Lecture 24: What is computer science

1
Lecture 24What is computer science?
2
What have we covered

• Superficially
• The basics of programming in Java
• More deeply
• Basics of programming are the same in any
  language
• Youve learnd how to think logically about
  problems
• How to break down tasks into steps
• How to model problems with objects

3
Whats next?

• COMP 114 Advanced Programming
• Youve learned how to program
• This will teach you how to program well
• More Java
• Lots of Object Oriented design
• Larger projects, and how to deal with large
  programs
• COMP 121 Data Structures
• How to organize data for efficiency
• Lists, Arrays, Trees, Stacks, Queues, etc.
• COMP 122 Algorithms
• How to do things efficiently
• How fast do programs run?
• Sorting, Searching, Finding shortest paths
• Mathematical proofs

4
What else is there?
Yeah right, Im taking all that. You couldnt
pay me, jackass.

• What is computer science?
• Programming Languages
• Graphics
• Algorithms
• Theory of computation
• Compilers
• Operating Systems
• Networking
• Artificial Intelligence
• Enterprise Applications
• Simulation
• Numerical Analysis
• Parallel Computing

5
Ok, lets pick a few

• These might matter to you in the future
• Programming Languages
• Different tools for different jobs
• Parallel Scientific Computing
• One processor isnt enough!
• Enterprise computing
• If you work in business, youll need to know

6
Programming Languages

• Java is just one of many!
• Youve learned in Java
• But most languages are pretty similar
• Some examples
• Python
• Perl
• C
• C
• MATLAB

Duke, the Java mascot
7
Python, Java, Perl, C
for i in range(10) print Hello world, i
public class HelloWorld public static void
main(String args) for(int i 0 i lt 10
i) System.out.println(Hello world
i)
foreach i (0..9) print Hello World i
include ltstdio.hgt int main(int argc, char
argv) int i for(i 0 i lt 10 i)
printf(Hello world d, i)
8
Think of languages as tools

• Different tools are for different jobs
• Pick the right one!
• C
• Good for writing operating systems
• very fast
• C
• Slightly higher-level than C
• Used for lots of Windows programs
• Also a systems language
• Java
• Easy to debug, easy to write applications fast
• Slightly slower than the above

9
Think of languages as tools

• Different tools are for different jobs
• Pick the right one!
• Python
• Very fast to develop python programs
• Sometimes slow, cant do really low-level stuff
• Perl
• Created as a glue language
• Great for processing text
• Great for taking some programs output and
  passing it to others
• MATLAB
• Good for doing matrix math
• Good for prototyping scientific applications

10
Simulation

• Scientists are increasingly in need of
  computational simulation for
• Weather
• Molecular Biology
• Bombs
• Physics
• Social networks?
• These are big problems
• You cant do this on your laptop

11
Parallel Computing

• Sometimes one processor isnt enough
• You might need to program for 130,000 or so.
• Enter BlueGene/L, pictured at left
• Weve got one rack (2048 processors)
• The unit at left is 64 racks
• Its at LLNL, in California

12
Parallel Computing
13
Moores Law

• Gordon Moore
• Founder of Intel
• Number of transistors on a chip doubles every
  18-24 months
• exponential growth for processor speed
• Held true for a while, but slowing down
• To keep it going, well need to go parallel

14
Moores Law

• Graph at left is extrapolated
• includes earlier technologies
• Not just microprocessors
• All paradigm shifts in computation

15
Parallel Computingon your Laptop

• Intels new chips contain more than one physical
  processor
• Youll be able to run parallel applications on
  your laptop
• Multithreaded applications will become more and
  more necessary
• Youve learned about threads
• Future programs arent just going to use these,
  theyll need them
• Programs without wont get the most out of the
  hardware

16
Enterprise computing

• If you go into business, computer science will
  probably matter to you
• IT is one of the biggest costs for businesses
• Growing number of companies have in-house
  programmers
• If youre not writing the code, youll work with
  the programmers!
• You can feel their pain

17
Enterprise computing

• Hopefully this course gave you an idea of
• what it takes to write these programs
• Whats feasible and whats not
• Businesses deal with these problems every day
• Time to write program vs cost
• Its only going to be more relevant as more
  things get automated.