The New Republic

Adequate numbers of polling places and longer and advance hours are essential. Plus a requirement that all citizens of a state have roughly equal access so that the poor neighborhoods -- with Democrats -- don't face larger hurdles. I don't know why the New York system, pre-optical ballot, is not the standard. Until the fad for electronics, we had for years these marvelous, fool proof machines. Each party line on a row, the offices and ballot initiatives in columns. You flipped levers down. The did not permit you to over-vote by flipping two levers for the same office. If you want to vote the party line, as I always do, you just go right across in a process that takes less than 10 seconds. You can change your voter until you pull the lever to open the curtain on the machine. I don't think it ever took me more than 30 seconds to vote. Those, alas, are now gone in favor of optical scanners. Much more time-consuming to vote. Even though many can fill them out at the same time, it is slow to do so and then they have to be read before the voter departs. I also saw on one occasion how this machines are tabulated. There are registers on the back. Election workers record the settings before the machines are sent out to polling places. The registers are then mechanically sealed with wires that are crimped shut. The backs of the machines are then locked. Not impossible to commit fraud with these machines, but damn near. It would take an extensive conspiracy to do it. Results are easily read off the registers of each machine.

- roidubouloi

November 8, 2012 at 12:03pm

This is not a technology problem. In Iowa we vote, if we choose, by mail on optically scanned paper ballots. The solution is easily scalable, and provides a permanent record for recount or fraud investigation purposes. There are real problems with American voting, but they are political problems, not technological problems: 1. Access: there is no excuse for inadequate access to voting. Early, by mail, voting ought to be mandatory and uniform. States and localities should be required to have overflow response plans and resources to make sure that long lines are quickly resolved when they do pop up, and they ought to be sanctioned when they don't do this. 2. Voter ID: the voter ID people are right, we ought to require people to identify themselves to vote, but the corollary is that we need to make it easy for people to acquire and maintain voter registration and id. 3. Gerrymandering: gerrymandering steals votes, period. Every state ought to be required to have either a non-partisan approach to re-districting, or proportional allocation of representation. The Supreme Court has already laid the growndwork for this by letting the Voting Rights Act stand for 50 years, and it's one-person - on-vote ruling on the Senate.

- IowaBeauty

November 8, 2012 at 12:09pm

Roi, I remember those same machines here. It could indeed take only 30 seconds to record a party line vote, but it took me never less than an hour to get to the polling place, stand in line, vote, and return home. Plus, more than once, I had to vote on paper provisional ballots when the one machine allocated to my precinct broke halfway through election day. Color me unimpressed with mechanical voting machines. In Iowa, you can fill in exactly one circle on a paper optical reader ballot to record a party-line vote, pop it in the envelope, sign it, and drop it in the mail in less time than it takes to boil water for tea - which you can do at the same time if you want. And, as I say, there is a permanent record of the vote. I don't think there's anything wrong with spending an hour every two years to vote - it's a small price to pay for democracy, and I never missed a national election - but on the other hand, why make it harder than it has to be?

- IowaBeauty

November 8, 2012 at 12:42pm

All for something that is a genuine improvement. Everyone should be allowed to vote by mail. NY has a good registration too. Your signature on the voting roll at the polling place is sufficient ID.

- roidubouloi

November 8, 2012 at 1:21pm

Good overview. We waited about an hour (between the morning & lunch rushes) at our polling place, which could easily been halved or better with a few more workers. Also, our site used to house 2 precincts, but one moved -- without adequate notification. A simple paper sign at the back of the line ("Check your precinct now") would have averted a lot of grief, given that so many folks had smartphones. (Ours is a mixed income area.) Bonus tip for the Republican voter fraud observers trying to go incognito: Not too many matchy-matchy Burberry purse & boot ensembles in this part of the state....

- Wonderland

November 8, 2012 at 1:24pm

I'm also not totally taken with the idea of an electronic solution, but I would point out that countries like Australia did the sensible thing and had a single system built with open source code which allowed a number of interested parties to find a few problems and fix them. Voting machine software is not really something I see the need for private enterprise to get involved with, but then again, I feel the same way about prisons. On the other hand, Obama could strip a number of things out of local control. Design of the ballot for example, absolutely needs to be simple and consistent across the country. This incredibly stupid idea that locals knows best is beyond idiotic. Minimum numbers of polling stations etc. and the funds to support them would be another. An independent electoral commission that manages these things, along with drawing districts following open and public rules based on geography and population density would also help.

- Nari224

November 8, 2012 at 1:30pm

Electronic tabulation of paper ballots seems to make the most sense - it creates a paper trail for recounts and also allows voting to continue if there are technical problems with the electronics at a polling station. I don't see ease of filling out a ballot to be an issue of any sort, although design standards to ensure clarity are needed. The most widespread problem I see is the underfunding of polling stations in many areas. When you consider the fact that over two billion dollars were spent on the last Presidential election it is criminal that we aren't spending more on ensuring that everyone has equal access to polling places. Maybe there should be a tax on campaign spending that would be dedicated to paying for modern and reliable polling stations for everyone.

