My interview process
on
Defining Google
·
· Score: 5, Insightful
So I was just hired by Google right out of college, I'm graduating from Carnegie Mellon Univ in May. Google came to our school to run interviews around October. Here's a summary of the process:
Google set up shop at one of our job fairs with about 120 other companies. Anyone could walk up to the booth and give them their resume and talk to the people there. If they liked you, they e-mailed you later for an interview
There are then 2 on campus interviews. The first on one day, and if the interviewer likes you, then you get called back to the next day. If that interviewer likes you, you get flown out to their head quarters by San Fransisco. For my trip, there were 8 kids from CMU, and about 25 total from 7 other schools.
Out in their headquarters, you have 3 interviews with different sets of people. One of my interviews had 2 people each asking questions. They also feed you though and give tours of the campus. They definitly treated the applicants very well out there, great hotel, very nice all around.
Finally a week later people found out if they had offers or not. I heard rumors that in interviews with Google, each interviewer has 'veto' power, so if just one person didn't like you, no luck.
As for interview topics, there was a large range. Most were data structure concepts and problem solving. One interview was very unique though, the guy had a sheet of general software eng questions ranging over topics such as application design and testing, server-client software design, internet concepts. He would ask you just to describe a general topic, and see how much you could explain about it. For instance, one question was like "If you wanted to improve one of your programs, what would you do?" So you had to talk about testing, bottle-necks, better hardware, etc, just about everything.
As for coding questions, some people have complained in this thread that they don't display if you are a good coder, and I quite disagree. The purpose of those questions are to find out how you think, not how you code. They look for if you can logically lay out a problem in entirety and solve it one step at a time. Yes it's under a stress you would normally not have, but I think the stress helps sometimes. The part that all my interviewers spent the most time with was if I could improve my current solution. To see if you could do it with less memory, less cpu. The hardest part is just not knowing if there's something obvious that you should see. But a hint, start with the worst solution, then 'think up' a better solution while you're writting out the first. Do not try writting out the optimal solution from scratch from your head. They want to first see that you can solve it, but then to make sure that you don't settle for that solution and instead cringe at every line to make sure it's perfect.
Other tips I would suggest, spice up your resume with team projects. Also, the breadth of experience you have, not depth. As for positions at Google, I was hired as a Software Eng, which means I can work on just about any project, so they wanted people with skills in many areas. Lastly, don't be afraid of saying 'I don't know' to a question. I did this a few times for 'quiz' questions where I knew I could just go look up the answer (for instance, one question was 'list and define all the different type casts in c++'). But just don't wait time trying to make up something or giving a wrong answer.
After that babble, I also wanted to mention that every interviewer seemed to love their job there, like some people in the thread have said.
I hope this might give some insight into the process, although it's specific for college grads. But the general idea I got was that Google was looking for genearally bright people with decent experience and good team skills.
Re:I feel that the professor let his students down
on
A College Guide to EA
·
· Score: 2, Informative
I just wanted to let you in on a little insight about the professor who wrote this article, since I know him personally.
He teaches a class at CMU called Building Virtual Worlds, which I'm in. The class is a boot camp for the ETC (Entertainment Technology Center) majors to get them to learn how to work in groups under extreme time requirements. Students put easily 40-60, and usually up to 80 hours into this single class. When EA came, many of the students were talking about the crazy time requirements for there and how it was just like the BVW class.
But the Prof never really talks about EA in class or during discussions and doesn't try to influence anyone at all. It's more of a real world experience thing for him than a 'this is how it should be'. Even though the class is also really time intensive.
But this prof is also a brilliant guy for his ability to get people to really want to work really hard to create something. Grades don't really matter to him, just creating something new, different, and that pushes the limits of what's has never been done before.
Sure, it might seem that way if you thought of the brain as a computer, but that is sometimes not a very good analogy. Here's a different way of viewing someone catching a ball: You see the ball leave the hand and recognized the shape. The ball now moves in relation to you, if you are to far to the left, the ball seems to go right. If you are right in front of it, it moves right up and down. So the brain sees that the ball is moving left, and you move right.
You are correct the brain can do many complicated calculations, but not nececarally vector calcualtions like a computer. The brain is extreamly good at taking 2-d images, and creating 3-d representations. Remember, our eyes only get us 2-d pictures (granted, two of them). But the brain allows us to perceive 3 dimensional objects and space.
