In computer science, typecasting means to convert a variable's data type to another data type to bypass some restrictions imposed on datatypes. Programming languages that exhibit such a feature of typecasting are C/C++, Java, Applescript, and many other modern or object-oriented programming languages. The expression used in human language is not very complicated to understand, however, understanding typecasting in programming languages is very difficult at first, as it can cause errors in programs if not done right.

Typecasting has risks because typecasting occurs at runtime, as in the compiler just checks to make sure the syntax is correct, but not that the cast will make perfect sense. Therefore, it is necessary to know what type of data a variable is, and to make sure that its target conversion type will make sense and not cause errors. Some examples of safe conversions are floating-point numbers to integers, characters to integers (returns the character's ASCII value), string types to other string types (such as character arrays to an official string data type). Some unsafe casts are floating-point values to characters (since ASCII values are all integers), numerical arrays to character arrays and vice-versa, and contrasting dimension arrays (such as a two dimensional array into a three dimensional array). Often times, a language will have a "safe-cast", in which the compiler/interpreter will try harder to guess at a cast's integrity, and may even have runtime checking so that if a cast throws an error, the cast is ignored. C++ has an implementation of several "safe-casts", such as "static_cast<>", "dynamic_cast<>", "reinterpret_cast<>", and "const_cast<>", all with the target type in the brackets and the source variable in brackets following the keyword.