Strict programming language
|This article does not cite any references or sources. (October 2006)|
A strict programming language is one in which only strict functions (functions whose parameters must be evaluated completely before they may be called) may be defined by the user. A non-strict programming language allows the user to define non-strict functions, and hence may allow lazy evaluation.
Nearly all programming languages in common use today are strict. Examples include C#, Java, Perl (through version 5), Python, Ruby, Common Lisp, and ML. The best known non-strict languages are Haskell, Miranda, and Clean. Languages whose ordinary functions are strict but which provide a macro system to build non-strict functions include C, C++, and Scheme.
In most non-strict languages the non-strictness extends to data constructors. This allows conceptually infinite data structures (such as the list of all prime numbers) to be manipulated in the same way as ordinary finite data structures. It also allows for the use of very large but finite data structures such as the complete game tree of chess.
Non-strictness has several disadvantages which have prevented widespread adoption:
- Because of the uncertainty regarding if and when expressions will be evaluated, non-strict languages generally must be purely functional to be useful.
- All hardware architectures in common use are optimized for strict languages, so the best compilers for non-strict languages produce slower code than the best compilers for strict languages.
- Space complexity of non-strict programs is difficult to understand and predict.
Strict programming languages are often associated with eager evaluation, and non-strict languages with lazy evaluation, but other evaluation strategies are possible in each case. The terms "eager programming language" and "lazy programming language" are often used as synonyms for "strict programming language" and "non-strict programming language" respectively.