1 Background

The algebraic approach to software specification originated in the early 1970's. Since then, dozens of algebraic specification languages have been developed--all of them supporting the basic idea of using axioms to specify algebras, but differing in design choices concerning syntax (concrete and abstract) and semantics. The lack of a common framework for algebraic specification and development has discouraged industrial acceptance of the algebraic method, hindered its dissemination, and limited tool applicability.

Why not agree on a common framework? This was the provocative question asked at a WADT meeting in Santa Margherita, 1994. At least the main concepts to be incorporated were thought to be clear--although it was realized that it might not be so easy to agree on a common language to express these concepts.

The following aims and scope were formulated at the start of the Common Framework Initiative, CoFI, in September 1995 [13]:

The aims of CoFI are to provide a common framework:

• by a collaborative effort
• for algebraic specification and development
• attractive to researchers as well as for use in industry
• providing a common specification language with uniform, user-friendly syntax and straightforward semantics
• able to subsume many previous frameworks
• with good documentation and tool support
• free--but protected (cf. GNU)

• specification of functional requirements
• formal development and verification of software
• relation of specifications to informal requirements and implemented code
• prototyping, theorem-proving
• libraries, reuse, evolution
• tool interoperability

• a critical selection of known constructs
• expressive, simple, pragmatic
• for specifying requirements and design for conventional software packages
• restrictions to sublanguages
• extensions to higher-order, state-based, concurrent, ...