What's the most exciting area of computer science

[Craig]

See title :nerd:

[Fireblade]

I say Robotics, cuz I wanna see MECHA for crying out loud!!! MECHA!!!
(wipes the sweat off his face.)

[537]

Mecha! (http://world.honda.com/robot/)


At its early stage, we have what you are looking for....

[SunWuKong]

wireless
i see it happening in asia first though

[Fireblade]

Bah... Honda. Sorry, lack of respect cuz of the poor high school ricers who know shit about their cars.

[loserbutt]

3d!

[537]

Originally posted by Fireblade@Dec 19 2002, 10:47 AM
Bah... Honda. Sorry, lack of respect cuz of the poor high school ricers who know shit about their cars.
I have that same lack of respect. But instead of directing that negative energy to the maker of the product those high school ricers abuse, I laugh at the perpetrators themselves.

Bullshit. I can't stand Honduhs, wtf am I saying.

[ism]

There's CE and CS stuff mixed up, so maybe you meant "computing" in general. I'd have to go with molecular computing. There is some groundbreaking research in recent months that show promise. From storage capacity to computational possibilities, as well as tangents such as replicating human chemical processes, I think it will revolutionize computing as much as the microchip did.

From the list, Data Mining shows the most possibilities in terms of affecting our daily lives. A good example of current, rudimentary data mining is Google's contest and beta projects such as Google Sets. That stuff hints at the possibilities and that's just dealing with public information on the web. Now apply that to purchases, geophysical information, routes, and basically any transaction that can be quantified. Brokerages data mine their databases to squeeze fractions of pennies out of each transaction, when totalled, reach massive amounts of money. There are sociological applications as well. Amazon's recommendations system is based on yours and others' purchase and browsing habits. It's just the beginning.

[achtungbaby]

Holy shit! Someone answered "Data Mining" with me.

[SunWuKong]

see, my opinion of data mining is that it's really not that much of a field. it just sounds like what marketing people have been doing all along.

[ism]

Unfortunately "data mining" is often used as buzzspeak, as is "knowledge discovery," the trendy thing in data mining right now. I'm not sure what most people define it as, but to me it is fairly clear: data mining is finding coherent patterns and correllations (i.e., useful information) in raw datasets without knowing beforehand what you are looking for.

Marketing people may have been doing it all along, but this goes beyond that. People can only look through so much data and discover a limited type of patterns. Data mining uses AI and statistics to discover things humans can't, or didn't even know they could, look for. The applications aren't limited to figuring out you're going to buy Ramen noodles the next time you go to the supermarket.

Concrete example: a brokerage a friend of mine works for deals with many stock, bonds, futures, etc. transactions. There is 10gb of data each day. They are developing a process by which this data is mined, and pre-programmed statistical algorithms are used to pinpoint where every cent can be squeezed out of the transactions. This is now. In the future, what would happen is the data miner would look at this data, not know how to save money, but yet, it would discover ways to do it.

Some other application may deal with sociology. Perhaps a database of New York City subway and bus usage, combined with a database of events and the weather can be mined to figure out patterns in people's movement. The mining program could then predict accurately which lines would be used, and when, and the routing and scheduling could be optimized on the fly. Currently, the only way to figure this out is to run statistical model simulations based on past data.

[Craig]

Originally posted by ism@Dec 19 2002, 02:43 PM
There's CE and CS stuff mixed up, so maybe you meant "computing" in general.
I actually meant primarily from a software perspective. All of these items are covered in the IEEE-CS/ACM CS curriculum guidelines (although the merged Neural Networks into a subset of Machine Learning). However, in all fairness as this poll is arguably biased, and along with my CS degree I did minor in EE. Firsthand I saw the most divergence in VLSI from the offerings in the ECE and CS departments. I haven't had the luxury of working professionally in most of these interesting and progressive fields.

[himura-dono]

i agree with data mining. ism's desrcription is what i've been hearing from people expressing interest. i originally thought it was training people to be the numbnutz that dig up info dropped off to companies via spyware and the rest of that crap.