Value (computer science)
 |
In computer science, a value is an interpretation of a sequence of bits according to some data type. It is possible for the same sequence of bits to have different values, depending on the type used to interpret its meaning. For instance, the value could be an integer or floating point value, or a string.
Some kinds of value are common to most programming languages (e.g., various kinds of number representations), while others are less commonly supported (e.g., Pascal supports a set type).
In declarative (high-level) languages, values have to be referentially transparent. This means that the resulting value is independent of the fact in which location a (sub-)expression needed to compute the value is stored. Only the contents of the location (the bits, whether they are 1 or 0) and their interpretation are significant. A value is the normal form of an expression, which means that an expression that is a value can not be reduced to a simpler expression. For example, the value of the expression "1 + 2" is "3", which cannot any longer be reduced to a simpler expression.
In C: lvalues and rvalues
Some languages use the idea of lvalues and rvalues. Lvalues are values that have addresses being programmatically accessible to the running program (e.g., via some address-of–operator like "&" in C++), meaning that they are variables or dereferenced references to a certain memory location. Rvalues can be lvalues (see below.) or non-lvalues—a term only used to distinguish from lvalues. In C, the term lvalue originally meant something that could be assigned to (coming from left-value, indicating it was on the left side of the assignment operator), but since 'const' was added to the language, this now is termed a 'modifiable lvalue'.
In C++0x a special semantic-glyph "&&" exists, to denote the use/access of the expression's address for the compiler only, i.e. the address cannot be retrieved using the "&"–address-of–operator during the run-time of the program (see the use of move constructors). This type of reference can be applied to all r-values including non-lvalues as well as lvalues. Some processors provide one or more instructions which take an "immediate value", sometimes referred to as "immediate" for short. An immediate value is stored as part of the instruction which employs it, usually to load into, add to, or subtract from, a register. The other parts of the instruction are the opcode, and destination. The latter may be implicit. (A non-immediate value may reside in a register, or be stored elsewhere in memory, requiring the instruction to contain a direct or indirect address [e.g. index register address] to the value.)
The lvalue expression designates (refers to) an object. A non-modifiable lvalue is addressable, but not assignable. A modifiable lvalue allows the designated object to be changed as well as examined. An rvalue is any expression, a non-lvalue is any expression that is not an lvalue. One example is an "immediate value" (look below) and consequently not addressable. Another example is the C expression (4 + 9). When executed, the computer generates an integer value of 13, but because the program has not explicitly designated where in the computer this 13 is stored, the expression is an rvalue. On the other hand, if a C program declares a variable x and assigns the value of 13 to x, then the expression (x) has a value of 13 and is an lvalue.
The notion of lvalues and rvalues was introduced by CPL.
In assembly language
A value can be virtually any kind of data by a given data type, for instance a string, a digit, a single letter.
Some processors support more than one size of immediate data, e.g. 8 or 16 bit, employing a unique opcode and mnemonic for each instruction variant. If a programmer supplies a data value that will not fit, the assembler issues an "Out of range" error message. Most assemblers allow an immediate value to be expressed as ASCII, decimal, hexadecimal, octal, or binary data. Thus, the ASCII character 'A' is the same as 65 or 0x41. The byte order of strings may differ between processors, depending on the assembler and computer architecture.