(no subject)
First year of uni is over. I'm just sticking around for the end of term party in a few hours, and they're kicking me out of halls tomorrow morning.
It's been a good year. Learned lots; had new experiences; met interesting people.
Exam results:
(Most of this was written before getting them.)
* Relativity: 96%
For whatever reason, I couldn't sleep the night before this exam. I sat down, picked up my pen, and found that my handwriting was slow and clumsy. Despite that, I found the paper quite easy and finished with ten minutes to spare. Most people seem to have found it difficult: I spoke to people who got 26% and 50%, and the second-years I spoke to got low marks when they did it as well.
* Linear Algebra: 77.39%
The exam paper was perfectly fair, and I'd done enough revision. I just wasn't thinking very well during the exam.
* Analysis: 81.91% overall; 85% on the exam
Also perfectly fair and enough revision. I didn't see question seven, which was probably easier than one of the questions I did.
I'm not really sure how my overall mark was lower than my exam mark, because I got 95% on the january exam. I suppose it's the assignments, but I wouldn't have expected them to be so significant. It's also possible that I got the exam and overall percentages the wrong way around.
* Probability A/B: 79.75%/62%
Getting a 2.1 in B was the most surprising aspect of these results. I wasn't confident I'd passed that. I didn't do enough revision, and everyone (including me) found part A quite difficult. Part B seemed fair, but there were large parts that I either failed at doing or didn't know how, and I used some of the time for it finishing part A.
* Number Theory: 80%
I could have done with better knowing some minor proofs. Also I couldn't decide between two questions, and ended up trying part of both before choosing.
* Differential Equations: 76.20%
This was fine, except one point where I was stupid and forgot that I'd calculated the value of lambda in a previous part of the question. That made it impossible to calculate the other stuff I needed.
* Geometry and Motion: 72.40%
Probably the scariest exam for most people: there's a lot of methods and formulae to remember, a lot of complicated algebra that you won't know if you're doing wrong, and you're expected to do it pretty fast. I found it fine, though there were a few things I should have revised better. Expected a higher mark than this.
* Intro to Geometry: 76%
Apart from the proof of Ceva's theorem, which I should have revised, this was really easy. That was worth either 8 or 12%, but I still expected higher.
* Discrete Maths: 96%
This was the first year of the module, so there were no past papers to revise from. Just a "sample paper", which turned out to have many of the same questions as the real exam, but slightly harder. I finished, then realised I'd made a mistake at one point when simplifying a predicate and ran out of time before I fixed it. It looks like that may have been the only place I lost marks.
These modules didn't have exams (or at least not in term 3):
* Foundations: 79%
* Experimental Maths: 85.40%
* Maths by Computer: 85%
* Vectors and Matrices: 94.51%
My unweighted average is 81.28%. Applying the Seymore formula (which takes into account that I took more modules than necessary), I get 85.16%. This puts me 15th in the year of 340.
tl;dr: WF remains pretty good at maths.
---
My mum wants me to phone both sets of grandparents, and David Race, to tell them my results. "Oh hi, I thought you might like to know how well I did in my exams." I suppose it's probably true, but that doesn't make it feel any less arrogant. And they'll probably want to congratulate me, but congratulations make me feel uncomfortable if they last longer than "well done", "thanks". It feels like the other person is more impressed than I am, and I'm not quite sure how to respond.
---
Immediately after exams finished, there was a 48 hour game design competition. The theme was procedural generation; I was teamed with Phil, and we did an idea I'd been throwing around my head of a competitive Life simulator.
My original idea was that it would be an RTS: each player attempts to kill/vitalise cells in such a way as to kill a specific cell of their opponent's which would be in a locally stable alive state (probably the middle cell of a 3x1 oscillator). Players' cells would be distinguished by colour; all live cells would be treated the same for the purposes of life and death, with conflicts in colours being resolved to black (neutral). You would gain resources by having lots of live cells, and use resources to make changes to the game board; you would also only be able to change cells that you had enough influence over, where influence would be some ill-defined function of the colours of nearby cells.
The problem with this is that it would have required either networking or AI. Neither of us had done networking before; Phil's third-year project was on AI, but he didn't think that what he learned would have generalised. And he wasn't sure it would really work as an RTS anyway. (I'm not convinced either, but I still want to try.) So we made it a puzzle game: you're given an initial configuration, and have to bring about certain states making as few changes to the game board as possible.
At ~24 hours things were looking pretty great: we had a Life simulator which worked, was fast enough, and had multiple colours; and we had a GUI which displayed it and allowed most of the desired interactive features. All we needed to do was put the puzzles in. But then Phil went to bed, and I looked at his GUI code and had no idea what to do. He'd made a finite state machine, which I don't think is a very good model for a video game with multiple levels: you need a state for each level, which means a lot of duplicated code and little dynamism (each state was a separate class). Or you can merge the "select level" and "playing level" states, which is stupid. So I decided I'd wait until he got back to see what he was planning; that would leave plenty of time to finish, and there was still lots else that I could work on.
