Comments on: "Where is Waldo?", or "Security by Origami" http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/ Algorithms, for the heck of it Tue, 21 Jun 2011 21:07:08 +0000 hourly 1 http://wordpress.org/?v=3.1.3 By: Algorithm Blogs » Origami crease-patterns are NP-complete http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/#comment-15 Algorithm Blogs » Origami crease-patterns are NP-complete Tue, 31 Jul 2007 21:59:04 +0000 http://www.algorithm.co.il/blogs/index.php/math/where-is-waldo-or-security-by-origami/#comment-15 [...] any case, this shows that my suggestion of using origami as a hash function is valid! Really cool stuff… I wonder if it could actually be useful [...] [...] any case, this shows that my suggestion of using origami as a hash function is valid! Really cool stuff… I wonder if it could actually be useful [...]

]]> By: lorg http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/#comment-14 lorg Mon, 04 Jun 2007 21:22:00 +0000 http://www.algorithm.co.il/blogs/index.php/math/where-is-waldo-or-security-by-origami/#comment-14 I talked again with the guy who gave me this riddle - Nadav Sherman. His solution to the problem seemed to be very similar to mine - use a cover with a hole right over Waldo's face. According to him this solution is problematic - I can cheat and plant another picture of Waldo underneath the cover. To make it work, he suggested adding another cover above the holed cover, this time complete. Now, in the manner of other zero-knowledge proof, the guy who didn't find Waldo can choose - either remove the upper cover - thus showing a part of Waldo, proving that I know where he is, or remove both of them - this showing the complete picture - proving I didn't cheat. It is important to note that this solution is a solution to the more interesting problem - of proving I found Waldo even if the other guy didn't find him himself. I talked again with the guy who gave me this riddle – Nadav Sherman. His solution to the problem seemed to be very similar to mine – use a cover with a hole right over Waldo’s face. According to him this solution is problematic – I can cheat and plant another picture of Waldo underneath the cover. To make it work, he suggested adding another cover above the holed cover, this time complete. Now, in the manner of other zero-knowledge proof, the guy who didn’t find Waldo can choose – either remove the upper cover – thus showing a part of Waldo, proving that I know where he is, or remove both of them – this showing the complete picture – proving I didn’t cheat.
It is important to note that this solution is a solution to the more interesting problem – of proving I found Waldo even if the other guy didn’t find him himself.

]]> By: Sylvan von Stuppe http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/#comment-13 Sylvan von Stuppe Sat, 02 Jun 2007 15:54:35 +0000 http://www.algorithm.co.il/blogs/index.php/math/where-is-waldo-or-security-by-origami/#comment-13 @lorg I had also considered that if you're rounding to the nearest centimeter (or some measurement), it might be easier to brute-force the solution. I picked the nearest centimeter because the two books would have to have been printed and bound exactly the same in order to get the measurements of a single point very exact. As long as it's okay for the proof to come after both parties have found Waldo, it's okay for the measurements to be exact so long as the first person can explain the coordinates used to gen the hash - the hash is only used as a way to keep the coordinate secret while the slower party finishes. @lorg

I had also considered that if you’re rounding to the nearest centimeter (or some measurement), it might be easier to brute-force the solution. I picked the nearest centimeter because the two books would have to have been printed and bound exactly the same in order to get the measurements of a single point very exact. As long as it’s okay for the proof to come after both parties have found Waldo, it’s okay for the measurements to be exact so long as the first person can explain the coordinates used to gen the hash – the hash is only used as a way to keep the coordinate secret while the slower party finishes.

]]> By: lorg http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/#comment-12 lorg Sat, 02 Jun 2007 12:43:33 +0000 http://www.algorithm.co.il/blogs/index.php/math/where-is-waldo-or-security-by-origami/#comment-12 This solution was the first solution I thought about. It works, but we (me and the guy who gave me the riddle, henceforth 'the riddler') discounted it as 'too obvious' :). Also - The 'riddler' also suggested that it might be faster for the guy who didn't find Waldo to brute-force the (discrete) coordinates then try and find Waldo. A possible improvement might be to give the hash of coordinates and some redundant information. For example - hashing the coordinates to x digits after the decimal point. There are also some possible proofs before the the guy finds Walso himself. For example, we take a black board with a small hole in it, and align the board such that a part of Waldo is underneath the hole. It still relies on you being able to give a part of the picture of Waldo, without giving any hints, just like the solution of cropping out everything besides Waldo, that appeared in <a href="http://blogs.securiteam.com/index.php/archives/917" rel="nofollow">Securiteam blogs</a>. This solution was the first solution I thought about. It works, but we (me and the guy who gave me the riddle, henceforth ‘the riddler’) discounted it as ‘too obvious’ :). Also – The ‘riddler’ also suggested that it might be faster for the guy who didn’t find Waldo to brute-force the (discrete) coordinates then try and find Waldo. A possible improvement might be to give the hash of coordinates and some redundant information. For example – hashing the coordinates to x digits after the decimal point.