It seems that every time a new technology comes out, people are relating that to the brain. 150 years ago when water pipes were a big thing, they thought that's how reflexes work. When telephones came out, they said the brain was just a large switching station. Now with the computer, we have computers in our head creating vectors and doing "a monstrous calculus problem". Analogies are great, but remember we really don't know that much on how the brain works.
Oh they'll let it happen, not much they can do. But they will make sure they make thier cut and themsome. I'm sure they will require a percentage of every blank disk sold, just like CD's. This is to compensate for copyright infringments. Then when people do copy mp3/ogg's to them, they will cry foul, and demand more money. 'tis the way.
Another great free site is thevax.org these guys have set up some VAXen machines on the internet for free for people to use. All you need to do is submit a form for a free account. So if you want some alternatives, here they are. Already a lot of users from around the world.
alright, just kinda found this ammusing seeing the story again. Here's the original link: here. My friend is the one who did this, when I saw the story here, I thought "Alrdy done, big whoop". But this kid is awesome, just in his free time he does this stuff. When he showed it to me (he lived a floor above me last year), It was really awesome. So many people thought it was a hoax, but it really wasn't. The greatest part is that the kid didn't think it was such a big deal. I'm shitting in my pants with harddrives doing this, and he just thinks of it as a week project for something to do. Very smart kid.
One other thing, the first post (submitted by me) took up 65% of Carnegie Mellon's bandwidth for that week. No slashdotting, but 65% of all the bandwidth cmu used up that week was people going to this page. The power we have.
I really think that the idea of songs for a quarter or a nickel really could work for the RIAA. Sure people are posting that there are still P2P networks and that idea would have worked pre-napster, but I think it still could. The problem with most P2P networks is that you really have no promisses about what you get, or how fast you get it. Usually with songs they are fair rips and are titled correctly. But imagine a site where you had loads of bandwidth, and had every new song (and old) out there. I'd pay money to have access to that. They could have good rips in a variety of formats, and also track better what people are really listening to.
What would I pay? I'd probably pay upto around 5 bucks a month. That's 60 a year, and get enough subscribers, I don't see the problem. Bandwidth costs could be covered easily and you really don't lose a whole lot. That is except the enormous profits from CD sales, what this really is all about. But you could offer so much on a website like this, music videos, interviews, bands could keep websites up there. At least we have one coherant writter among our point of view, which I'm so pleased about. For people who don't RTFA, she got over 2200 emails, and responded to every one. Even got her account suspended twice for spamming while she was responding back. Insane.
The SOHO OmniView USB/Audio switches USB keyboards and mice between computers without having to reinstall them after each switch, but the same can't be said for USB peripherals. Each time you select a new PC, the OS detects new USB peripherals.
I don't think that this point was made enough by the reviewer. I bought a 4 port KVM (forget which maker), but it took USB mouse and keyboard inputs, instead of PS/2. Basicly the fact that the computer needs to install the devices again is THE MOST ANNOYING THING IN THE WORLD. After I discovered this little problem, I returned the product and got a ps/2 input one. In the review, the first product doesn't have to install the mouse and keyboard, but I would imagine for the USB devices, it would be the same annoyance. If you have a KVM, you want to be able to switch from one computer to the next instantly. Here you cannot, I just don't think the reviewer spent enough time with it to really see how annoying this fact is.
Think about the potential for abuse if it falls into the wrong hands.
Right, but you could say that with just about anything. What about nuclear weapons? The government has special classifications for information dealing with anything that has to do with nuclear weapons. I'm sure these things will be similar, and no one will really get at them.
About LINEAR (the guys who found the big rock)
on
A Rock Moves In Space
·
· Score: 5, Informative
It was first seen on the night of 5 July, picked up by the Linear Observatory's automated sky survey programme in New Mexico, in the southern US.
I work at Lincoln labs and acutally know the people running the LINEAR project (they are so proud that they are the best in the world, let me tell you). But for the rest of you, here is their website.
They find more than half of the new NEO (Near earth orbit) asteroids each year that are found. They have a telescope down in New Mexico and have the largest CCD (2560x1960 res) in the market. That's the thing that takes a digital image of the sky and compares it to past images to see if any "stars" have moved...i.e asteroid. The higher resolution you can get, the further out you can see. From their webpage, you can see they have found at least 951 NEO's. So there are a LOT of asteroids comming near us. But in space, near is still very far away. So unpack those bunkers and return to Real Life, we're still safe for a while. Also, the rate of finding new NEO's is decreasing, so that means that we've (humans) found most of the asteroids that can endanger us.
