Voters News Service: What Went Wrong

ddtstudio writes "Baseline Magazine has a pretty good recounting of how even the national TV networks can have a computer network go wrong -- in this case the night of the last U.S. election. From the article: "VNS had been trying to rewrite and retool the system for years. This was just the most recent attempt and it failed miserably." Oracle, IBM, BEA Systems -- all crashed."

Perceived bias also doomed VNS?

[MtViewGuy]

I think another thing that doomed VNS is the fact that it might be possible the way the system was programmed could have been biased towards one political party or another. Unfortunately, this could have bad effects on the election, as the 2000 Presidential election fiasco showed.

Come to Brazil and see it working!

[lotrfan]

Just to mention, aside from the obvious advantage of our elections in Brazil over the US elections: the TV networks could manage to deliver almost instant voting data for the public, including statistics, pre-voting predictions and so on.

If the USA voters want a clean, fast and effective election, send the people responsible for it to Brazil, put your pride away and admit it works nicely.

Re:Oh BooHoo

[mccalli]

So, what this means is that people were able to go late to the polls, and cast a vote free from the influence of network

I'm hoping someone with more up-to-date knowledge will fill in for my sketchiness here, but...

In the UK, there are laws about broadcasting political material during (and I believe immediately preceding?) an election. Additionally, I seem to remember that you are not allowed to report on the progress of that election whilst the voting booths are still open. I'm open to correction on that last point though - I'm sure some news programmes broadcast latest exit polls during the last few General Elections. However, it's a rule I definitely recall from somewhere.

Regardless of my shaky memory, they both seem like a very good rules to me. An election's point is not to win ratings for some TV programme, and it really won't kill you to know the result a couple of hours later.

Cheers,
Ian

Re:Oh BooHoo

[jonathan_ingram]

In the UK the media are not allowed to report any exit poll information until *after* the polls have closed, precisely in order to remove any possibility of the media influencing the votes of the populace. I'm very surprised that the same isn't true in the US.

Of course, we don't have the amount of different time zones in this country, so we don't have quite the pressure for early information to satisfy the ravenous need for statistics.

Re:First hand account

[Anonymous Coward]

tell us more! tell us more! like specific details! about the systems the causes, etc!

and if you have an account, and your uid is triple digit or less, then i'm blaming the fiasco on /. distraction!

Re:Oh BooHoo

[Anonymous Coward]

I guess this would be why the British people still have a monarchy.

A constitutional monarchy. As opposed to the US patriarchy...? Remind me, who was Bush's dad again, and who ran the state which made the court decision over the recount...? A shining example of democracy in action it was, oh yes. Heartwarming to see.

Re:Perceived bias also doomed VNS?

[jayayeem]

I certainly perceived bias in VNS when I read that article. In both the 2000 presidential and 2002 North Carolina senate races, the system erroneously showed Democrats winning against republicans. Since this corresponds to the widely perceived bias in the media, it could easily look like a fix to a lot of people.

Re:Oh BooHoo

[sql*kitten]

In the UK, there are laws about broadcasting political material during (and I believe immediately preceding?) an election

I don't know about any laws, but there is certainly an unwritten rule that the BBC will broadcast whatever political material the Labour party tell it to.

Rrrriiiggghhhtt...

[Talisman]

Assuming you aren't grossly misinformed about Brazil's voting system (which you probably are), they have much bigger problems to deal with.

For example, what good is a technologically sound voting system when all the candidates are shit?

I guess if you don't mind your savings account being frozen by the president [civnet.org] (de Mello), or a 35% currency devaluation [cnn.com] (Cardoso), or a president without a high school diploma [worldpaper.com] (da Silva), it's not so bad...

And I won't even start on the rampant corruption in Brazil. Slashdot's database wouldn't be able to hold so much information.

We'll put our pride away when Brazil puts away its complete joke of a government and stops forcing its masses to live in abject poverty [oecd.org].

You can lecture us on technology when Brazil stops doing asinine things like blowing up its own oil platforms [acusafe.com].

Verdade?

Talisman

Wanna get pissed [remail.org]?

Actually I think the VNS needs to be saved or...

[cottonmouth]

perhaps we should go back to using hand written ballots. The computerized voting that this administration is pushing will only lead to more vote fraud because the count will not be audited. In the 2002 elections in 5 different states Republican candidates won by the same score of 18181 (a prime number). Now, I am willing to give Diebold the benefit of the doubt and say this is a programming error but the fact that this isn't being discussed in the media is a problem. This voting machine code should be GPL'ed so we can all look at it and make sure it works. If they do go to computerized voting there needs to be an audit trail.

