Huzzah for exams.

[chess]

As I confidently predicted, today's paper went fine. However, I still feel rather like I cheated somehow, because it only went fine because I revised the Exact Correct Thing To Revise this morning, which I feel is almost certainly the effect of the people who are praying for me... credit should also go to tienelle who explained my Data Structures and Algorithms course to me far better than the notes or the textbook managed (or, indeed, my supervisor, who suffered from English Not First Language problems).

Question One, I thought about attempting, and then when I actually got around to doing it I thought 'No, actually, I don't know what a D flip-flop *is*', and as most of the question was about the specifics of D flip-flops, that rather ruled out answering it.

Question Two is the question which I revised quite thoroughly this morning for no good reason. It was about Huffman codes and arithmetic coding. I would have done even better on it had I remembered how to deal with log2 (2^foo + (2^2)^bar) gracefully, but hopefully only by two marks at most. I did this question second, and dealt with the bits on arithmetic encoding last. I expect 15-19 marks on this one, depending how many marks they were offering for 'assumptions for optimally compact code' (who with the what now?) and algebra.

Question Three is the question on which I produced piles of steaming bullshit, but after I had done all the questions I could actually answer well, it was what I was left with that I could give some kind of waffly answer to all the parts of. So I did. I'm not expecting more than about 10 marks because my whole answer can be summed up as 'waffle waffle waffle'.

Question Four I thought 'yay, trivial Prolog stuff' and won at it. If I don't get 20 marks, it will be because the examiners got lost somewhere in my reams of prolog evaluation and/or they wanted more than one line for five marks for part (d) (but that would have been a huge and terrible evaluation and I'd already spotted the pattern, damn it).

Question Five I was a little wary of because it was stated in rather an informal way and I wasn't entirely sure what some of the little bits were getting at, but the only bit I was very unsure of was (b)(i) which was only two marks out of the question. 12-20 marks depending on how good my guesses at what they wanted were.

Question Six was another 'yay, here's the easy question', although I couldn't remember whether x/(minus infinity) was zero or not a number. 15-19 marks depending on how lousy my solution to the last part actually was. But mostly it was all about Floating Point, and that's nice and easy and fun.

I didn't do Seven, Eight or Nine. Seven I didn't do because it had the words 'rigorous proof' in, and also because I could not for the life of me remember any undecidable problems which were not the Halting Problem. Eight I didn't do because it had the irritating 'calculate the maximum resolution needed by a foo, stating your assumptions' and I couldn't even remember a reasonable assumption for Cones In Eye, let alone how to put it all together, although if I'd been desperate for a question I could have got some marks out of it and the other parts. Nine I didn't do because even though it was basically the cryptographic signing thing which ToothyWiki does, I was less confident of the details of that than I was of the AI question stuff.

I should probably do some last-minute revision for the next few papers, since today's last-minute revision worked so well, but first - Spod!