By Herbert J. Bernstein

© Copyright 2000 Herbert J. Bernstein

Two houses built on the same size lot may have very different organizations of rooms. One may be a colonial, another an A-frame, another a ranch. These are different types of houses. Similarly we may have different types of variables that we use in allocating computer memory (see the discussion of computer memory). We abstract the organiztions of data in variables and speak of data types. Implict in any consideration of variable types and data types is an understanding of what operations may be performed, and how those operations will be performed. When we make the operations, actions or methods associated with a type explicit, we speak of it as a class and the instances of the class as objects.

The most elementary data type on most computers is the bit, which is not often provided as a variable type in programming languages, but which is used as the type in terms of which all other types are directly or indirectly defined. (There have been computers based on non-binary number systems, but the common practice these days is to reference everything to an underlying binary representation.) See the discussion of numbers.

In any given programming language, some data types are considered basic parts of the language and are called primitive or elementary types, and other organizations of data are derived from those types.

• Primitive Types
  • May reflect constructs supported by hardware operations
  • May require extensive software support
  • Integer
    • Whole numbers
    • Unsigned
    • Signed
      • Two's complement
      • One's complement
      • Signed magnitude
      • Biased
  • Reals
    • Fixed point
    • Floating point
    • Decimal (decimal digit strings)
    • Excess-6
  • Character data
    • Choice of character representation
    • Fixed or variable length
    • Design of character strings
      • List or array
      • Packed or not
      • Counted, terminated, blocked
      • Direct or indirect references, descriptors
      • Static or dynamic allocation
      • Accessibility of substrings
      • Collation issues
  • Logical data
    • Boolean operations
    • Whole word vs. Bit-packed
    • Distinguish from integers or not
• Simple derived types
  • Enumerations
  • Ranges
• Complex derived types
  • Arrays
    • Storage mapping issues
    • Indices and dope vectors
    • Associative arrays
    • Interaction with secondary storage
    • Array slices
  • Records, structs and unions
    • Strongly related to direct access secondary storage
    • Early languages lacked operations on records
    • Unions intorduce type checking issues
  • Sets
    • May be ordered or unordered
    • Usually handled with bitmaps or character strings
  • References and pointers
    • Powerful tool
    • Difficult to check
    • Arithmetic on pointers