There’s a fine line between game freak and computer geek, and I think I may be standing on it. I first discovered Visual Basic on a web forum based around an online game called Neopets, and taught myself using online tutorials. I made programs which utilised the winsock control to make playing the game much easier, but it quickly led to a race to be the best, as it was very competitive amongst the programmers to have the most efficient programs. I then started programming for myself and for friends (and even once for a teacher) for other uses, ranging from seeing if a website is available (e.g. to get tickets to a Led Zeppelin concert) to calculating the lowest common multiple of two numbers (thank you, Euclid!). More recently, I started looking at PHP and CSS in addition to HTML to increase my knowledge of web design, and I started a website to test out my coding, and a blog on events in my life.

In the future, I hope to be a creative software engineer, but even if I don’t, the ability to program efficiently will be extremely useful, as almost all fields of work can benefit from some form of programming. This is a fact that was proven to me during my various work experiences over the past few summers: at Royal Bank of Canada, I saw they wrote specific programs when they needed them for tasks like searching their enormous database of clients and complex products; at Citibank, I saw how they could have used a programmer on their team in order to make the complex data and product structures in spreadsheets much faster to access and manipulate. This past summer in India, I visited my uncle’s car parts factory, which I discovered still uses Visual Basic for all of their systems. However, they had a programmer on site to convert it to the new .NET framework. I spent time with him, and I was impressed with the complexity of the VB programming in use. I also volunteered at a rural development NGO called Chirag, where I lived with villagers, helped in the local primary school and planted grass as fodder for the livestock. In order to improve my programming knowledge, during the long journeys to and within India I read some of Donald Knuth’s famous ‘The Art of Computer Programming’ which I found both extremely interesting and at times confusing. It did however reinforce my thirst to learn more. I am doing computing as an A level to help me understand better how to apply the subject to real-life situations (e.g. I programmed a simulation of a petrol pump system for my AS coursework and am writing a program to gather and search Tweets from the social networking website Twitter).

I am president of the school computer, magic and bridge societies, and captain of the bridge team. In the computer society, we discuss technological advancements and test out various projects on the computers we assembled specifically for that purpose. Some of us entered a competition to make a 2 minute video about computer science and won a trip to the Google headquarters in London. In the magic Society, we learn tricks and organise shows at local primary schools. In bridge club we strive to improve our bridge skills and occasionally play against nearby schools. I sing in the Concert Choir and am also on the Charities Committee. My main sporting interests are fencing, cricket and aikido, and I am currently in the middle of completing the Duke of Edinburgh gold award. In order to get a taste of university Computer Science, I attended Computer Science taster courses at Imperial College and Royal Holloway, where I benefited from very interesting lectures on topics such as artificial intelligence, game theory and the travelling salesman problem. I also had the privilege of visiting Princeton University where I met Brian Kernighan, famous co-author of ‘The C Programming Language’. These lectures and meeting this great professor further strengthened my conviction that without a doubt Computer Science is what I want to study.

I woke up at 9. Got out of bed at 10. Had breakfast. Then spent two hours deciding what to wear (I eventually ended up in a suit). I was worried that I’d be the only person there in a suit and look stupid, but fortunately I wasn’t.

The day went pretty well; at first we just hung around in a waiting room but then we waited in a different room which had food in it (this is a big deal). Dr. Bradley came along after a while and we went off to a talk by him about Imperial in general and a bit about Computer Science at Imperial. Then the JMC people separated from the straight CS people and we went to a talk about Computer Science at Imperial (which I had seen before).  We were then split into groups based on the timings of our interviews; I was in group 6. We were shown various projects that students had done and taken on a short tour of the campus (”The bars are here, here and here. There is alcohol in them.”).

I was the second last person to be taken to their interview, so the wait was pretty stressful.

The interview actually didn’t go that badly… We started by talking a little about The Art of Computer Programming by Donald Knuth which was pretty fun. I was then asked to work out something about the sum of all fibbonaci numbers up to k, and then prove it by induction (this made me lulz inside, because last night I tried re-reading some of Knuth but got bored after the first section (which happened to be on proof by induction)). It went okay but I was nervous and so kept making stupid mistakes, and then correcting them.

