I. COMPUTER FUNDAMENTALS

A) Computer Hardware

A.1) Machine Model

• Input Unit: Function, speed and examples.
• Processor: MHz and examples.
• Output Unit: Function, speed and examples.
• Memory: Its distinctive role, types and capacities, bytes and the hierarchy of storage.

A.2) Evolution

• Brief History, Stored Program, Generations
• Processing: Batch, Time-shared, LAN, Client-Server (TCP/IP)

B) Computer Software

B.1) Programming & Languages

• Levels of Abstractions
• The 'Best' P/L!
• Source, Object and Executable Files

B.2) Classification of Softwares

• System Software: Operating systems, compilers, ....
• Application Software: WP, SS, Database; horizontal vs. vertical
• User Software.

B.3) Reusability

• Subprograms
• Libraries
• Objects

C) Data Representation

C.1) Objectives & Techniques

• Memory Model and the 2^n Rule
  • logical view of memory: One dimensional row of cells. Each is called a byte and has an address and a content.
  • Each byte is made up of 8 smaller cells, called bits.
  • A bit has two states: on or off, 0 or 1.
  • We "store" a thing by choosing a pattern for the bits that represents it.
  • Multiples of bytes: KB=1024 bytes, MB=1024KB, GB=1024MB.
  • In a cell of n bits we can represent any one of 2^n things.
• Mapping Data to Numbers: Encoding and Digitization
  • Sampling sound waves turns sound to a list of real numbers
  • Digitizing a picture turns it to a list of integers
  • Encoding characters by numbers turns each into one integer.

C.2) Data Types, Declaration & Round-off

• Representing Integers
  • Use binary numbers. 2's complement not covered in this course.
  • In 8 bits, the 2^n rule leads to 256 different numbers. They can be in the range -128 to +127. We'll take them to be in the +/- 127 range.
  • Can combine more cells to increase the range. Two and four cell combinations lead to 2 and 4-byte integers.
  • integer representation is exact; no rounding.
• Representing Reals
  • Using an exponent and normalization, we need only store the sign, the fraction and the exponent.
  • Computers use base-2 exponents instead of 10.
  • Due to the limited size of the fractional part, round-off errors are bound to occur. With n bits for the fraction, accuracy (i.e. relative error) is about 2^(-n).
• FORTRAN Declarations

           integer*1 code          ! +/- 127,    exact
           integer*2 count         ! +/- 32K,    exact
           integer*4 sample        ! +/- 2G,     exact
           real*4 amount           ! +/- 10^38,  accuracy 10^-6
           real*8 density          ! +/- 10^308, accuracy 10^-15
           logical*1 done          ! two values: .true. and .false.
           character*20 title      ! a string of upto 20 characters
           !---------------------------------------------------------
           ! You can optionally combine declarations of the different
           ! sizes of the same type:
           !---------------------------------------------------------
           character title*20, name*30, college*1
           integer code*1, count*2, sample*4
           real amount*4, density*8
           !---------------------------------------------------------
           ! You can also initialize as you declare:
           !---------------------------------------------------------
           character title*20 / 'York' /
           integer*2 count / 0 /
           real*4 amount / 1.0 /
           !------------------------------------------------- --------
           ! Note that title will have 20 characters even though
           ! we stored only 4 (they will be padded by blanks). 
           !------------------------------------------------- --------
           

C.3) Arrays