- Attrill

November 8, 2012 at 2:04pm

The single most important thing, however, is to put an end to gerrymandering. Independent commissions are one approach. However, not the best approach. Better is a mathematical solution that requires congressional districts to have equal population and be maximally compact within states. That means, they have the shortest total boundary length. This is something that computers can now do and should do. It would mean that the voters who share a district are as physically proximate to one another as possible and therefore most likely to share common circumstances. Any opportunity for manipulation for partisan advantage would be eliminated.

- roidubouloi

November 8, 2012 at 2:44pm

Roi, As appealing as your mathematical approach might seem, it is not that simple. There is no unique solution to the partioning problem based on just the constraints "equal population" and "shortest total boundary length" Neither is that a workable definition for compactness or likely to lump people together who "share common circumstances." Believe me, I do this kind of optimization and analytics professionally, and I would love to say there is an algorithm that is obviously fair, understandable and useful across the very different circumstances of states as different as Texas and Connecticut, but there isn't. I think a simple rule that uses a non-partisan commission to create the lines, and requires all-party approval of the results will give better results.

- IowaBeauty

November 8, 2012 at 3:06pm

What's wrong with Florida? They're still counting votes?? California and New York both have a bigger population and they've reported results. New York just endured the superstorm. Fortunately the election doesn't hinge on Florida this year and we avoid the 2000 drama. Perhaps instead of worrying about purging voter lists, the Florida Secretary of State should focus on the state's abysmal performance. Concerning gerrymandering, there's a report that the total number votes cast for Democratic House candidates exceeded Republican by 500,000, but yet the House remains under Republican control

- dubyadoubte

November 8, 2012 at 3:08pm

I believe Florida is waiting for absentee ballots to work their way through the mail. There aren't enough absentee ballots in California and NY to affect the outcome of the Presidential race, so they don't have to wait.

- Attrill

November 8, 2012 at 3:17pm

Check out the Oregon election system - vote by mail. No polling places, no long lines, one can vote early and end robo-calls, no taking time off work to vote. It seems to work and as an Oregonian, I love it.

- deutsche

November 8, 2012 at 3:32pm

Roid -- I'm a New Yorker and also miss the old voting machines, but largely because I loved being surrounded by the grand, red curtain and that great ka-chung sound when you pulled the lever to record your vote. I never had a machine break but definitely heard of it happening a lot, so the optical scanning ballots seems like an improvement. Now they just need someone to create a sound effect that mimics the ka-chung once you feed your ballot into the scanner and it records your vote. But on a more serious note, it does seem like expanding access and ease of voting should be a bipartisan cause, as David Frum has noted. (But I guess he's persona non grata among Republicans so they wouldn't listen to him anyway.)

- shellski

November 8, 2012 at 4:16pm

That's interesting, Iowa, but I'm not convinced. There may not be a closed solution, but computers can do fast enough simulations to be sure they are asymptotically close to the minimum boundary length. You should end up with districts that are very compact, very close to the optimum if not exactly at it. Does geographical proximity ensure common interest? As well as anything else, better than most criteria. As it stands now, there are bizarre aggregations to satisfy partisan interests. Even bi-partisan commissions play games. The parties don't necessarily want competitive districts. I do. They may be willing to agree to a sweetheart deal that gives both parties mostly non-competitive districts. I would rather vote with my neighbors than with someone geographically removed out of partisan gerrymandering, even if it is bi-partisan gerrymandering. I wouldn't be adverse to giving humans a 1% fudge factor, allowing them to have up to 1% differences in population and 1% excess boundary length to satisfy peculiar concerns such as adjustment to natural or other political boundaries.

- roidubouloi

November 8, 2012 at 4:46pm

It is pretty outrageous that the Dems are down 40 seats in the House although they appear to have garnered more congressional votes nationally. That isn't democracy, it is a rigged game. It the one person, one vote decision means anything, this ought to be unconstitutional. Of course, with this Supreme Court, totally unprincipled, nothing means anything.

- roidubouloi

November 8, 2012 at 4:47pm

Roi, In Iowa the maps are developed by non-partisan employees of the legislature who are enjoined by the law to consider respect for existing political boundaries (that is counties and municipal borders), contiguity, and compactness. They are not allowed to have or use data related to party affiliation, voting patterns, or addresses of existing legislators in drawing the map. The legislature then votes to accept or reject a map. The only thing I would change is to make the vote require a supermajority to pass the plan. Could the system be subverted - of course it could, but not easily, and in practice it has worked quite well. As for the algorithm - my point was that given only the constraints you wish to impose, there is not necessarily a unique solution. So how do you figure out which of many possible solutions you use? There seems to me to be a lot of opportunity for mischief there. Obviously you could impose a single algorithm and just say that's what you get, regardless of how dumb the result, but that seems unnecessarily obtuse to me. I'd prefer it to what we have today in TX and PA, but not to what we do in IA.

- IowaBeauty

November 8, 2012 at 5:33pm

"It is pretty outrageous that the Dems are down 40 seats in the House although they appear to have garnered more congressional votes nationally. " It is absolutely outrageous, and had we a Supreme Court that gave a shit about democracy, grounds for a challenge to every districting scheme in the country that was not non-partisan and based on objective criteria of voter equality, based on Baker v. Carr.