We talked a little about my PS but only a bit about my visit to Princeton to visit Kernighan and about how being a “computer geek” would be beneficial but it wouldn’t be good to be *too* into it.  I’m glad she didn’t point out the typo I made (unlike RAB who took great pleasure in informing me of it).

I was then asked a programming question. It was about an array full of As and Bs. It was okay. I suggested the least efficient method of switching the As and Bs (bubble sort) as a joke but then went on with the real thing. It was all good.

All in all it was good. We’ll find out in a month.

“Even though we have fast internet, it’s wasted speed since you can’t torrent… that was a lie; as Computer Science students, you can torrent all you like.”

I have my Imperial interview on Wednesday. I think I’ll take some of Douglas Adam’s very good advice. DON’T PANIC.

1. Would it be useful to be able to search all of the Tweets that anyone has ever made (including private feeds of people you are following)?

2. Would it be useful to be able to search Tweets by date posted?

3. Would it be useful to be able to download all Tweets of anyone who you can view the feed of?

4. Assuming the answer to 3. is a yes, would you prefer the downloaded file to be one that is easily understandable in a spreadsheet software package, or that can be read by the program that is being made and searched from within that?

5. Would you expect a program such as this to run quickly or would you allow for it to be quite slow (given that the current programs which do similar things can take more than 10 minutes to run)?

6. Would a Twitter login that is integrated into the program be better than a browser window opening within the program (which would show the user that the information is (more likely to be) secure?

7. Any suggestions?

Twitter Archiver

A program which downloads tweets from an inputted Twitter user. It gets all Tweets ever made by the person and saves it to a file which can be viewed and modified in any spreadsheet software, which would eliminate the old problem of Tweets not being searchable, and would let you do graphs to show the history of various people. It would be able to collect tweets from all people you are following and collate them all in a single file. This would get around the problem that is normally faced which is that you can only look at the last 40 pages of tweets. It would not use the Twitter API so that it is not restricted by the API limits (70 queries per hour or something similar), although this would make it harder to code and slightly slower, the benefits outweigh the small loss in speed. I would probably use Winsock or INET to get the information. If a user’s updates are protected then it will ask for you to log in to Twitter, and if then you aren’t following them, then it would come up with an error saying so. It may also be able to archive all direct messages (though this function is fairly useless). Will be able to save login info (locally, of course ;)) so that you don’t have to type it in every time you use the program. Potential users are pretty much any Twitter users and I could probably interview @farhan, @dolphonia or if needs be @preoccupations if he isn’t too busy, and I could do a survey over Twitter. Hurrah for social networking! This is actually something which was requested before at least once, and made by the Twitter team but now no longer exists (sadly). Another example of a similar program is this Wordpress plugin which does almost the same thing but saves the data to a database instead of a manipulable (is that a word?) file, and is only usable as a Wordpress plugin and not a standalone program. Another example is this, but it only takes 250 tweets (silly API, why do people even bother?) and only downloads your tweets, and doesn’t let you download those of anyone else. 

If anyone can think of any other useful functions, please Twitter them to me (@vpjayant). 

Bridge Bidding Assistant

This one was more complicated, but would be really useful if I managed to succeed. There are several possibilities for a program; a program in which you input a hand and it suggests what to big, a program which tests your bidding skills by showing you hands and asking you what you would bid. It would be limited by the fact that there are thousands of different conventions to choose from, and so would either have a huge database of them (difficult to implement) or it would be limited to only a common few. It would be based on the ACOL bidding system, and for users I could look to members of Bridge Club and some of the bridge playing teachers in the school such as @RAB. It would use a simple input method of either a drop-down list of card magnitude and then another for suit, or a code based one (E.g. 2C for the 2 of clubs, KH for the king of hearts). For the first idea, you would input your hand using that method, it would count your points, check whether it is balanced or not and then suggest an opening bid. However, the rebid will be considerable more complicated (see above) and so I don’t know how in depth the rebids will be able to get until I begin the design (if I choose to do this project). For the second idea, it would simply be a matter of having it generate a hand and ask you questions about it (”how many points does this hand have?”, “If your partner bid XX, what would you respond?”, “What would you open with with this hand?”). If you get the answer right, it will take you to the next question, but if you get it wrong, it will explain why you got it wrong. 

