The trick now is that a lot of people don't believe them.
I write software for a living. This sort of thing happens all of the time, and generally has less to do with malice and more to do with a mistake in programming. It's easy to tweak a variable--especially in any software that involves pattern recognition--or implement a filter that "seems like a good idea at the time" and find that it has echoing consequences that you didn't predict. It is not hard for me to see how the following sequence of events might have happened:
Amazon decides to implement/tweak an adult filter for some reason. I can think of (quite a few) business reasons they might decide to do this.
They implement/tweak the filter, run it against a test set of data, see that it looks good. It doesn't get rid of "The Last Unicorn" or Chaucer's "Canterbury Tales," but it gets rid of some random "Big Book of Erotic Stories." Great, it works! Let's deploy it!
Unintended consequence: A bunch of gay, lesbian, etc books get deranked. People call in protest, twitter ends up with #amazonfail, within a day the thing has spread to near-epic proportions and people are calling in protest.
Amazon's PR people do what PR people do: they reach for the nearest ready explanation for the first few people who call and recite the form response. "We are implementing something to protect teh children." This--predictably--only inflames things further since there is a real bug lurking in the code.
Someone actually looks at the number of calls and the books being deranked (e.g., Brokeback Mountain, I mean, really, how can anyone think that was deliberate?) They realize there is a bug. It gets put on their bug tracker and they give a press release saying they'll fix it.
This is a wholly predictable response. I've seen this happen with software designed for customers before: someone made a change; client calls back and says its broken; immediate response goes back that it isn't broken, it's a feature; client provides copious documentation on how it is broken; software company says "well, whoops, damn... we'll get right on that." Fairly normal, I've seen it multiple times in companies of all sizes, it doesn't surprise me that it happens at Amazon as well.
So why are people convinced this is being done deliberately? One person on LJ commented "Interesting how as soon as they're caught red-handed it becomes a 'computer glitch.'" A lot of others came out with similar statements (e.g., #glitchmyass), but really, as I mentioned before: if this weren't a glitch, would Brokeback Mountain really have been on the hatchet list? Really? I have trouble believing that--especially when a movie has been made which to the best of my knowledge was not deranked.
As Robert J. Hanlon put it: "Never attribute to malice that which is adequately explained by stupidity."
The Wild Hunt blog listed some of the deranked books that still remained deranked. As of this moment, Brokeback Mountain, Gay Witchcraft, Rites of Pleasure, Sexy Witch, The Double Goddess, and Lesbian Rites are all, at the time of this writing, ranked again.
So two possibilities remain: Either it is not a glitch, but they learned their lesson and are unlikely to repeat it or (much more likely) it was a glitch, and they have modified their test-cases to try and keep it from happening again.
Either way it is being fixed as we speak, so can we please stop trying to burn them in effigy?
Further Reading
AmazonFail: An inside look at what happened, which gives an extremely plausible scenario for what actually happened, based on an inside source
The day Google banned the internet
The Failure of #amazonfail by author Clay Shirky, which provides an excellent conclusion to the moral outrage we've seen.
Posts
Meh. I don't use Amazon for other reasons which are largely personal biases, but I was not convinced that this was some evil, anti-GLBT conspiracy, either.
ReplyDelete