That night I was unshockingly unproductive; I came up with some level ideas, improved the interface a little, and started looking at audio generation. (With the RTS idea, I thought it would be cool if the music playing would be a function of the living cells, with each player represented by a different set of instruments. So if you were getting overrun, you'd hear a lot of drums instead of woodwind, say, and that could be pretty cool. But maybe less cool if you were playing as drums and the "oh shit I'm about to die" music was heavy-duty piccolo stylings.) But tiredness made me slow and stupid; I really should have got some rest, but Tom had been on MSN telling me that I was being an idiot for not sleeping. So I had to stay up to spite him. Besides, I had energy drinks.
At about 5AM the last person in the room other than myself also left to get sleep. Which meant I really couldn't sleep now, because there were so many laptops in the room and no way of locking it. I also couldn't leave the room to go to the toilet. And that meant I didn't want to drink too much. (I think I did drop off for an hour or two anyway. But nothing went missing, and there's no other way I would have been functioning by the end of the competition.) I was just slacking off at this point; I expected Phil to be back soon anyway. (Hah.)
At 11am Sarah returned, which meant I could use the toilet and shower. By the time I got back, Phil had also returned. But now we only had six hours left.
It turned out Phil had no plan for the level selection. He ended up breaking out of the FSM to let the level state query the level selection state to find out which level to play. Which was simple enough, but in the end I only managed to add two of the five levels that I'd thought of, plus the sandbox. And didn't really test either of them; I have no idea how reasonable they are. And this is despite that we finished three hours late. I'm not really sure what happened towards the end...
tl;dr: WF needs more practice at making games.
---
Airsoc won the award for best new society. This was not surprising to me, because Jake's girlfriend was on the judging committee or something. I decided it would be polite to not tell anyone else. Some of them were impressed that I'd managed to keep it a secret, but it wasn't exactly difficult. I didn't care too much one way or the other if we'd won, although it does mean we get something like £100 extra in our budget for next year.
---
At the start of the year, everyone in Go society was giving me nine stones and lots of komi except Rob, who was only slightly stronger than me, and Jonathan, who was slightly weaker. Now some new people have joined, and I'm giving them most of them nine stones and some komi (while receiving even more from Bucko). I don't know how to play at that level; I can make two attacks before there's basically no room for me to do anything. But I seem to win about 50% of the time anyway, so it must be accurate.
It's been a good year. Learned lots; had new experiences; met interesting people.
Exam results:
(Most of this was written before getting them.)
* Relativity: 96%
For whatever reason, I couldn't sleep the night before this exam. I sat down, picked up my pen, and found that my handwriting was slow and clumsy. Despite that, I found the paper quite easy and finished with ten minutes to spare. Most people seem to have found it difficult: I spoke to people who got 26% and 50%, and the second-years I spoke to got low marks when they did it as well.
* Linear Algebra: 77.39%
The exam paper was perfectly fair, and I'd done enough revision. I just wasn't thinking very well during the exam.
* Analysis: 81.91% overall; 85% on the exam
Also perfectly fair and enough revision. I didn't see question seven, which was probably easier than one of the questions I did.
I'm not really sure how my overall mark was lower than my exam mark, because I got 95% on the january exam. I suppose it's the assignments, but I wouldn't have expected them to be so significant. It's also possible that I got the exam and overall percentages the wrong way around.
* Probability A/B: 79.75%/62%
Getting a 2.1 in B was the most surprising aspect of these results. I wasn't confident I'd passed that. I didn't do enough revision, and everyone (including me) found part A quite difficult. Part B seemed fair, but there were large parts that I either failed at doing or didn't know how, and I used some of the time for it finishing part A.
* Number Theory: 80%
I could have done with better knowing some minor proofs. Also I couldn't decide between two questions, and ended up trying part of both before choosing.
* Differential Equations: 76.20%
This was fine, except one point where I was stupid and forgot that I'd calculated the value of lambda in a previous part of the question. That made it impossible to calculate the other stuff I needed.
* Geometry and Motion: 72.40%
Probably the scariest exam for most people: there's a lot of methods and formulae to remember, a lot of complicated algebra that you won't know if you're doing wrong, and you're expected to do it pretty fast. I found it fine, though there were a few things I should have revised better. Expected a higher mark than this.
* Intro to Geometry: 76%
Apart from the proof of Ceva's theorem, which I should have revised, this was really easy. That was worth either 8 or 12%, but I still expected higher.
* Discrete Maths: 96%
This was the first year of the module, so there were no past papers to revise from. Just a "sample paper", which turned out to have many of the same questions as the real exam, but slightly harder. I finished, then realised I'd made a mistake at one point when simplifying a predicate and ran out of time before I fixed it. It looks like that may have been the only place I lost marks.
