The 2011 DARPA Shredder Challenge | Hacker News Reader

The 2011 DARPA Shredder Challenge

medium.com

122 pointsjarcane28 comments

I was on the team that won the shredder challenge. I talked to one of the UCSD people afterwards and he practically convinced me that my team did it. After a while I realized that it was so damn easy to sabotage their effort that I didn't really care if my team did it because that's what happens when you invite the whole internet to your party. Now I guess they managed to clear us of that. :) Unfortunately, when this Adam guy wrecked their program, he wrote them a letter which simultaneously implied that he was our team and the group "Anonymous" at the same time. Some people at DARPA seemed to believe that claim (even though I used my real name to sign up) and so DARPA was very hesitant to announce that we had won the competition. It was a fun competition though.

I worked with Keith at the time. Wasn't it true that you guys figured out that the yellow paper had a regular blue dot pattern watermark, and that you were able to use them to help solve that one? If I recall correctly, DARPA didn't know about that.

Yes, one of the keys to reconstructing the harder puzzles was that we found a yellow dot pattern on the paper. At the time, I was reading HN and I saw a headline that said something about the government tracking photocopies using little yellow dots. DARPA had photocopied all the docs before shredding them, so they all had a very high-res repeating dot pattern on them. We made our program snap the pieces to the dot pattern. With that in place, the puzzles came together relatively quickly. Some people at DARPA knew about the concept of little yellow dots, but didn't think much of it. I don't blame them. The shredded pieces were so small that it would be hard to imagine any secret pattern helping with the reconstruction. But I guess it helped enough. :)

Interesting response from Harper Reed, who led Obama's 2012 online campaign.

https://medium.com/@harper/crowdsourcing-isnt-broken-5681da9...

WOW Thanks. There are some very good advice in there. Do you know any more articles in this area? Or how to manage a community in general!?

The Harper Reed article reminded me of this article [1] from a while back about how StackOverflow handles similar issues. Pretty similar, but maybe you will get something additional from it.

[1] - http://blog.codinghorror.com/suspension-ban-or-hellban/

Fantastic. It's really unfortunate that the article paints a simplistic Game-theory failure for all community efforts when there are already solutions out there that do work.

If you think about it, corporations are essentially crowd sourcing machines with sophisticated recruiting, hierarchies, regulations and rewards. One could say that corporations are on the far right of the spectrum while the contest itself with minimal rules, regulations and vetting for participants is at the far left.

Which is more efficient Corporations or crowd sourcing? Perhaps the sweet-spot is somewhere in the middle. Just add some rules and structure for vetting the players or distinguishing the saboteurs and maybe it might work!

"Game theorists have found that systems where individuals can build up a good reputation, are (probably) not as prone to devastating attacks from within.

But wily humans are good at finding their way around even the most secure digital systems."

/me eyes HN's karma system suspiciously...

So a crowdsourcing platform that knew it had security holes, but decided to risk it anyway... got trolled.

Interesting, and a good story, but it hardly invalidates the concept of crowdsourcing.

Did you read the whole article?

The studies cited point to this being the case in a lot more than just one instance.

Even 'successful' crowdsourced efforts like Wikipedia are scarcely without their problems.

The article does highlight that there is an interesting game-theoretic issue here. Given a crowd of N people, with such-and-such dynamics, is the damage that 1 person can do bounded or unbounded? Is it better to control the number of bad actors, or to restrict the actions any one actor may take?

I'm un-aware of results in this area, but as online communities become more important, and bots become more sophisticated, the stakes will raise. Having some community models, analysis tools, or heuristics, would be very interesting.

Does anyone know why a high end shredder would shred the paper in such regular pieces, and of such size that there are still letters visible on each piece?

I am speculating here, but I would imagine that the smaller the pieces, the more likely the machine is to jam and the fewer sheets it can shred at a time. Regular pieces are used because of how the shredding mechanism works.

I think they did what they did because until this competition, it was always sufficient. If you really don't want paper to be recovered, burn it.

Here's the real-world case involving the East German Stasi that I remember first hearing about years before this article:

http://www.npr.org/2012/10/08/162369606/piecing-together-the...

Why such regular pieces?

"Petter says the shredding machines were under such strain they eventually burned out... Panic-stricken, the Stasi's agents resorted to ripping up files with their bare hands."

Or as another example, quoting from http://en.wikipedia.org/wiki/Paper_shredder :

> After the Iranian Revolution and the takeover of the U.S. embassy in Tehran in 1979, Iranians enlisted local carpet weavers who reconstructed the pieces by hand. The recovered documents would be later released by the Iranian regime in a series of books called "Documents from the US espionage Den".[8] The US government subsequently improved its shredding techniques by adding pulverizing, pulping, and chemical decomposition protocols.

What about burning the shreds?

It's a good addition to any suite of destruction, but creating any kind of fire is a higher amount of risk for collateral damage than the other methods. Plus, unless the combustion is thorough, you may still be able to discern information off the paper.

Yes, I meant burn after shredding. If you have time, composting (after shredding) should work well.

At the very least, you can mix your shreds and soak them in water.

That's pulping.

It looks like the US people locked themselves in the vault in order to destroy the documents. After they were done, they left. There might not be the oxygen in the vault to burn all of the documents. But that's conjectural.

According to http://en.wikipedia.org/Burn_bag , burning isn't so common due to environmental concerns and the high value of recycled paper, though some places have burn bags with built-in thermite charges, in case of emergency.

I wonder if self-organisation can help?

If people can reenforce matches, and people who match alike are probable to see the same board, then the trolls will most match the other trolls and be invisible to the serious competitors?

I've played with collaborative filtering for sites like HN and I wonder if the same approach might sort trolls from the serious in these kinds of things?

http://williamedwardscoder.tumblr.com/post/15581427232/self-... <- my blog on collaborative filtering and trolls

A bit like pull-requests in git?

Interesting article but a completely hyperbolic title and paragraph headers. "The end of crowdsourcing?"

The best way to complain about a title on HN is to suggest a better title. If you're right, we'll probably use it.

Edit: We've changed the HN title to a phrase taken (mostly) from the article's subtitle. If anyone suggests a better one, we can change it again.

The article itself is clearly good and on-topic, so let's not fret about the title too much.

Alternate title: "Crowdsourcing Vulnerabilities: DARPA Shredder Challenge Post-Mortem."

I find it interesting that you do this. I think that the title is a part of the original work. After all the author did put energy into selecting the title, just as he selected the words to use for the content of the article. I suppose the person that submits the link might not be the author, which is probably what happens the most. I'm still torn.

Respect for the original content is a high value at Hacker News, but so is intellectual substance. These two values sometimes conflict, especially when it comes to titles. When they do, intellectual substance wins.

It has to. Otherwise threads would get sucked into the black hole of arguing about titles. In that world we would probably be wading through angry denunciations and counter-denunciations of this one, for example, instead of pure HN gold like https://news.ycombinator.com/item?id=9025587.

Reply on news.ycombinator.com