If you could, did you VOTE?

[Inquartata]

Why can't we have the same standards of accuracy in voting that we have in basic econimic transactions?

Because unlike commerce there is no profit motive in voting. In the public sector, the invisible hand is weak, crippled and withered. ( Hence all those feebly punched ballots in Florida four years ago. )

[BOliver]

Because unlike commerce there is no profit motive in voting. In the public sector, the invisible hand is weak, crippled and withered. ( Hence all those feebly punched ballots in Florida four years ago. )

I agree that there's less money involved in the election process than, say, Visa, but I read that over $4 billion was spent on this election. I'd think that such a tremendous expendature of cash would motivate the candidates to push for greater accuracy.

[jeff]

This touches somewhat on my professional activities, so....

MasterCard, Visa International, American Express and similiar have had many years to refine credit card authorization and funds transfer (a web transaction is similar to a pre-Internet "card not present" transaction, similar to sales made over the phone or via mail). It's taken a long time for them, and there are sharp immediate losses due to fraud hence motivation. Yet, it still happens: identity theft and credit card fraud are very serious problems, so it's not a solved problem in the credit card world either.

The problem is a different from voting: the risk is limited by the size of the size of the transaction and credit limit and risk is pushed back to merchants. Transaction history can be used to detect exceptions from a pattern: if Grandma in Grand Rapids suddenly starts creating large purchases in Mozambique it's possible to deduce it might not be her, and ask for further proof of identity (as in the TV ad where somebody attempted to buy surfboards on a stolen credit card). This klnd of pattern matching just started in the '90s - I think AmEx was first - and it's not the kind of thing you can apply to voting: we don't keep track of people's prior votes (thank goodness), and even if we did it wouldn't help: it would just mean that somebody changed affiliation, a normal part of voting, not a sign of fraud.

Best of all: there's a paper record the customer can use on a monthly basis to see if there are unexplained debits. That's why the experts recommend a paper trail for voting as well.

Voting presents additional problems: your credit card company issues you a card with a number and a PIN, but we don't have an airtight national ID system - Social Security isn't a substitute. With voting you have to make sure that the person at the booth is properly registered and is voting at most one time (tough without an ID number), while with credit cards they're happy to issue you multiple cards so there's no requirement of uniqueness. Also, the vote is over with right away: you don't have the chance to get a charge reversed days or weeks later. There were plenty of voting "anomalies" in this election, as in elections past - but the election's over and what's done is done.

In summary, this is harder than it looks, and the efforts to do it in any serious manner have only recently started and are immature. For people interested in this subject, I suggest going to the essential "Forum On Risks To The Public In Computers And Related Systems", at http://catless.ncl.ac.uk/Risks/ It's not exclusively about electronic voting, but it's (ahem) been an active topic recently. It's an eye-opening mailing list to read regardless of whether its voting season or not (and you'll probably be made more nervous about your credit card transactions and use of Windows, too). A copy of the current issue can be seen at http://www.csl.sri.com/users/risko/risks.txt with comments on electronic voting by moderator Peter Neumann. I also suggest reading the document at http://www.notablesoftware.com/evote.html plus anything by Bruce Schneier.

To Inq's point: there is tremendous money involved, but it's a consequence of the policies of those who get voted in, and not directly and measurably tied to any individual voter. This means that financial or political incentives for fraud are not counter-balanced sufficiently by the individual voters ability to even detect when an action has been made to his or her disadvantage, as is clearly visible on a credit card statement.