There are also some possible proofs before the the guy finds Walso himself. For example, we take a black board with a small hole in it, and align the board such that a part of Waldo is underneath the hole. It still relies on you being able to give a part of the picture of Waldo, without giving any hints, just like the solution of cropping out everything besides Waldo, that appeared in Securiteam blogs.

]]> By: Sylvan von Stuppe http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/#comment-11 Sylvan von Stuppe Fri, 01 Jun 2007 23:00:56 +0000 http://www.algorithm.co.il/blogs/index.php/math/where-is-waldo-or-security-by-origami/#comment-11 This solution requires technology and you and your friend are still on the "honor system" because the proofs could cross paths in the air. Each friend picks a word or passphrase. When you've found Waldo, you use the colors of the five closest people, or the x,y coordinates of Waldo's hat or something like that - something that explains where he is. When you've found Waldo, you append your public key (the agreed upon word) to the coordinates or other distinguishing characteristic, then calculate a hash for that (MD5, SHA-1, SHA-256 - depends on how important the game is). Give the hash to your friend. When your friend finds Waldo, they use the agreed upon characteristics (location works better) and your password, then calculates the hash in the same way. If they match, then you solved it first. The biggest downside to this is that your friend MUST find Waldo in order to prove that you found him first. Suppose we agree on coordinates of the tip of Waldo's hat as the objective. And my public word is SYLVAN. When I find Waldo, he's at X 8.75cm (round to the nearest cm - 9) and Y 4.1cm (rounded to 4cm). So I calculate: MD5('SYLVAN:9:4') and pass that to my friend. When my friend finds Waldo (at the same coordinates), he has enough information to calculate the same hash - neither of us can calculate it without finding Waldo. This solution requires technology and you and your friend are still on the “honor system” because the proofs could cross paths in the air.

Each friend picks a word or passphrase. When you’ve found Waldo, you use the colors of the five closest people, or the x,y coordinates of Waldo’s hat or something like that – something that explains where he is. When you’ve found Waldo, you append your public key (the agreed upon word) to the coordinates or other distinguishing characteristic, then calculate a hash for that (MD5, SHA-1, SHA-256 – depends on how important the game is). Give the hash to your friend. When your friend finds Waldo, they use the agreed upon characteristics (location works better) and your password, then calculates the hash in the same way. If they match, then you solved it first.

The biggest downside to this is that your friend MUST find Waldo in order to prove that you found him first.

Suppose we agree on coordinates of the tip of Waldo’s hat as the objective. And my public word is SYLVAN.

When I find Waldo, he’s at X 8.75cm (round to the nearest cm – 9) and Y 4.1cm (rounded to 4cm). So I calculate:

MD5(‘SYLVAN:9:4′) and pass that to my friend. When my friend finds Waldo (at the same coordinates), he has enough information to calculate the same hash – neither of us can calculate it without finding Waldo.

]]> By: SecuriTeam Blogs » “Where is Waldo?”, or “Security by Origami” http://www.algorithm.co.il/blogs/math/where-is-waldo-or-security-by-origami/#comment-10 SecuriTeam Blogs » “Where is Waldo?”, or “Security by Origami” Fri, 01 Jun 2007 15:43:41 +0000 http://www.algorithm.co.il/blogs/index.php/math/where-is-waldo-or-security-by-origami/#comment-10 [...] is an interesting excersize in security: A friend of mine gave me a riddle this morning regarding “Where’s Waldo?”. The riddle is as [...] [...] is an interesting excersize in security: A friend of mine gave me a riddle this morning regarding “Where’s Waldo?”. The riddle is as [...]

]]>