Champaign County Illinois, USA uses a pair of ES&S machines to prepare and count (plus physically store) the ballots. Use of the ballot preparation machine is optional—one can fill in the bubbles manually with a pen or pencil. This first machine can also (at the voter’s option) scan a completed ballot and report to the voter how it read the ballot, informing the voter of their vote as well as any mistakes such as voting for too many or too few candidates in a race. But all voters must feed their voter-verified paper ballot into the counting+storage machine. I despise the use of the counting machine.
I also despise that both of these machines run on proprietary software; citizens in Urbana, Illinois are fighting for instant runoff voting (IRV) for local elections. You should help them in their fight. IRV requires voters to rank the candidates instead of voting for one candidate; if the voter’s first choice doesn’t win, the vote rolls over to the second choice, and on down the line. With IRV, voters don’t need to fear that they’re “throwing their vote away” on an unpopular candidate who supports their political goals. IRV is a great step to increasing participation in elections both for broadening the candidate pool and encouraging voting.
Unfortunately there are dark clouds in the forecast: If Urbana and Champaign County stick with computer vote counting, they’ll have to convince the proprietor (ES&S) to change the vote-counting software to work with instant runoff voting. This is one reason I endorse the use of free software, software that respects a user’s freedom to run, share, and modify programs. Urbana ought to have the freedom to get whomever they want to alter the software to support IRV. Urbana can pay to send their modified software through the government-required approval process and then use the software in citywide elections.
The silver lining in this cloud is the Champaign County Clerk, Mark Shelden: When I was part of the recommendation committee that evaluated electronic voting machines for Champaign County, I discussed this issue with Mark Shelden and he agrees that a free software voting machine is preferable. Free software voting systems also mean jobs for our community: Champaign County could become a hub of voting software development. We didn’t have any such machines to choose from back then, and ES&S was not interested in selling us a license to their software under a free software license. But as more people evaluate voting machines and find serious problems with them, I think this position will change.
Clay Shentrup | 11-Aug-07 at 1:27 pm | Permalink
As is typical of the IRV advocates, this post contains typical IRV myths, such as
“With IRV, voters don’t need to fear that they’re ‘throwing their vote away’ on an unpopular candidate who supports their political goals.”
This is incorrect. With IRV, a sincere vote can be a “wasted” vote, such that IRV incentivizes top-ranking ones’ favorite of the *front-runners*, rather than one’s sincere favorite. This helps to understand why IRV has historically produced two-party duopoly, just like our plurality voting system does now.
http://rangevoting.org/TarrIrv.html
http://rangevoting.org/IRV.html
The solution for breaking up duopoly, picking better winners (especially among strategic/insincere electorates), reducing spoiled ballots, and diminishing the importance of cash in elections, is the simpler system called Range Voting.
http://rangevoting.org/CFERlet.html
See an example of what a Range Voting election would look like, here:
http://polls.zoho.com/brokenladder/2008-u-s-presidential-straw-poll-using-range-voting
Clay Shentrup
clay@electopia.org
415.240.1973
J.B. Nicholson-Owens | 11-Aug-07 at 4:45 pm | Permalink
Despite Adam Tarr’s remarkably poor choice of words, I posted your followup in the interests of educating myself and my readers about first-past-the-post alternatives.
In the future, I suggest finding alternate sources to link to or rewriting the arguments about Range Voting to be more approachable to the newcomer. It’s awfully hard to reach people while belittling them (”Nope, logical arguments aren’t going to be enough for you.”) or referring to them as “a suicidal idiot”. It is possible to disagree with someone without running the risk of being dismissed out of hand for one’s approach.
Almost immediately I can see that the Center for Range Voting (CRV’s) claim about working with extant Plurality voting systems is untrue.
Yet even the CRV’s own demo doesn’t bear this out. This demo is said to be a sample range-voting election as one would run on a Plurality machine (”See what this range-voting election instead would be like on Plurality machine” according to the link on http://rangevoting.org/quickdemo/PresBallot.html) but the Plurality machine demo doesn’t look or work like any plurality voting system we tested in Champaign County, Illinois.
Champaign County didn’t pay ES&S big money so we can take on what computer programmers refer to as “a hack”. From one perspective, I can see how the CRV-endorsed transformation system that ostensibly allows Plurality systems to be used (without modification) for Range Voting is a neat idea (using what you’ve got for something new is sometimes an attractive option). But if the interface for using (and possibly counting, if people insist on machine counting ballots) isn’t smooth and seamlessly integrated, the new voting method won’t be adopted. I have to conclude that switching to any other voting system will require software modifications and this places Urbana or Champaign County in the same precarious position with regard to negotiating with the monopolist (all software proprietors are monopolists because you can’t go anywhere else to get the software changed).
The user interfaces are considerably different. The Plurality machine demo has you “award each candidate a numerical score between 0 and 99″ but every Plurality system Champaign County tested has you place a non-numeric mark next to the candidate(s) you wish to vote for in each race (some races let you vote for more than one candidate). Even non-programmers can see how changing marks to numbers will require software changes. The paper ballots Champaign County uses don’t have spaces big enough to hold a clearly legible number (the space where the mark goes is quite small when compared to the space one needs to write a legible number). A new ballot redesign for this will require changes to the current ES&S system software.