Had my SATs this morning, woke up nice and early to go to Hillingdon to do the test. It went pretty well, I managed to complete my goal of putting a reference to Bill Gates in my essay, so I was happy. After the exam I went to Melissa’s for the afternoon, then went to the concert.

I don’t actually know who I went to see at the concert. Originally I wanted to see DragonForce, but after listening to some Turisas I decided that they are awesome (I had also heard bad things about DragonForce live) and that they may be better. I was right. In terms of general awesomeness and showmanship, Turisas won hands down, but DragonForce still got the crowd going. DragonForce was super loud, and I was standing on the right hand side, so I now can’t hear out of my right ear; just a constant dull ringing. It was so worth it though. Sadly Turisas won’t be back for a loooong time :(. 

Finally decided to write a new post after far too long. Nothing interesting enough to write about has happened. Now it has. The end.

20/07/08: Day 4, on Sundays Chirag is closed, so we had no work to do. We spent most of the day at the Chirag centre but later went up to Sitla later for some food (candy and Sprite, of course!). In the evening we sorted out who was staying in which village. Aoi, Esther, Kate and I stayed in Ritha, and Viveka, Ushashi, Ben, Zaffie and Josh stayed in the other one.

21/07/08: Day 5, today we went to our villages in the morning. I went to Ritha with Aoi, and we started our “work”. Initially we just sat in the Chirag office for a little while then went out to a nearby plot and planted a little grass until lunch. It was raining really hard, so after lunch we waited a while in the office and started planting again. Aoi and I went back to our house which was at the bottom of a very big hill (actually, I think it was at the bottom of a valley) and we killed flies for a few hours. It was… fun.

22-25/07/08: This week we basically worked every day in different places, doing pretty much the same thing; separating and planting grass. Aoi used whatever he could find (a big rock) as digging implements to help loosen the ground, and I used a very nice stick I found to make holes for the roots.

26-28/07/08: No work on these days; weekends there is little to do and Chirag is closed on Sundays. We went up to Sitla a few times and made one shopkeeper very happy with many, many orders of Maggie and this awesome omelette. The weather really cleared up, and I was able to find Pokemon-shaped clouds everywhere! After a lot of stress, we finally finished all the admin for the trip, organised tickets for the train journey, and paid for everything.

29/07/08: Final day, it was interesting. We got driven to the station (a 2 hour drive) by jeep and had dinner with at a south-Indian restaurant near the station. The journey was scary, I sarcastically said that we would be split up into several different carriages as a joke to scare Viveka. Her reaction was funny. What was funnier was the fact that I was right. The journey was terrible for the first half, it smelled, and we had one seat for four of us (Zaffie, Ben, Kate and I). After a few hours it improved as we were officially designated seats and everyone but me went to sleep. I managed to almost listen to every song in my iPod on that journey.

Overall it was a good experience, we got wet and tired and hurt, but it was good. I learned a lot. For example, western toilets are a godsend.

If you have 12 coins and one of them weighs more or less than all of the others. You have a balance scale; what is the smallest number of balances that can be done in order to find it?

Farhan at first said 11, where he took any coin and weighed it against each of the others which was possibly the least efficient method. It worked, but it was way too complicated.

Erroll and I both thought that he was wrong and that we could get it in a smaller number of iterations. We thought it would be 7. Both of our methods were fairly similar, so I’m only going to talk about mine. I thought that you could name them from 1-12, and then measure 1 and 2, 3 and 4, 5 and 6 etc until you find an uneven pair. This takes a maximum of 6 iterations. Then measure one coin in that pair against another coin which you know is normal; if the coins balance, then the leftover coin is the fake, if they don’t, then the one you measured was the fake. This makes the number of iterations be 7.

