Skeleton programming is a style of computer programming based on simple high-level program structures and so called dummy code. Program skeletons resemble pseudocode, but allow parsing, compilation and testing of the code. Dummy code is inserted in a program skeleton to simulate processing and avoid compilation error messages. It may involve empty function declarations, or functions that return a correct result only for a simple test case where the expected response of the code is known.

Skeleton programming facilitates a top-down design approach, where a partially functional system with complete high-level structures is designed and coded, and this system is then progressively expanded to fulfill the requirements of the project. Program skeletons are also sometimes used for high-level descriptions of algorithms. A program skeleton may also be utilized as a template that reflects syntax and structures commonly used in a wide class of problems.

Skeleton programs are utilized in the template method design pattern used in object-oriented programming. In object-oriented programming, dummy code corresponds to an abstract method, a method stub or a mock object. In the Java remote method invocation (Java RMI) nomenclature, a stub communicates on the client-side with a skeleton on the server-side.

A class skeleton is an outline of a class that is used in software engineering. It contains a description of the class's roles, and describes the purposes of the variables and methods, but does not implement them. The class is later implemented from the skeleton. The skeleton can also be known as either an interface or an abstract class, with languages that follow a polymorphic paradigm.

Modern software is often complicated due to a host of reasons. This can mean that not just a single programmer can develop it, or that other modules or parts have to be separately imported. The programs can also be too complex on their own, some with multiple methods accessing a single variable at the same time or even generating pixels for displays. Skeleton code is used to assist programmers to develop their code with the fewest errors during the time of compilation.

Skeleton code is most commonly found in parallel programming, but is also applied in other situations, like documentation in programming languages. This helps to simplify the core functionality of a potentially confusing method. It can also be used to allow for a small function within a larger program to operate without full functionality temporarily. This method of programming is easier than writing a complete function, as these skeleton functions do not have to include main functionalities and can instead be hardcoded to use during development. They usually involve syntactically correct code to introduce the method, as well as comments to indicate the operation of the program. This is not always necessary to call a piece of text skeleton code.

Pseudocode is most commonly found when developing the structure of a new piece of software. It is a plain English portrayal of a particular function within a larger system, or can even be a representation of a whole program. Pseudocode is similar to skeleton programming, however deviates in the fact that pseudocode is primarily an informal method of programming. Dummy code is also very similar to this, where code is used simply as a placeholder, or to signify the intended existence of a method in a class or interface.

Computer programmers are extremely dependent on pseudocode, so much so that it has a measurable impact on their psyche. A typical programmer is so conditioned with the idea of writing simplified code in some manner, be it by writing pseudocode or skeleton code, or even just by drawing a diagram, that this has a measurable impact on how well they can write their final implementation. This has been found over a number of applications, with different programmers working in different languages and varied programming paradigms.