These modules didn't have exams (or at least not in term 3):
* Foundations: 79%
* Experimental Maths: 85.40%
* Maths by Computer: 85%
* Vectors and Matrices: 94.51%
My unweighted average is 81.28%. Applying the Seymore formula (which takes into account that I took more modules than necessary), I get 85.16%. This puts me 15th in the year of 340.
tl;dr: WF remains pretty good at maths.
---
My mum wants me to phone both sets of grandparents, and David Race, to tell them my results. "Oh hi, I thought you might like to know how well I did in my exams." I suppose it's probably true, but that doesn't make it feel any less arrogant. And they'll probably want to congratulate me, but congratulations make me feel uncomfortable if they last longer than "well done", "thanks". It feels like the other person is more impressed than I am, and I'm not quite sure how to respond.
---
Immediately after exams finished, there was a 48 hour game design competition. The theme was procedural generation; I was teamed with Phil, and we did an idea I'd been throwing around my head of a competitive Life simulator.
My original idea was that it would be an RTS: each player attempts to kill/vitalise cells in such a way as to kill a specific cell of their opponent's which would be in a locally stable alive state (probably the middle cell of a 3x1 oscillator). Players' cells would be distinguished by colour; all live cells would be treated the same for the purposes of life and death, with conflicts in colours being resolved to black (neutral). You would gain resources by having lots of live cells, and use resources to make changes to the game board; you would also only be able to change cells that you had enough influence over, where influence would be some ill-defined function of the colours of nearby cells.
The problem with this is that it would have required either networking or AI. Neither of us had done networking before; Phil's third-year project was on AI, but he didn't think that what he learned would have generalised. And he wasn't sure it would really work as an RTS anyway. (I'm not convinced either, but I still want to try.) So we made it a puzzle game: you're given an initial configuration, and have to bring about certain states making as few changes to the game board as possible.
At ~24 hours things were looking pretty great: we had a Life simulator which worked, was fast enough, and had multiple colours; and we had a GUI which displayed it and allowed most of the desired interactive features. All we needed to do was put the puzzles in. But then Phil went to bed, and I looked at his GUI code and had no idea what to do. He'd made a finite state machine, which I don't think is a very good model for a video game with multiple levels: you need a state for each level, which means a lot of duplicated code and little dynamism (each state was a separate class). Or you can merge the "select level" and "playing level" states, which is stupid. So I decided I'd wait until he got back to see what he was planning; that would leave plenty of time to finish, and there was still lots else that I could work on.
That night I was unshockingly unproductive; I came up with some level ideas, improved the interface a little, and started looking at audio generation. (With the RTS idea, I thought it would be cool if the music playing would be a function of the living cells, with each player represented by a different set of instruments. So if you were getting overrun, you'd hear a lot of drums instead of woodwind, say, and that could be pretty cool. But maybe less cool if you were playing as drums and the "oh shit I'm about to die" music was heavy-duty piccolo stylings.) But tiredness made me slow and stupid; I really should have got some rest, but Tom had been on MSN telling me that I was being an idiot for not sleeping. So I had to stay up to spite him. Besides, I had energy drinks.
At about 5AM the last person in the room other than myself also left to get sleep. Which meant I really couldn't sleep now, because there were so many laptops in the room and no way of locking it. I also couldn't leave the room to go to the toilet. And that meant I didn't want to drink too much. (I think I did drop off for an hour or two anyway. But nothing went missing, and there's no other way I would have been functioning by the end of the competition.) I was just slacking off at this point; I expected Phil to be back soon anyway. (Hah.)
At 11am Sarah returned, which meant I could use the toilet and shower. By the time I got back, Phil had also returned. But now we only had six hours left.
It turned out Phil had no plan for the level selection. He ended up breaking out of the FSM to let the level state query the level selection state to find out which level to play. Which was simple enough, but in the end I only managed to add two of the five levels that I'd thought of, plus the sandbox. And didn't really test either of them; I have no idea how reasonable they are. And this is despite that we finished three hours late. I'm not really sure what happened towards the end...
tl;dr: WF needs more practice at making games.
---
Airsoc won the award for best new society. This was not surprising to me, because Jake's girlfriend was on the judging committee or something. I decided it would be polite to not tell anyone else. Some of them were impressed that I'd managed to keep it a secret, but it wasn't exactly difficult. I didn't care too much one way or the other if we'd won, although it does mean we get something like £100 extra in our budget for next year.
---
At the start of the year, everyone in Go society was giving me nine stones and lots of komi except Rob, who was only slightly stronger than me, and Jonathan, who was slightly weaker. Now some new people have joined, and I'm giving them most of them nine stones and some komi (while receiving even more from Bucko). I don't know how to play at that level; I can make two attacks before there's basically no room for me to do anything. But I seem to win about 50% of the time anyway, so it must be accurate.