(most of that was taken from this post of mine from a while ago)
I was in the Odyssey of the Mind program for six seperate years. My team made it to the world finals 3 of those. The greatest thing about OM is the aspect that it encourages doing your own work, and working as a team. One of the unique things about OM is the fact that NO OTHER PERSON THAN THE TEAM MEMBERS can help. If a team wants to do something, they must do it themselves. They must research, learn, experiment, build, whatever. A few small exceptions, where safety is a factor apply. This has deffinitly changed my life by increasing the ability to listen to other's ideas, and learn something yourslef, instead of getting someone else to do it. I judge at OM competitions when I can, and just like giving back to the program. Any program that is brought in by the schools should have this kind of aspect to it, to teach kids both technical, and social skills. And OM definitly does teach technical skills, I can build a balsa wood structure now that weights less than 18 grams, 8 inches tall, and holds over 1000 pounds.
I actually have a 1 farad cap. that's maybe 2 cm long, and 700 mm diameter. Powerful little bugger. It only handels 2 Volts though. So technology has really gotten these things pretty damn small, holding a damn lotta charge. I got this thing in a physics set in my physics class at college.
Who pays $9500 for a piece of software?
Well, clearly not the average consumer. Not all software can be bought from the local compUSA for 50 bucks or less. So much commercial software costs much more than $9500. I am actually working with a software package which costs over 100 K per seat. Now, yes these are expensive, and usually they are worth it. And yes, they might only get 1000 sales, but this is the exact point. With that few sales, every one counts. If a person from a company could go and download this program for free, then that 1 sale means a lot of loss.
Also, these expensive programs are usually very unique and if a company was going to download to crack it, they usually were going to buy it anyways. This is compared to adobe photoshop, which almost any warez user has, but very few of them would have bought it anyways. Your point about people learning it, and possibly buying a legal copy later fits in perfectly with photoshop though.
One thing that really pisses me off is these companies compaining that they are getting ripped off when they don't put as much effort as they could into securing their program. If a program costs 500 (photoshop), maybe you should have a more secure system than just putting in a single CD key without checking it with a database. Now, I know that there really is no suching thing as securty in these type of situtations. But if you make a program not worth the cracker's time to crack, by making it too hard, then they might give up. And it isn't too hard to make a program that difficult to crack. If you have 1000 sales total, then you could do any number of things, like compiling unique binaries, secure-server key checking...etc. These companies are like the RPAA that want their money with no work done on their part. Times change, people change, so just hire one of these crackers, and friggen make a program that takes too long to break.
For the cheaper programs, I hate those statistics that millions are lost in sales. Where did they get lost? How can you say that people would have bought the program? Just like the figurative losses by the RPAA from online trading. I can't say that what the crackers do is right, but it's the companies problem to try harder to prevent their programs from being cracked.
Re:Current bug in Windows I/O....
on
Pet Bugs?
·
· Score: 1
Acutally, I was just working with files like that too. I would open a file in binary for writing in one program, and the same file for reading in another program. When I wrote to the file, the file pointer in the reader moved sometimes, but not always to the new end of file, without even reading anything. Oh well.
Actually, that's true, and not true. The courts can kinda create new law. Maranda rights are something that the courts created, not congress. These rights came at a time where the supreme court started giving more rights to the accused. While it is not a law, it is still enforced by the courts very strongly. It is help up by precident, which sometimes is stronger than some laws.
You are correct. In neural networks, you take a large part of your data set and denote it your "train" set. Next you take a chunk and call it your "test" set. Lastly the rest of the data is in the "validate" set. In this example, maybe the years 1950-1995 could be used for test and train, then the last 5 years for validate. The all the good predictions with accurate validate sets could be used reliably for the next 50 years to come. What they are doing now is just guessing and using the power of the distributed power very brute-force ish.
While I agree with you, the problem is that so many parents do not know their children. How many times have we seen parents like those of the Columbine kids. Many parents are good and know when a game should be allowed to be played by their kid. But there are SO many parents who don't give a shit, and don't want to know. I think the government doing this is the lesser of two evils. This way, the kids can not buy it, but the parents can if they deem it ok. This is no banning letting kids play it, just selling to them.
My university, Carnegie Mellon, competes in this competition. We usually do very well, I think second last year or the year before. But if anyone would like to see a video of the action, here's one here. Very cool stuff.
