hardware platform; Minimum requirements.
>     - Supported/recommended scanners. (I've seen problems with buggy Twain
> drivers in my work).
>   Legal
>   - What licence it will use.
>   Management
>   - Development procedures.
>       e.g. How do we ensure security and quality of the codebase? Maybe a
> policy that all source code commits are peer-reviewed.
>     Technical
>     - Third party code (libraries) used, and policy on allowed licences.
>     - Hosting of the codebase, and source control system (Sourceforge?;
> Google Code?; Private?).
>     - Other development and online collaboration tools. (e.g. Wiki).
>     - Coding convention.
>   Marketing
>   - Name of the system.
>   - Web site hosting and domain.
>   - Publicity / advertising.
>
> Have you decided any of these already?
> I'll post my ideas on them later, and I'm interested in everyone else's.
> There are some important tradeoffs involved in some of them.
>

Most of these are going to require a bit of discussion to tease out,
and there are some dependencies between them. My friend suggested Free
and Fair as a name, which I quite like :)


> There was a mailling list, free-prstv, for writing an open source
> implementation of PR-STV, which hasn't had any mails receently.
> (I'll get the address later).
> There might be someone on it who might join this project, if we are going to
> support PR-STV.
>

That would be great. Even if we leave PR:STV till later, it's a good
place to look for interested people. Once the mailing list is going,
we can start spreading the word and asking people to join.

Do you have a Wave server running on your host? I think it could be a
very good way to organise our discussions.

Margaret



More information about the E-voting mailing list