Comments | bryant: Amazon Troll Busting

bryant

Amazon Troll Busting

Apr 13, 2009 12:20

The original post which prompted this one is now inaccessible accessible again. C'est la vie.

Hopefully you weren't thinking of this as a news post, but just in case ( Read more... )

Leave a comment

Back to all threads

jonthegm April 13 2009, 16:52:23 UTC

The code he POSTED didn't work... but you can immediately see that given a slightly more powerful perl or python script that you could do it without links.

bryant April 13 2009, 16:55:03 UTC

I do! Heck, I wrote that code in about five minutes while I was fooling around with this. I'm willing to believe that he screwed with it for some reason.

Buttt the automated complaint page doesn't exist, which is a bigger problem.

jonthegm April 13 2009, 17:07:18 UTC

What about the tagging system? Couldn't that be the culprit as well? (This hinges on Amazon having a heuristic that says "adult" tag combined with some other tags == highly likely to need blacklisting... therefore blacklist and send for review)

It seems a bit easier to blame a Bantown style exploit the more you look at it. Even if this guy is just a fake.

bryant April 13 2009, 17:14:56 UTC

Oh, I will not at all be surprised if it turns out to be a third party responsible. Users can absolutely tag items, and if there's something turning those tags into decisions about which books should have sales rank? That'd explain it.

The tag in question would have to be something fairly innocuous, like "lesbian," given that Heather Has Two Mommies was de-ranked. It didn't have any adult tags, etc. -- lesbian was the one tag I could see which could be interpreted by someone dumb as adult content. So I dunno. Maybe.

But at this point there's nothing conclusive -- the post in question has nothing that would prove that the author was responsible. This puts it back in the pure speculation category. I'm holding off on pretending I know what happened until there's something real to bet on. Maybe Amazon screwed up; maybe Amazon was malicious; maybe someone else was malicious. Lord knows.

lynxreign April 13 2009, 17:28:34 UTC

I've read that part of the problem is that publishers can add tags and Amazon adds tags. The publisher's tags for many of these contained words that those that got through did not.

bryant April 13 2009, 17:35:00 UTC

That's an interesting differentiation.

lynxreign April 13 2009, 17:38:32 UTC

If I were a publisher and I didn't know that certain tags are being filtered out and especially if I don't know much about how computers work, I'd tag books with every tag I could think of that remotely has something to do with the book so it'd show up in more searches. And if I thought "Adult" meant "Not a kids or YA book" then I'd throw it around with wild abandon.

bryant April 13 2009, 17:40:13 UTC

Heh, yeah. And if the filtering occurred on one set of tags but not the other? Maybe.

There are so many possibilities. It's way more fun thinking about them than it would be if we were sitting there trying to fix it.

... man, they may not even know yet.

lynxreign April 13 2009, 17:46:02 UTC

Yeah, I was thinking they currently have 2 problems and I'm really happy I'm not trying to hunt down either one. There's the tech problem, how was this done and how do we prevent it from happening again and there's the human problem, was this achieved by someone outside, someone inside, our policies and how do we prevent that from happening again. If it turns out it was an internal person acting technically within policy that's a royal mess to clean up, far harder than any code fixes.

sylvanstargazer April 13 2009, 23:08:02 UTC

What I would do if I were slightly more bored is mine the tags, categories, publishers and so on from the stripped books and do a correlation to see where it came from. My initial suspicion was that categories are based on hidden tags, which is why they are the most visible shared features, and that it's something like, "if a book has at least two of the following tags it is flagged."

Then something like "sexuality" got added to the list, but it is used in ambiguous contexts by publishers to mean "sexual orientation" rather than "sexual nature", and it rapidly spiraled out of control. Of course, the real answer is to not censor book results, and certainly not by the "remove their Amazon ranking" hack. Really people, add the extra boolean and scatter if statements liberally. It may be ugly, but if they had done so it might never have been noticed.

bryant April 13 2009, 17:55:10 UTC

More on this:

http://dearauthor.com/wordpress/2009/04/12/amazon-possibly-using-category-metadata-to-filter-rankings/

lynxreign April 13 2009, 18:37:57 UTC

Now that looks like it makes sense. From a "why were certain things filtered" sense anyway.

sylvanstargazer April 13 2009, 23:08:57 UTC

Not only that, but I had the initial thought of "you know, if such capability did exist, I would abuse it like so..." and it's pretty similar to what he threw together. I assume he had the same thought, tossed in some homophobia to play off the panic and figured he'd see how much attention he'd get.

Personally I just went and tagged Ann Coulter's books as "sexuality", "bestiality", "hard core porn", and "adult content". Apparently other people had the same idea, since today the sixth most popular tag for her latest book is "gay porn". If my beliefs are going to be tagged as adult content, I certainly want everyone I disagree with to come with me.

Back to all threads