Re:Oh BooHoo

[kmellis]

In the UK the media are not allowed to report any exit poll information until *after* the polls have closed, precisely in order to remove any possibility of the media influencing the votes of the populace. I'm very surprised that the same isn't true in the US.

Well, we have that pesky Constutional guarantee of "freedom of the press". You can ask people how they voted (why wouldn't you be able to, and why wouldn't they be able to tell you?), and you can tell other people what you find out.

This is a good example--of which there are many, many more--of a situation where the strict and broad Constitutional prohibition makes less sense than a nuanced and particular law tailored to the situation. It would be better if exit poll results could be suppressed.

The thing that non-USAians don't quite understand about the USA and USAians is that built into the very fabric of our culture is a paranoia about abuses of power by the government. (Periodic lapses into naive trust during wartime, like now, notwithstanding.) All of the Bill of Rights are built upon the same sort of slippery-slope thinking that the gun rights folks use in talking about the Second Amendment: if you cut holes into the brick wall of blanket protections, the government is sure to come barreling through and effectively destroying the whole barrier. How libertarian-minded conservatives can tolerate Ashcroft is beyond my limited ability to comprehend human irrationality.

Anyway, I'm pretty sure that the reporting of exit poll data has been legally found to be protected speech in prior law. I could be wrong. A better answer is just to encourage a civic-minded sensibility among the news reporting agencies so that they voluntarily refuse to report exit poll data until after the polls close. Or even after all the polls close.

Quality of data issues

[Anonymous Coward]

I read somewhere that the famous "Dewey vs. Truman" predictions were skewed by the fact that telephones were something of a luxury at the time, so telephone polling resulted in more Republicans than Democrats responding to a survey. The theory is that Republicans were over-represented in the survey, thus leading to the "Dewey Wins" newspaper headlines.

If you tried a telephone survey today, you would discover that Republicans are likely to have money to spend on "Caller ID" and "Privacy manager". Now it's the Democrats who are over-represented in a telephone survey.

Even if you bag the telephones entirely and survey people outside the polls, the Republicans are perhaps more likely to decline a survey response because of the steady stream of marketing that is directed towards people with money.

Some times the old ways are best

[mwood]

Ya know, the news agencies *could* just wait and announce the *real* results after *all* of the votes are counted, instead of spending millions on guessing. I never believe their conjectures anyway; I always wait for the next day's paper to tell me what actually happened.

Agile anyone?

[john_roth]

This is utterly predictable. Long projects with a
single deliverable at the end are notorious for
not delivering on time.

Several things come to mind immediately from the Agile methodologies
playbook:

1. The customer should not set technical requirements.

2. A working (but not feature complete) version
of the product should be delivered no less
frequently than every three months.

3. The customer should set business requirements
with one voice. If that means that the various
customers have to vote on what's most important,
then so be it.

4. Features should be implemented in the order
that it's most important to the customer.

And we haven't even gotten into the software
engineering yet!

John Roth

Why so quick to dump the mainframe?

[bigirondawg]

I've seen this story, in one version or another, many times over the past decade or so - Some executive leader of a company or organization learns that the system their service is based on (which has probably been up and working nearly as long as they've been alive) is running on a mainframe, and they sniff their nose in horror and say, "Mainframe?!??!!! OS/390??? That dinosaur? We must get rid of this junk immediately!"

Then, they proceed to fix the service that's not broken by a)completely junking the proven, tested old system before a quality, fully-tested replacement solution is ready, and b)leaning hard on their (poor, overworked ;) programmers to slam the new service into place in less time that it probably took to upgrade the OS on the old solution. That mold has never worked, and I think never will.

I'm not a troll, so I won't dwell on how Java (and WebLogic) runs well on OS/390, and Linux runs on the mainframe just as well as on any other platform (and Java and WebLogic run there,of course, also); but those solution possibilities are there, needless to say.

Even if they were going to replatform the whole system, why in God's green earth did they junk the old system before the new system was in place? I mean, it doesn't take a brain surgeon to figure out that you don't completely scrap your #1 bread-and-butter application before its replacement is ready and in place. Even if the new system would be light-years better... some information is better than no information (from the point of view of the networks)!!

I agree with many of these previous posts... this is, among other things, a bad case of project managers and clueless executives getting caught with their pants down -- big time.