What is referred to as an “undervote” in a Plurality system is a valid vote in Range voting. In a Plurality voting system, not voting in a race is something to warn the voter about—maybe they forgot to vote and had intended to. In Range Voting, not voting is valid; specifying no vote is how a voter says “I don’t care to rank that candidate”. You can specify no votes in the aforementioned CRV demo and get no warning (simply visit that link and press the submit button, thus submitting an “empty” ballot). Changing no votes from raising a warning to specifying “I don’t care” will require a change in Plurality system voting software.
Finally, the votes are tallied considerably differently in the two voting systems. Knowing that Range Voting involves averaging numbers from voters means the Champaign County’s extant ES&S software would have to be modified.
As I get more time to spend on the subject, I’ll read more about Range Voting from the Center for Range Voting. But for now, I am left thinking that the CRV should spend some time cleaning up its own promotional material.
Hallett Weaver | 18-Aug-07 at 3:15 pm | Permalink
Nicely put, JB. There’s a reason why no governnment in the US or any nation in the world uses approval voting for any of elections at any level.
IRV has a long history, and well-deserved fans. http://www.instantrunoff.com is a good resource, as you may know, along with fairvote.org/irv
Steven Severinghaus | 18-Aug-07 at 4:34 pm | Permalink
Hi there, Jeff. The pamphlets for the IRV For Urbana movement claim that the machines we use in Champaign county will “accomodate IRV” and that implementation will be provided “at no cost to the county”.
I, too, would be much more comfortable with these sorts of changes if they were open to scrutiny under a free software license, but at least it doesn’t sound like we’ll need to “convince” the proprietor to make the change for us–we just need to request it be done.
Do you have any additional insight or information on that topic?
Thanks for covering this issue!
-Steve
J.B. Nicholson-Owens | 18-Aug-07 at 9:29 pm | Permalink
Hi Steve, yes I do have more to share on this. For those of you who don’t know I’ll give some background: I was a part of the (appointed, not elected) group which recommended an electronic voting machine system (a ballot preparation machine plus a ballot counting machine) for Champaign County, Illinois.
We saw machines from three vendors (ES&S, Diebold, and Hart Intercivic), took a field trip to Tippecanoe County, Indiana (which has been using the then-Diebold machines since they were Global Election machines), ranked the 3 systems we saw, and reported our recommendation to the elected County Board. The County Board made the final decision on which system Champaign County would get (taking our recommendation under consideration) and the County Board signed the final contract with ES&S (which we recommended). I gave ES&S the nod because they were the least worst, not because I’d choose them if I had my druthers. Someday I’ll write about the fun I had with the Diebold reps from a company called Fidlar (pronounced like “fiddler”) and the needlessly complicated Hart Intercivic system. We would have seen more machines if other vendors hadn’t backed out and if the state hadn’t narrowed the choices for us.
I mention elected versus appointed because if the appointed recommendation group had any purchasing/contract power that would be horribly wrong. People with power to make payment and sign contracts must be democratically accountable. Of course, at the time I invited people to come and participate in our meetings. I spoke with people interested in the issue including respondents to my Counterpunch article on the need for free software voting machines, happy to take input or correction along the way.
Mark Shelden and I asked the vendors how much they would charge to change the software to meet our needs. When the younger of the two ES&S reps asked what changes I meant, I said that we couldn’t determine all of our needs up front just as you can’t determine everything you’ll want a software program to do or do differently when you initially design the program. ES&S gave a figure that struck me as about right–I don’t recall it precisely now (and I’m away from my notes) but I recall it being over one hundred dollars per hour (some commercial programming firms get more for modifying more complex programs such as GCC, a programming tool called a “compiler” which is used to build other programs). It was clear that some changes would cost us money and some would be covered under ES&S’ typical agreement with clients (most if not all bugfixes would be covered under any agreement).
When I asked for specifics on past customers who wanted voting system changes our ES&S reps couldn’t name any customers who had requested this. The scope of what ES&S was willing to change was not clear, but I came away thinking that voting system changes were sufficiently complex that this was going to cost us money and possibly not happen if the proprietor decided that they didn’t want to do it. Simply put, all of the details were not forthcoming from any of the vendors.
Toward the end of our discussions on what we had seen, I reiterated my warning to the group and the assembled vendors about changes to voting machines saying that we were essentially picking a monopolist. The county has nobody else to go to for changes on this order of magnitude (as opposed to making sure someone’s name is spelled properly on a ballot. So it was troubling that we knew so little about how much this will cost up front.
It’s entirely possible that changing the voting system is listed in the contract. But I’d want to know precisely what is covered before I’d concur that this has been sufficiently taken care of. I have to go based on my experience with ES&S’ reps so I remain skeptical until I see more details.
Assuming the Urbana IRV proponents are correct, it still sets up a disturbingly short-term affair: Let’s assume that Urbana votes to change city-wide elections to use IRV and let’s assume that ES&S makes the appropriate changes gratis. What about next time? Electronic voting machines are intended to work well past the lifespan of a home computer, so the next big change could come up decades from now when this issue has been forgotten by being the norm, an ordinary part of the firmament of city civic life. This is the timespan we had to consider when weighing what we were hearing from vendors (which is an unusual situation regarding anything computerized). Will ES&S be willing to do more work for no additional money? Under the plans we discussed during vetting the machines, the county would only get source code access if ES&S went out of business (admittedly narrow terms, but I tried to do far better). If ES&S is in business for another 50 years will ES&S distribute source code under a free software license under some condition that comes up sooner? It’s hard to plan in detail for something that could come up in 30 years but I’m very uneasy about placing important systems at the mercy of lucky breaks, as I’m sure are all of you.