Oftentimes these days, I've been sitting studying and wanted to let out a primal scream. If there are people around, I usually manage to keep it in, but SAC has pushed me over the edge.
Rijndael. Rijndael. Rijndael. Rijndael. Rijndael. Rijndael. Rijndael.
I shall spare you a detailed description of the paper: Chris has beaten me to it. What it suffices to say is that it was a difficult exam, no question, and far more difficult than the 2004 paper (insert usual caveat about hindsight and sitting with a sample answer), which, given the bipolar nature of these things, is perhaps to be expected.
I was initially surprised that question 1 (modular arithmetic, knapsacks, secret key safekeeping) proved so unpopular: to me, it seemed the most straightforward, though I don't and probably never will know how to construct a superincreasing knapsack cipher with a keyspace entropy of 512. Question 2 looked unpleasant to me because of all the marks for elections (not my area of expertise), and I wasn't overly keen to talk about the Oakley improvements to Diffe-Hellman, but I hedged my bets, and went for question 3 where the marks seemed a little easier to achieve.
The problem with this exam was the number of fiddly details that had to be remembered. I commented on the way in today that Ron could contrive an exam that would utterly stump me (a big question on EMV, a big question on elections, something intricate about Rijndael, a vague question on PGP and a detailed question on Oakley cookies ought to have done it). Fortunately, he didn't have the audacity to do that :P, but it was still one of the hardest SAC exams I've seen.
It should be noted that today marked the start of the infamous week of death for the unfortunate people who chose SAC4, SCS4, NCT4, DBIT4 and SEP4. Good luck to everybody doing SCS tomorrow!*
PS. It's too bad nobody doing SCS is blogging the exams, since they could put details of the questions up tonight!