Dictionary.net

An evaluation strategy under which anexpression is evaluated by repeatedly evaluating its leftmostinnermost redex. This means that a function's arguments areevaluated before the function is applied. This method willnot terminate if a function is given a non-terminatingexpression as an argument even if the function is not strictin that argument. Also known as call-by-value since thevalues of arguments are passed rather than their names. Thisis the evaluation strategy used by ML, Scheme, Hope andmost procedural languages such as C and Pascal.See also normal order reduction, parallel reduction.