Principles of Commitment Protocols

存在于网络环境中的自主各方（包括公司和个人）必须进行互动。协议构建并简化了这些交互。协议的传统表示指定了合法的操作顺序，但没有规定其内容。因此，传统的表示不能充分支持灵活的交互（例如，处理异常和利用机会）。该项目将通过一种称为承诺机器的推理机制开发更丰富的协议模型，其中根据参与者的承诺为协议状态和行为制定声明性内容。由于其内容表示和基于推理的操作规则，承诺机器有效地编码了协议的灵活版本，从而允许比传统表示更广泛的合法操作。该项目旨在研究协议和承诺机器，获取实际协议（如电子商务）所需的各种推理。具体来说，该项目将1)为应用于开放网络应用的协议制定语义，2)将协议表示为广义承诺机，3)在协议分析和设计工具中实现承诺机，4)通过将该工具应用于流行的和新兴的电子商务协议来评估该方法，5)分类有用的和病态的承诺机。

Autonomous parties - both companies and people - who exist in networked environments must perforce interact. Protocols structure and streamline those interactions. Traditional representations of protocols specify legal sequences of actions but not their content. Thus, traditional representations cannot adequately support flexible interactions (e.g., to handle exceptions and exploit opportunities.) This project will develop a richer model of protocols via an inference mechanism, called commitment machines, in which a declarative content is formulated for protocol states and actions in terms of the participants' commitments. Because of its representation of content and its inferencing-based operational rules, a commitment machine effectively encodes a flexible version of a protocol, thereby allowing a wider variety of legal moves than traditional representations. This project seeks to study protocols and commitment machines, capturing the kinds of reasoning needed in practical protocols, such as for e-commerce. Specifically, this project will 1) formulate a semantics for protocols that applies in open, networked applications, 2) represent protocols as generalized commitment machines, 3) implement commitment machines in a tool for protocol analysis and design, 4) evaluate the approach by applying the tool on popular and emerging e-commerce protocols, and 5) classify useful and pathological kinds of commitment machines.