Re:Oh BooHoo

[Anonymous Coward]

Up until the 1960's, most US citizens were able to vote just fine, all by themselves, without the need for knowing why *INSERT DEMOGRAPHIC HERE* people voted for *INSERT CANDIDATE NAME HERE*

Up unitil the 1960s, most US citizens to vote just fine for *INSERT POLITICAL BOSS' SLATE HERE*, it was just a matter of *INSERT TOKEN PAYMENT HERE* or your *INSERT BODY PART HERE* might get broken.

Seeing political candidates marketed like soap is depressing, but let's not go overboard and glorify the bad old days.

And anyone with the slightest clue about historical ethnic and racial divisions in this country wouldn't so easily dismiss "demographics". Enjoy your Beaver Cleaver fantasies.

IBM Didn't Let 'em Down

[Bob Uhl]

Contrary to the write-up, if one RTFA one finds nowhere is it stated that International Business Machines let anyone down. The only mentions of IBM are that the old system used DB2 (along with Oracle 7) and S/390 mainframes, while the new system used Oracle 9i and BEA WebLogic (on what platform, it doesn't say).

Reading further, I think that one can tell what went wrong with this project: rather than relying on proven technology, they wanted to make it all snazzy: voice recognition, Java, web application, XML &c. &c. &c. ad infinitum. Instead of sticking to what works, they went with what doesn't. It's like replacing a rock-solid program written in Lisp and running on a Unix system with something written in Visual Basic written on Windows. Don't have high hopes: it may run as well or better, but I'm not betting on it. The likely answer is that this system was over-designed and under-implemented. Too much fun, cutting edge technology and not enough old-fashioned engineering.

Disclaimer: I work for IBM--when I saw the writeup, I read the article. I have nothing to do with our OS/390 division or our DB2 division. I'm a Unix admin, that's all.

VNS has its own structural problems

[real gumby]

VNS itself is screwed up. It's a cartel of newspapers and news services so they can cover lots of elections. Sounds good, right? But they normally only cover the "real candidates" -- so that a two-party race between a and a green will be reported as " candidate running unopposed."

This combination leads to skewed, pro-establishment news reporting.

So I wouldn't be surprised at all if they had a specification problem (as reported by the message from the guy who worked on it). It's completely consistent with the charter of the organization.

Re:Oh BooHoo

[Anonymous Coward]

Quote from article: "Also, the networks would be unable to give the type of detailed explanations as to why voters voted the way they did this time around. For example, according to TV network analysts working the election, the networks wouldn't be able to tell viewers why particular demographic groups voted for specific candidates nor the issues that they considered most or least important when voting."

So, what this means is that people were able to go late to the polls, and cast a vote free from the influence of network prognostication.

No. What it means is that we lose [slashdot.org] what was probably the only [slashdot.org] audit [slashdot.org] system [slashdot.org] we [slashdot.org] might [slashdot.org] have [slashdot.org] had [slashdot.org] for [slashdot.org] electronic [slashdot.org] voting [slashdot.org].

More suspicious [alternet.org] folk can speculate about connections between this fiasco and the many other things [alternet.org] that Battelle Memorial Institute [battelle.org] does [nbcindustrygroup.com] for our government, and how well it does them.

Several groups already systematically Lying to 'em

[Ungrounded Lightning]

As a Western-state voter, I fully intend to lie to any exit pollers who ask me how I voted.

There are already several groups systematically lying to pollsters - especially exit pollsters, but also telephone pollsters, etc.

One of the points is to foul up the "parallel election" you mentioned, in the hope that they'll either stop it or be discredited and ignored.

But others are making various political points - again trying to get the pollsters to make wildly inaccurate predictions and lose credibility with the public (some of whome try to conform with the crowd and/or "vote for the winner") and/or with politicians (who, once elected become isolated from their constituents and tend to believe the media and the polls).

Meanwhile: When listening to poll results, take careful notice of WHICH polling outfit is doing the work. Some polls (including most that hit the airwaves) are commissioned by people who want to create an image and use it to swing popular opinion and/or legislators, rather than to actually measure public opinion. And some poll operations cater to this market to make money.

My impression:

Zogby: Actually tries to predict elections, and does perhaps the best job of it.

Field: Their results seem to completely mirror the Democratic Party line and are often wildly at odds with the actual results once an election is held.

Gallup: Depends on the poll and the season. Better than Field come election time (though not as good as Zogby) but still seems to whore on other issues between elections.