Okay, to clarify: My source for the comment about why Australia hasn't adopted voting machines was an article published in the late 1970s or early 1980s (sorry, can't recall more details) where the author, an Australian, commented that the voting machines used by most US States at that time -- correct me if I'm wrong -- usually had one lever per candidate and, each time a voter pulled that lever, a mechanical counter turned the tally dial one for that candidate more click. Once polling closed, officials simply unlocked the machine and checked the numbers on the respective candidates' dials.

Now, obviously this works fine when you choose only one candidate to fill one seat, or even three candidates (at once) to fill three seats. But when the voting system requires voters to indicate an order of preference, it becomes more complicated both to cast and count votes, because you'd need to know the number who voted for A only, then the number who voted for [1] A and [2] B, the number who voted for [1] A and [2] C, and so forth. In an STV election with (say) 20 candidates for 5 seats (which is the average in Ireland and Malta, I understand), you would have thousands of possible permutations -- making it not feasible for a "click the counter" mechanical voting machine.

The only way these older voting machines might be able to accommodate preferences at all would be to limit the number of preferences, eg to a maximum of two (as in London Mayoral elections), three (as in Sri Lankan presidential elections) or four (as in San Francisco elections -- is that correct?). The machine could then have two/ three/ four columns and the voter would pull the lever for a different candidate in each column. This would make a rough form of preferential voting feasible for single-seat elections, although with many candidates the candidate with most votes on the final count might still fall short of 50%, and it still wouldn't be able to deal with the fractional values of transferred surplus votes in a multi-seat election using proportional representation.

Obviously, though, more recent technology -- eg, optically scanning the ballot papers, entering the voter's ranked list of candidates directly via a keyboard or touch screen -- would make preferential systems more feasible, even allowing voters to indicate preferences for all the candidates. So the complexity argument might no longer hold, although the concerns about cost and fraud might remain.

At 10:48 24-09-2002 -0500, smulroy@memphis.edu wrote:

I'm not sure I understand why paper ballots would be more appropriate
for single transferable vote systems than they would be for plurality
voting.  I would think that the vote tranfers are sufficiently
complicated that the use of a computer tabulation would really speed
things along.  Conversely, plurality counting is simple enough that it
could be done relatively easily and quickly by hand using paper
ballots.  What am I missing?

----- Original Message -----
From: Tom Round <T.Round@mailbox.gu.edu.au>
Date: Monday, September 23, 2002 5:45 pm
Subject: Re: paper ballots

> Addendum to Graeme's post: The Australian Capital Territory (our
> smallest
> jurisdiction in both population and area -- constitutionally
> equivalent to
> Washington DC, but socio-economically equivalent to Cambridge,
> Massachusetts) experimented with touch-screen electronic voting at
> its
> latest Territorial election in October 2001. The initial press
> releases
> were at
>
> http://www.elections.act.gov.au/media9901.html
>
> http://www.elections.act.gov.au/media0008.html
>
> http://www.elections.act.gov.au/media0104.html
>
> -- although these are old links: I could find more recent details
> if
> anyone's interested.
>
> Like Cambridge, the ACT uses single transferable vote in multi-
> member
> constituencies (5, 5, and 7 seats), which makes them even braver
> for
> experimenting with electronic voting -- since one common reason
> offered in
> Australia for keeping to paper ballots, apart from cost, is that
> all our
> jurisdictions use preferential voting in some form. Voting
> machines as used
> in the USA do seem to simplify voting when a society is using
> first-past-the-post for a number of positions, but adapting them
> to
> preferential systems would be more complicated.
>
> At 12:12 20-09-2002 +1000, Graeme Orr wrote:
>
> >Australia too is all paper-ballots.  And we even provide pencils,
> not
> >pens, to mark ballots!    The AEC long ago realised the best way
> of
> >avoiding Floridian debacles was to stay with the simplest,
> 'failsafest'
> >technology.       The anti-fraud safeguards, they would say, lie
> in
> >systemic reform - eg in having the most independent,
> incorruptible,
> >professional and co-ordinated (if not nationalised) electoral
> >administration you can achieve.    At least no-one would claim
> >Canadian>
> >(Paper ballots also ensure that the bedrock search for 'the
> intention of
> >the voter' remains central.  Australians by and large found it
> hard to
> >fathom how US electoral law could deny Gore's plaint, at least if
> he had
> >been demanding a manual recount of the whole state - which after
> all was
> >the relevant electorate, and not just selected counties).
> >
> >Graeme Orr
> >Lecturer, Law
> >Griffith University
> >Brisbane  4111
> >Australia
> >
> >Rick Hasen <rick.hasen@lls.edu>
> >Sent by: owner-election-law_gl@majordomo.lls.edu
> >
> >20/09/2002 03:42 AM
> >Please respond to rick.hasen
> >
> >         To:        "election-law@majordomo.lls.edu"
> > <election-law@majordomo.lls.edu>
> >         cc:
> >         Subject:        Computers Criticized in Fla. Voting
> >
> >Here's another call for paper ballots.  Isn't anyone worried
> about the
> >potential for fraud in the hand counting of all ballots? I know
> Canada>conducts its national elections with paper, but they don't
> have the same
> >history of voter fraud.
>
>