This paper examines the past, present, and future of the programming languages course as reflected by its textbooks.

/pdf/the-evolution-of-the-programming-languages-course-532ofthfna.pdf

The evolution of the programming languages course

Traditionally beginning undergraduate computer science students are taught a high level programming language. Since the early 1970's Pascal has become the de facto standard introductory programming language. With the creation of Ada, its standardization, and the availability of validated compilers, there is considerable discussion concerning the introduction of Ada into the computer science curriculum. This paper discusses the approach taken by the Computer Science Department at Harvey Mudd College.

On the teaching of Ada in an undergraduate computer science curriculum

This thesis presents a type system which supports the strong static type checking of programs developed in an applicative subset of the Common Lisp language. The Lisp subset is augmented with a guard construct for function definitions, which allows type restrictions to be placed on the arguments. Guards support the analysis and safe use of partial functions, like CAR, which are well-defined only for arguments of a certain type.
A language of type descriptors characterizes the type domain. Descriptors are composed into function signatures which characterize the guard and which map various combinations of actual parameter types to possible result types. From a base of signatures for a small collection of primitive functions, the system infers signatures for newly submitted functions.
The system includes a type inference algorithm which handles constructs beyond the constraints of ML-style systems. Most notable are the free use of CONS to construct objects of undeclared type and the use of IF forms whose two result components have unrelated types, resulting in ad hoc polymorphism. Accordingly, the type descriptor language accommodates disjunction, unrestricted CONS, recursive type forms, and ad hoc polymorphic function signatures. As with traditional type inference systems, unification is a central algorithm, but the richness of our type language complicates many component algorithms, including unification. Special treatment is given to recognizer functions, which are predicates determining whether an object is of a particular type. Type inference in this setting is undecidable, so the algorithm is heuristic and is not complete.
The semantics of the system are in terms of a function which determines whether values satisfy descriptors with respect to a binding of type variables. The soundness of each signature emitted by the system is validated by a signature checker, whose properties are formally specified with respect to the formal semantics and proven to hold. The checker algorithm is substantially simpler than the inference algorithm, as it need not perform operations such as discovering closed recursive forms. Thus, its proof is both more straightforward to construct and easier to validate than a direct proof of the inference algorithm would be.

Strong static type checking for functional common lisp

The learning of programming can take manifold starting points, e.g. via friends or family, a school course, a compulsory requirement in vocational education (e.g. industrial robotics) or even in an academic setting (e.g. formalization and execution of mathematical algorithms in numerical mathematics). Within these approaches, context-specific programming languages are used, which might support different paradigms. In this paper, a proposal for a language and paradigm overarching competency structure model is developed from theoretical considerations. By analyzing different programming languages, textbooks and papers three content dimensions are derived. Principles and the application of high-Level paradigms build the first dimension. The second dimension elements represents paradigm and language independent concepts of program components, data types, basic data structures and algorithmic paradigms. The last content dimension language is defined by syntax, semantics, standard library and the build/compile and run process of programming languages.

Competency structure model for programming for the transition from school to university

A limitation of current selection-entry syntax-directed editors is that a particular editor is limited to manipulating programs in one particular programming language. A primary goal of the IMEGS research project was to construct a selection-entry syntax-directed editor which can manipulate programs in a wide variety of languages. An IMEGS prototype has been developed which contains a multi-language editor.

A multi-language syntax-directed editor

http://ieeexplore.ieee.org/iel5/5/31357/01457716.pdf

M.B. Bilodeau

Papers

Fundamentals of programming languages, Second edition