Contracts in Programming and in Enterprise Systems

In the first part of this report, we present an extension of the programmingby-contract (PBC) paradigm to a concurrent and distributed environment. Classical PBC is characterized by absolute conformance of code to its specification, assigning blame in case of failures, and a hierarchical, cooperative decomposition model – none of which extend naturally to a distributed environment with multiple administrative peers. We therefore propose a more nuanced contract model based on quantifiable performance of implementations; assuming responsibility for success, and a fundamentally adversarial model of system integration, where each component provider is optimizing its behavior locally, with respect to potentially conflicting demands. This model gives rise to a game-theoretic formulation of contract-governed process interactions. We introduce an abstract model of communication, which assumes no common model of computation at the peers in the distributed system. An implementation consists of a set of processes and possible delegation, which defines a strategy for the set of games specified by the set of committed contracts. Verification of correctness is introduced via contract portfolio conformance, which generalizes the idea of Hoare triple validity (underlying classical PBC). Contract portfolio conformance corresponds to a mix between partialand total correctness, called timed total correctness, and we show that portfolio conformance supports compositional reasoning. In the second part of this report, we have conducted a survey on both theoretical and practical approaches to contract formalization. By contract formalization we mean representation of (business) contracts in computer systems, to encompass automatic validation, execution, and analysis of contracts. These activities are collectively referred to as contract lifecycle management (CLM), and the typical aspects of a CLM system include (1) contract creation, (2) contract negotiation, (3) contract approval, (4) contract execution, and (5) contract analysis. Our survey concludes with a summary of the pros and cons of the covered approaches, and a perspective towards future work.