- IowaBeauty

November 8, 2012 at 6:15pm

Maybe. I think there is a unique solution to the shortest total boundary, but not an algorithm for finding it. It would have to be done by an iterative process that shows the obtained solution cannot be far from the optimum because there is little variation, meaning asymptotically close. I suppose if the boundaries are really developed without partisan info it can work, but I don't really see why a human solution is better than a machine solution. Given population density variations, it is rarely possible to have congressional districts follow much of other local boundaries unless you are giving a lot of latitude on the requirement of equal numbers in districts.

- roidubouloi

November 9, 2012 at 12:53am

Does someone have a link to an article proving the claim that pro-Dem House votes outnumbered pro-Republican ones? I don't doubt it could be the case, but would like to see the proof.

- Thunderroad

November 9, 2012 at 2:23am

Roi, On real maps, with defined political boundaries you have to follow, and falsely precise populations, there may well be unique solutions. I looked at the problem last night, and can tell you that there is no requirement that this be true - counterexamples with simple maps aren't very hard to find. But it's beside the point. You clearly could specify a deterministic algorithm and use it to derive districting solutions that are nearly optimal with respect to population equality and path length. The solution would be unique for any given data set, whether or not optimality guarantees a unique solution. So what you're proposing is possible. Clearly it would be better than what we have. My long experience suggests it's a tough sell though - it's hard,even in business, where you have clear dollar denominated yardsticks, to get people to trust a black box solution even when it's demonstrably better than the ones they come up with through their own imagination. I'm certain it's impossible when there is no clear yardstick. So, to answer your question "why a human solution is better than a machine solution" - it almost certainly is not objectively better, if your only criteria are population equality, contiguity, and compactness. But the fact that it is a process people understand, and they get to vote on it, makes it a lot easier to "sell."

- IowaBeauty

November 9, 2012 at 8:27am

Thunderroad, I think I saw something on Political Wire to that effect. Google? Iowa, Now we are getting somewhere. I think we agree on the objective aspects. It goes without saying that the solution will be incorrect insofar as the data are incorrect, but that is neither here nor there. The data is the best we have (well, not really, there should be statistical adjustment to census data, another subject) and no one will know inwhich direction things are off or be able to manipulate based on the errors in the data. I do not disagree about your view of the politics. Personally, I happen to have a lot of faith in a proper black box because I have one, designed by me and my partner, that has been successfully trading securities, based on statistical analysis that no human, indeed no team of humans no matter how large. could do in real time. If I were really le roi, I would impose a uniform Federal mathematical standard, for congressional districts as a matter of course and for all local districting as an implementation of Article IV and the 14th Amendment under Baker v. Carr. I would give local jurisdictions a Federally designed program that would solve the problem so they don't have to implement their own and prove it works to, without prohibiting the states from implementing their own solution if they could demonstrate it meets the Federal standard. To address the political issue, I would allow a limited amount of human intervention, or override, that met specific purposes, such as conforming to a local political boundary or geographical feature or transportation issue(so that, for example, some people don't end up in a small enclave on the opposite side of a river, separated from the community in which they live). I would prohibit using partisan data when doing so, as you say is the rule in Iowa. That ought to give enough room to satisfy the subjective concerns. My goal would be not only to end manipulation and achieve equal and compact districts, but to maximize the number of competitive districts without having to gerrymander to that end. The reason the House is elected every two years is so that it can be as responsive as possible to the popular will. Gerrymandering for partisan purposes, or for any purpose, leads not only to unrepresentative outcomes, but frustrates the purpose of allowing people to turn out the government if it is not performing -- or gives the effective decision to the handful of people who live in competitive districts, replicating the problem of a few "battleground" states that results from the Electoral College system. Undemocratic in many ways, not just because gerrymandering is hyper-partisan. If we had such a system, I doubt very much we would see the craziness in the House that we have been living with for the past two years. They wouldn't dare.

- roidubouloi

November 9, 2012 at 9:40am

"I happen to have a lot of faith in a proper black box because I have one, designed by me and my partner, that has been successfully trading securities, based on statistical analysis that no human, indeed no team of humans no matter how large. could do in real time." Yeah, I'm in a similar business, but as a vendor of such solutions. No question they work. We saved a custom manufacturer tens of millions/yr by increasing the utilization of macines in their fabrication plants. People looked at the proposed routings in the plant, and said we were nuts, that they could do better than that with their eyes closed. The solution was adopted only by executive fiat because it demonstrably improved utilization in simulations, even though the routing made no sense to anyone. $1M worth of effort, 20 or 30X ROI the first year alone.

- IowaBeauty

November 9, 2012 at 12:39pm

To answer my own question about a link to an article confirming that Dems won more votes for House seats than Republicans did, here's one: http://thinkprogress.org/justice/2012/11/07/1159631/americans-voted-for-a-democratic-house-gerrymandering-the-supreme-court-gave-them-speaker-boehner/?mobile=nc

- Thunderroad

November 9, 2012 at 1:19pm