Computers in Science and Computer Networks

We discuss two computer-related topics today: the application of computers in science and the ways that computer networks operate.

Computers in Science

The early history of computing mostly involves scientific and engineering applications. Only in the last thirty years have business, home, and educational applications overshadowed scientific and engineering applications. But the applications of computers to science are hugely important: their share of the pie is smaller, but the pie itself is much bigger. From the point of view of the scientists, computers play an ever-larger role.

How are computers used in science? An approximately chronological list:

• Data manipulation
  • y = f(x), z = f(x₁, x₂, . . .)
  • Used extensively even in Manhattan-project era (R.P. Feynman)
  • It's important to verify the machine calculation with an occasional manual calculation!

• Decision-making calculations
  • We need to know how to set an instrumental parameter;
  • compute what ought to be, given various known or suspected values

• Process automation
  • We want to collect data overnight;
  • the computer performs the repetitive calculations

• Instrument control
  • move motors, open and close shutters; well-understood timings crucial
  • automatic recording of instrument status
  • alerting humans to fault states

• Plotting and graphing
  • Straightforward application of data manipulation
  • . . . but a huge time-savings over manual graphing!

• Visualization
  • Computer-generated model of physical system
  • Allows viewing system at arbitrary magnification or demagnification
  • Allows visualization of sub-structures
  • Allows modification of model without replacing physical model parts

• Simulations
  • use computer to mimic a physical experiment
  • allows mimicry of physically impossible experiments
  • provides prediction of results of physical experiment
  • narrows focus for real experiments down to small number of possibilities
  • pharmaceutical molecules, mechanical objects, planets

• Conventional Applications
  Scientists use computers for the same things that others do:
  • Word processing (papers, lectures, reports, . . . )
  • Spreadsheets (graphing, table-formatting)
  • Communications (e-mail, websites)
  • Information organization (databases . . .)
  • Finances

Computer Networks

Computers tended to be isolated from one another up until fifteen years ago. This changed:

• Technology for networking got better
• More single-user computers--communication required networks
• More consciously-designed distribution of processing loads

Kinds of Networks

• Local area network (LAN)
  • well-defined list of computers
  • under at least nominal control of an administrator
  • addresses assigned locally
  • usually relying on locally-maintained cabling
• Wide area network (WAN)
  • Network of networks--built up of LANs
  • no single administrator (happy chaos?)
  • the Internet is the biggest WAN, but not the only one
  • WWW is a way of organizing network resources on the Internet

Network Hardware

• Physical connections
  • Dedicated wiring: thick-wire, thinwire, twisted-pair
  • Piggy-backing on existing wiring: phone cables, even power lines
  • Non-cable connections: fiber-optic links, infrared, radio, satellite
• Network processor electronics
  • Provide hardware interface to the rest of the computer
  • Package up messages going out
  • Unload messages coming in

Network Software

• Lowest-layer
  • Information packaged up into small bundles called packets
  • Each packet carries enough header information to identify it as distinct from other packets
  • Packets arrive at destination and are identified as to source
• Intermediate layer
  Software determines what kind of data are being sent and which application should be receiving it
• Higher layers
  • Commands, text, displays
  • High level of interactivity now possible (CGI scripts, Java, . . . )
  • File-sharing