> The data were quickly recalled. The official story is that the app's developer had mistakenly sent out the 2008 election results as part of a test. But that's a bit flimsy, given that the released totals show the candidates from this week, not from 2008.
This is horseshit.
I spent years building US and international election results maps for a large company, and I had exactly the same thing happen to me on at least one occasion.
When you build an election map or any kind of election reporting site or app, you have to have test data ahead of time. You can't use test data from a previous year, for the simple reason that it has the wrong candidates and parties.
How do you know your code even works with the current candidates? Maybe there's an encoding problem with one of the candidate names this year. Happened to me.
You have to use test data from the current election, yes, the election that has not yet happened. Because you have to test your app with the current candidates, photos, parties, news feeds, electoral boundaries including all the latest redistricting, and all of that.
So yes, it is made up test data, with "predicted" results based on whatever. Recent polling plus a good dose of randomization, perhaps.
You just hope and pray that your test data never leaks out onto a live feed.
But it's a balancing act. You need to test your code on real devices, real browsers, and the whole works. And you have to be ready to swap in your live feed election night and have it all work seamlessly.
Should be easy, right?
Now consider the dynamics of election reporting. You have a deadline. A deadline that will not budge. You do your best, but you're probably not going to get much sleep the week before the election. And sometimes you make a mistake.
That happened with one of the US primary elections. We had a test feed leak out onto a live page for a couple of hours, and man did it make the news. They said we'd released the election "results" early - i.e. we'd made up the results.
Well of course we made up the results. It was test data, and we had to do it that way. So yeah, sorry we goofed, but anybody with an ounce of sense who wasn't looking for a news scoop should have realized that the election hadn't happened yet and it was just a stupid bug.
People talk about deadline pressure. You should try election work sometime!
If the data was generated for testing and escaped accidentally, why not say that? Their claim that it's last year's data runs directly counter to this perfectly reasonable explanation.
That's a very good point, and of course you flag your test data with a special "TEST" flag.
Now remember the lack of sleep for a week before the election, and imagine a possible bug where your client code somehow fails to put up that big TEST DATA message that you thought was there.
I'm with you on taking the more cynical view on this, but there are plausible explanations. Perhaps they based their testing on 2008 data and just people wouldn't understand the testing process so didn't include that part in the explanation.
It's certainly plausible it could have been a screwup using test data. However given the situation in that country it also seems plausible that it could have been a screwup with the fraudulent data. The article did say "flimsy" after all not "demonstrably false."
"In 2009, following his reelection as president, Aliyev passed a referendum which removed the presidential consecutive term limit, thereby allowing him to run for president as many times as he wishes. Opposition claimed this to be a violation of the Azerbaijani constitution and the European convention on human rights.[13] It is widely expected that Aliyev will win the 2013 presidential elections.[14]"
Those "plenty of things" change dramatically from one news source to another. I stopped listening/reading to political news 3 years ago and I feel so happy.
Definitely! It's plausible to me anyway, having been there and done that. :-)
But I can imagine myself in the place of a PR person who doesn't fully understand the technical issues. And somehow it sounds better in the moment to say "that was real data from a few years ago" rather than "that was fake data that we just made up!"
Of course I don't know what actually happened in Azerbaijan, but I can sure relate to the possible scenarios.
This is horseshit.
I spent years building US and international election results maps for a large company, and I had exactly the same thing happen to me on at least one occasion.
When you build an election map or any kind of election reporting site or app, you have to have test data ahead of time. You can't use test data from a previous year, for the simple reason that it has the wrong candidates and parties.
How do you know your code even works with the current candidates? Maybe there's an encoding problem with one of the candidate names this year. Happened to me.
You have to use test data from the current election, yes, the election that has not yet happened. Because you have to test your app with the current candidates, photos, parties, news feeds, electoral boundaries including all the latest redistricting, and all of that.
So yes, it is made up test data, with "predicted" results based on whatever. Recent polling plus a good dose of randomization, perhaps.
You just hope and pray that your test data never leaks out onto a live feed.
But it's a balancing act. You need to test your code on real devices, real browsers, and the whole works. And you have to be ready to swap in your live feed election night and have it all work seamlessly.
Should be easy, right?
Now consider the dynamics of election reporting. You have a deadline. A deadline that will not budge. You do your best, but you're probably not going to get much sleep the week before the election. And sometimes you make a mistake.
That happened with one of the US primary elections. We had a test feed leak out onto a live page for a couple of hours, and man did it make the news. They said we'd released the election "results" early - i.e. we'd made up the results.
Well of course we made up the results. It was test data, and we had to do it that way. So yeah, sorry we goofed, but anybody with an ounce of sense who wasn't looking for a news scoop should have realized that the election hadn't happened yet and it was just a stupid bug.
People talk about deadline pressure. You should try election work sometime!