So, we had it down to 7, which I thought was pretty good, but after some thought, I discovered I could bring it down even further. Labelling them from 1-12 again, you measure 1,2 against 3,4; 5,6 against 7,8 etc. This will take a maximum of 3 iterations. From this you can find that there is one set which is not balanced. Of the four that are left, remove one from each side and measure them (for example, if the remaining ones were 1, 2, 3 and 4, then you would take 2 and 4 away and measure 1 and 3). This is the 4th iteration. There are two possible outcomes from this measurement:

1. They balance: if they balance, then you know that they are normal, and so measure one of the other two coins against any of the coins that you know is normal. If these balance, you know the remaining coin is the fake, and if they don’t, then you know the coin you just balanced is the fake.
2. They don’t balance. Then you just measure either of the coins against any of the coins that you know is normal. If these balance, you know the remaining coin is the fake, and if they don’t, then you know the coin you just balanced is the fake.

Yes I did just copy and paste that last bit. But that was 5 iterations.

I stopped thinking about it for a while, but whilst in India I was quite bored so looked at it again, this time trying strange measurement combinations so that you can find the odd one without actually having to eliminate any. I thought that this would be the only way to get it in less than 5, as if you get it down to a choice of two, that’s at least one iteration, and bringing it down would be difficult. After some thought, and a few false methods, I managed it in 3.

The explanation is extremely complicated and I wrote it out in some strange form of pseudo-code which probably only I could understand, so I’m going to try to simplify it here. I’m going to use vague pseudo-code because otherwise this will be too long.

First, as usual, label them 1-12. The first measurement is 1,2,3,4 against 5,6,7,8. There are three outcomes from this.

1.     1,2,3,4 = 5,6,7,8: In this case, you then measure 1, 9 against 10, 11. There are then three more outcomes:

a.     1, 9 = 10, 11:  12 is the fake

b.    1, 9 < 10, 11: you now measure 10 against 11, with three more possible outcomes:

i.    10 = 11: I is the fake

ii.    10 < 11: 11 is the fake

iii.    10 > 11: 10 is the fake

c.     1, 9 > 10, 11: you now measure 10 against 11, with three more possible outcomes:

i.    10 = 11: I is the fake

ii.    10 < 11: 10 is the fake

iii.    10 > 11: 11 is the fake

2.     1,2,3,4 > 5,6,7,8: In this case, you then measure 1,2,5 against 3,6,12. There are then three more outcomes:

a.     1,2,5 = 3,6,12: you now measure 7 against 8, with three more possible outcomes:

i.    7 = 8: 4 is the fake

ii.    7 < 8: 7 is the fake

iii.    7 > 8: 8 is the fake

b.    1,2,5 < 3,6,12: you now measure 3 against 12, with two more possible outcomes:

i.    3 = 12: 5 is the fake

ii.    3 > 12: 3 is the fake

c.     1,2,5 < 3,6,12: you now measure 1 against 2, with three more possible outcomes:

i.    1 = 2: 6 is the fake

ii.    1 < 2: 2 is the fake

iii.    1 > 2: 1 is the fake

3.     1,2,3,4 < 5,6,7,8: In this case, you then measure 1,2,5 against 3,6,12. There are then three more outcomes:

a.     1,2,5 = 3,6,12: you now measure 7 against 8, with three more possible outcomes:

i.    7 = 8: 4 is the fake

ii.    7 < 8: 8 is the fake

iii.    7 > 8: 7 is the fake

b.    1,2,5 < 3,6,12: you now measure 1 against 2, with three more possible outcomes:

i.    1 = 2: 6 is the fake

ii.    1 < 2: 1 is the fake

iii.    1 > 2: 2 is the fake

c.     1,2,5 < 3,6,12: you now measure 3 against 12, with two more possible outcomes:

i.    3 = 12: 5 is the fake

ii.    3 < 12: 3 is the fake

Wow, that was long. I’m finally done typing that up, I hope it is understandable.

Now to get back to writing my personal statement. Bah.

I also got a new Western Digital HD which I used to move all my music from my laptop to my PS3, finally using that 40GB of storage (or at least some of it). I started learning PHP and after a lot of fail, I managed to get it set up in Ubuntu. I’m going to be using vpjayant.com as my testing ground for my PHP stuff.