Slashdot Mirror
So I was just hired by Google right out of college, I'm graduating from Carnegie Mellon Univ in May. Google came to our school to run interviews around October. Here's a summary of the process:
Google set up shop at one of our job fairs with about 120 other companies. Anyone could walk up to the booth and give them their resume and talk to the people there. If they liked you, they e-mailed you later for an interview
There are then 2 on campus interviews. The first on one day, and if the interviewer likes you, then you get called back to the next day. If that interviewer likes you, you get flown out to their head quarters by San Fransisco. For my trip, there were 8 kids from CMU, and about 25 total from 7 other schools.
Out in their headquarters, you have 3 interviews with different sets of people. One of my interviews had 2 people each asking questions. They also feed you though and give tours of the campus. They definitly treated the applicants very well out there, great hotel, very nice all around.
Finally a week later people found out if they had offers or not. I heard rumors that in interviews with Google, each interviewer has 'veto' power, so if just one person didn't like you, no luck.
As for interview topics, there was a large range. Most were data structure concepts and problem solving. One interview was very unique though, the guy had a sheet of general software eng questions ranging over topics such as application design and testing, server-client software design, internet concepts. He would ask you just to describe a general topic, and see how much you could explain about it. For instance, one question was like "If you wanted to improve one of your programs, what would you do?" So you had to talk about testing, bottle-necks, better hardware, etc, just about everything.
As for coding questions, some people have complained in this thread that they don't display if you are a good coder, and I quite disagree. The purpose of those questions are to find out how you think, not how you code. They look for if you can logically lay out a problem in entirety and solve it one step at a time. Yes it's under a stress you would normally not have, but I think the stress helps sometimes. The part that all my interviewers spent the most time with was if I could improve my current solution. To see if you could do it with less memory, less cpu. The hardest part is just not knowing if there's something obvious that you should see. But a hint, start with the worst solution, then 'think up' a better solution while you're writting out the first. Do not try writting out the optimal solution from scratch from your head. They want to first see that you can solve it, but then to make sure that you don't settle for that solution and instead cringe at every line to make sure it's perfect.
Other tips I would suggest, spice up your resume with team projects. Also, the breadth of experience you have, not depth. As for positions at Google, I was hired as a Software Eng, which means I can work on just about any project, so they wanted people with skills in many areas. Lastly, don't be afraid of saying 'I don't know' to a question. I did this a few times for 'quiz' questions where I knew I could just go look up the answer (for instance, one question was 'list and define all the different type casts in c++'). But just don't wait time trying to make up something or giving a wrong answer.
After that babble, I also wanted to mention that every interviewer seemed to love their job there, like some people in the thread have said.
I hope this might give some insight into the process, although it's specific for college grads. But the general idea I got was that Google was looking for genearally bright people with decent experience and good team skills.
I just wanted to let you in on a little insight about the professor who wrote this article, since I know him personally.
He teaches a class at CMU called Building Virtual Worlds, which I'm in. The class is a boot camp for the ETC (Entertainment Technology Center) majors to get them to learn how to work in groups under extreme time requirements. Students put easily 40-60, and usually up to 80 hours into this single class. When EA came, many of the students were talking about the crazy time requirements for there and how it was just like the BVW class.
But the Prof never really talks about EA in class or during discussions and doesn't try to influence anyone at all. It's more of a real world experience thing for him than a 'this is how it should be'. Even though the class is also really time intensive.
But this prof is also a brilliant guy for his ability to get people to really want to work really hard to create something. Grades don't really matter to him, just creating something new, different, and that pushes the limits of what's has never been done before.
Sure, it might seem that way if you thought of the brain as a computer, but that is sometimes not a very good analogy. Here's a different way of viewing someone catching a ball:
You see the ball leave the hand and recognized the shape. The ball now moves in relation to you, if you are to far to the left, the ball seems to go right. If you are right in front of it, it moves right up and down. So the brain sees that the ball is moving left, and you move right.
You are correct the brain can do many complicated calculations, but not nececarally vector calcualtions like a computer. The brain is extreamly good at taking 2-d images, and creating 3-d representations. Remember, our eyes only get us 2-d pictures (granted, two of them). But the brain allows us to perceive 3 dimensional objects and space.
It seems that every time a new technology comes out, people are relating that to the brain. 150 years ago when water pipes were a big thing, they thought that's how reflexes work. When telephones came out, they said the brain was just a large switching station. Now with the computer, we have computers in our head creating vectors and doing "a monstrous calculus problem". Analogies are great, but remember we really don't know that much on how the brain works.
