What is data type and why is it useful ?

What is Data Type ?
- Data type or type for short, is a class of value.
- Each variable in a program must have a data type (or type)
- Each constant in the program must also have a type
Categories of Types
- Built-in data types: these types come with the language and you can use them without having to "define" them
- User-defined data types: these data types are defined by the programmer to serve some purpose in his program
Built-in Data Types
- The purpose of built-in data types in programming langauges is to allow the programmer to represent commonly used values with ease
- Common built-in data types:
  - character
  - byte integer
  - short integer
  - normal integer
  - long integer
  - single precision float
  - double precision float
  - logical value type (TRUE/FALSE)
- NOTE: not every common types need to be provided by the programming language
  You can often use one type in place of another, e.g., logical value type is not provided in C, you can use character type to replace it: 0 means FALSE, non-zero means true.

C++ built-in data types:

Syntax	Name	use
char	character	Character or very short integer (-128..127)
short	short integer	uses 2 byte memory, value between -32768 and 32767
int	ordinary integer	uses 4 byte memory, value between -2147483648 and 2147483647
long	long integer	uses 8 bytes memory, value between -9223372036854775808 and 9223372036854775807
float	single precision float	uses 4 byte memory, absolute value between 1.4E-45 and 3.4E38
double	double precision float	uses 8 byte memory, absolute value between 4.9E-324 and 1.8E308

Built-in type and data encoding
- Recall that different type of values are encoded differently inside the computer: click here, here, and here,
Operations involving values from same built-in type
- Variables and/or constants of the same bulit-in type can be combined "naturally" - i.e., without having to perform any conversion operations - in operations such as +, -, * or /
- Example:
  float A, B, C; - Defines 3 float variables A = 4; B = 5; C = A + B;
  This is the similar situation as "adding apples and apples".
Operations involving values from DIFFERENT built-in types
- Variables and/or constants of the different bulit-in type CANNOT be combined WITHOUT having to perform any conversion operations FIRST !!! - in operations such as +, -, * or /
- Example:
  int A; - Defines integer float B, C; - Defines 2 float variables A = 4; B = 5; C = A + B; - value of A must be converted to FLOAT first !!!
  This is the similar situation as "adding apples and oranges". We must convert the apples to oranges first before we can add
- Example Program: (shows type conversion) --- click here
  - Save the program
  - Compile using: CC -S type-conv.C
  - Look at the output using: vi type-conv.s
  - Notice the additional instructions needed to convert the integer A2 to a float value before the addition (floating add - fadds) can take place.
Safe (widening) and unsafe (narrowing) conversions
- As you have seen, the computer can convert one data type to another data type.
- But not all conversions are safe
- The following conversions are safe because the destination type contains all values of the original type:
- C/C++ will automatically perform all (safe and unsafe) conversions as needed:
  - When an operation involves values of 2 different data type, the "lower" type value will be first converted (automatically, i.e., without asking or telling the programmer) to the "higher" data type
  - A "higher" type value can be converted to a "lower" type in an assignment operation
    When you assign a value of a higher type to a variable of a lower type, you may NOT obtain a correct value !!!
    Example:
    int i; short s; i = 9827563; s = i; // <-- This conversion is // DANGEROUS !!!
  - Example Programs: (convert a higher type to a lower type) -- click here
User-defined types and their usage
- In C++, a programmer can define his/her own data types
- Why would you want to define your own type ?
  - In complex programs, the programmer often have to deal with values from different domains
  - Operations involving values from different domains usually do NOT make sense
- Example: banking
  - A bank management program may operate on float values that represent the balance of the customer
  - The program may also operate on integer values that represent the number of days that the account has been inactive (long inactivity will make an account become dormant)
  - It does not make any sense to add or subtract balance and number of days values....
- If the balance is defined as an float and number of days defined as an integer, then the built-in automatic conversion rules in C++ will allow the program to add a balance to the number of days....
- To prevent such absurd operation, the programmer can define one's own user-defined types along with the ALLOWED operations and prevent balance and number of days to be added together.
- Example Programs: (shows user-defined type)
  - This program defines a user-defined type "my_int" without any additional operations.
    C++ will NOT allow this new type to operate with a float variable: click here
    Save program and compile with "CC user-type1.C" - you will see error messages
  - This program defines a user-defined type "my_int" now WITH an additional operations + that tells C++ how to add "my_int" and "float" values.
    C++ will now allow this new type to operate with a float variable: click here
    Save program and compile with "CC user-type2.C" - you will NOT see error messages
Reason for using user-defined data types:
WHY do we need so many INTEGER and FLOAT types ?
1. Space:
  The most common computer application is NOT numerical analysis, but data management - billions of pieces of information are stored.
  If you can represent the values by a smaller integer type, you save a lot of space (and can store more data)
2. Speed:
  In computation (especially heavy numerical math applications), operations on lower type values are faster.
  Reason: higher datatypes use more bits in representation. It takes longer time to move the data around to be operated (need to move the value from memory to CPU and back in operations).
  Example Programs: (shows difference in speed for float and double type)
  - Here is a program that uses floats: click here
  - Here is a program that uses doubles: click here
  - Compile and run with:
    - CC -o speed1 -fast speed-comparison1.C
    - CC -o speed2 -fast speed-comparison2.C
    - time speed1
    - time speed2