Oh they'll let it happen, not much they can do. But they will make sure they make thier cut and themsome. I'm sure they will require a percentage of every blank disk sold, just like CD's. This is to compensate for copyright infringments. Then when people do copy mp3/ogg's to them, they will cry foul, and demand more money. 'tis the way.
Another great free site is thevax.org these guys have set up some VAXen machines on the internet for free for people to use. All you need to do is submit a form for a free account. So if you want some alternatives, here they are. Already a lot of users from around the world.
alright, just kinda found this ammusing seeing the story again. Here's the original link:
here. My friend is the one who did this, when I saw the story here, I thought "Alrdy done, big whoop". But this kid is awesome, just in his free time he does this stuff. When he showed it to me (he lived a floor above me last year), It was really awesome. So many people thought it was a hoax, but it really wasn't. The greatest part is that the kid didn't think it was such a big deal. I'm shitting in my pants with harddrives doing this, and he just thinks of it as a week project for something to do. Very smart kid.
One other thing, the first post (submitted by me) took up 65% of Carnegie Mellon's bandwidth for that week. No slashdotting, but 65% of all the bandwidth cmu used up that week was people going to this page. The power we have.
I really think that the idea of songs for a quarter or a nickel really could work for the RIAA. Sure people are posting that there are still P2P networks and that idea would have worked pre-napster, but I think it still could. The problem with most P2P networks is that you really have no promisses about what you get, or how fast you get it. Usually with songs they are fair rips and are titled correctly. But imagine a site where you had loads of bandwidth, and had every new song (and old) out there. I'd pay money to have access to that. They could have good rips in a variety of formats, and also track better what people are really listening to.
What would I pay? I'd probably pay upto around 5 bucks a month. That's 60 a year, and get enough subscribers, I don't see the problem. Bandwidth costs could be covered easily and you really don't lose a whole lot. That is except the enormous profits from CD sales, what this really is all about. But you could offer so much on a website like this, music videos, interviews, bands could keep websites up there. At least we have one coherant writter among our point of view, which I'm so pleased about. For people who don't RTFA, she got over 2200 emails, and responded to every one. Even got her account suspended twice for spamming while she was responding back. Insane.
I would say it would be more of $101, gotta make a profit don't ya know (maybe a little more for lawyer fees)
The SOHO OmniView USB/Audio switches USB keyboards and mice between computers without having to reinstall them after each switch, but the same can't be said for USB peripherals. Each time you select a new PC, the OS detects new USB peripherals.
I don't think that this point was made enough by the reviewer. I bought a 4 port KVM (forget which maker), but it took USB mouse and keyboard inputs, instead of PS/2. Basicly the fact that the computer needs to install the devices again is THE MOST ANNOYING THING IN THE WORLD. After I discovered this little problem, I returned the product and got a ps/2 input one. In the review, the first product doesn't have to install the mouse and keyboard, but I would imagine for the USB devices, it would be the same annoyance. If you have a KVM, you want to be able to switch from one computer to the next instantly. Here you cannot, I just don't think the reviewer spent enough time with it to really see how annoying this fact is.
That site is ok, but I still prefer slashdot.org. Your site is so unstable, it seems to revert back to mine.
And for those who don't think that this might affect them in the slightest, look at the current webpage in your location bar.
Think about the potential for abuse if it falls into the wrong hands.
Right, but you could say that with just about anything. What about nuclear weapons? The government has special classifications for information dealing with anything that has to do with nuclear weapons. I'm sure these things will be similar, and no one will really get at them.
It was first seen on the night of 5 July, picked up by the Linear Observatory's automated sky survey programme in New Mexico, in the southern US.
I work at Lincoln labs and acutally know the people running the LINEAR project (they are so proud that they are the best in the world, let me tell you). But for the rest of you, here is their website.
They find more than half of the new NEO (Near earth orbit) asteroids each year that are found. They have a telescope down in New Mexico and have the largest CCD (2560x1960 res) in the market. That's the thing that takes a digital image of the sky and compares it to past images to see if any "stars" have moved...i.e asteroid. The higher resolution you can get, the further out you can see. From their webpage, you can see they have found at least 951 NEO's. So there are a LOT of asteroids comming near us. But in space, near is still very far away. So unpack those bunkers and return to Real Life, we're still safe for a while. Also, the rate of finding new NEO's is decreasing, so that means that we've (humans) found most of the asteroids that can endanger us.
(most of that was taken from this post of mine from a while ago)
I was in the Odyssey of the Mind program for six seperate years. My team made it to the world finals 3 of those. The greatest thing about OM is the aspect that it encourages doing your own work, and working as a team. One of the unique things about OM is the fact that NO OTHER PERSON THAN THE TEAM MEMBERS can help. If a team wants to do something, they must do it themselves. They must research, learn, experiment, build, whatever. A few small exceptions, where safety is a factor apply. This has deffinitly changed my life by increasing the ability to listen to other's ideas, and learn something yourslef, instead of getting someone else to do it. I judge at OM competitions when I can, and just like giving back to the program. Any program that is brought in by the schools should have this kind of aspect to it, to teach kids both technical, and social skills. And OM definitly does teach technical skills, I can build a balsa wood structure now that weights less than 18 grams, 8 inches tall, and holds over 1000 pounds.
I actually have a 1 farad cap. that's maybe 2 cm long, and 700 mm diameter. Powerful little bugger. It only handels 2 Volts though. So technology has really gotten these things pretty damn small, holding a damn lotta charge. I got this thing in a physics set in my physics class at college.
Ya, long day, my bad with MPAA.
Who pays $9500 for a piece of software?
Well, clearly not the average consumer. Not all software can be bought from the local compUSA for 50 bucks or less. So much commercial software costs much more than $9500. I am actually working with a software package which costs over 100 K per seat. Now, yes these are expensive, and usually they are worth it. And yes, they might only get 1000 sales, but this is the exact point. With that few sales, every one counts. If a person from a company could go and download this program for free, then that 1 sale means a lot of loss.
Also, these expensive programs are usually very unique and if a company was going to download to crack it, they usually were going to buy it anyways. This is compared to adobe photoshop, which almost any warez user has, but very few of them would have bought it anyways. Your point about people learning it, and possibly buying a legal copy later fits in perfectly with photoshop though.
One thing that really pisses me off is these companies compaining that they are getting ripped off when they don't put as much effort as they could into securing their program. If a program costs 500 (photoshop), maybe you should have a more secure system than just putting in a single CD key without checking it with a database. Now, I know that there really is no suching thing as securty in these type of situtations. But if you make a program not worth the cracker's time to crack, by making it too hard, then they might give up. And it isn't too hard to make a program that difficult to crack. If you have 1000 sales total, then you could do any number of things, like compiling unique binaries, secure-server key checking...etc. These companies are like the RPAA that want their money with no work done on their part. Times change, people change, so just hire one of these crackers, and friggen make a program that takes too long to break.
For the cheaper programs, I hate those statistics that millions are lost in sales. Where did they get lost? How can you say that people would have bought the program? Just like the figurative losses by the RPAA from online trading. I can't say that what the crackers do is right, but it's the companies problem to try harder to prevent their programs from being cracked.
Acutally, I was just working with files like that too. I would open a file in binary for writing in one program, and the same file for reading in another program. When I wrote to the file, the file pointer in the reader moved sometimes, but not always to the new end of file, without even reading anything. Oh well.
This was actually already posted today in the supermodels in space bit. Please guys, at least try to read the articles.
Actually, that's true, and not true. The courts can kinda create new law. Maranda rights are something that the courts created, not congress. These rights came at a time where the supreme court started giving more rights to the accused. While it is not a law, it is still enforced by the courts very strongly. It is help up by precident, which sometimes is stronger than some laws.
You are correct. In neural networks, you take a large part of your data set and denote it your "train" set. Next you take a chunk and call it your "test" set. Lastly the rest of the data is in the "validate" set. In this example, maybe the years 1950-1995 could be used for test and train, then the last 5 years for validate. The all the good predictions with accurate validate sets could be used reliably for the next 50 years to come. What they are doing now is just guessing and using the power of the distributed power very brute-force ish.
While I agree with you, the problem is that so many parents do not know their children. How many times have we seen parents like those of the Columbine kids. Many parents are good and know when a game should be allowed to be played by their kid. But there are SO many parents who don't give a shit, and don't want to know. I think the government doing this is the lesser of two evils. This way, the kids can not buy it, but the parents can if they deem it ok. This is no banning letting kids play it, just selling to them.
At least one company is accepting the way things are. MPAA, you listening?
My university, Carnegie Mellon, competes in this competition. We usually do very well, I think second last year or the year before. But if anyone would like to see a video of the action, here's one here. Very cool stuff.
Just a little note, but CD sales actually went up after Napster hit the big time.