Today, business rules are the focus of various efforts in industry and academia. The recent work on e-business, web services, the Semantic Web, rights description and management, privacy, security and trust, is reinforcing the idea of a global approach to deal with the management, interchange and sharing of policies, regulations and business rules between various systems on the Web and in distributed environments.
The RuleML Initiative is proposing the formation of a Policy RuleML Technical Committee and the use of RuleML as a way to interoperate between different Policy systems.
Please send any comments you may have to Said Tabet and Bruce Spencer. The discussion should take place on the RuleML Mailing List.
Today, business rules are the focus of various efforts in industry and academia. The recent work on e-business, web services, the Semantic Web, rights description and management, privacy, security and trust, is reinforcing the idea of a global approach to deal with the management, interchange and sharing of policies, regulations and business rules between various systems on the Web and in distributed environments.
The RuleML Initiative is proposing the formation of a Policy RuleML Technical Committee and the use of RuleML as a way to interoperate between different Policy systems.
The Policy RuleML TC will need to define a comprehensive prioritized list of focus areas.
The overall motivation for the proposed TC is that to employ RuleML as a semantic interoperation vehicle for heterogeneous policy languages, standards, protocols, and mechanisms, both currently existing and those developed in the near future. RuleML will provide intermediate markup syntax, with associated deep knowledge representation semantics, for interchange between those languages, standards and mechanisms. The RuleML TC will also develop translation and implementation tools for interchange between policy languages/standards/protocols/mechanisms that are already XML-based. This can, for instance, be achieved using XSL transformations (XSLT), e.g., to translate into and then out of RuleML.
The RuleML Initiative has been approached by a number of major players in the financial services area. We will discuss the relevance of financial services as an initial focus. As another application, we will look into manufacturing (supply chain and policies).
The initial goal of the TC is to investigate scenarios of usage of RuleML as an interchange vehicle for policy languages, and to develop standards for such interchange suitable for a useful set of such scenario areas. Industry Vertical: Financial Services Financial services represent a very good initial prospect. Other industry-vertical areas will also be explored to some extent in parallel, e.g., electronics/computer or automotive manufacturing supply chain. Financial services give us a unique opportunity to explore many areas and domain scenarios where policies are used. The Securities Exchange Commission (SEC) defines and produces many rules and regulations for trading and compliance. These can be of various complexities and usually implemented in many different policy systems. The results of the RuleML Policy TC can be directly applied to encode such policies from SEC and other government and regulation agencies. Furthermore, the RuleML Policy language will offer ways to translate and integrate between disparate Policy systems. The RuleML Policy TC is anticipated to work and collaborate with other policy and regulations efforts on this specific domain as well as with other standards committees. The RuleML Policy TC will deliver an interoperation vehicle between other standards' policy specifications.
Initially, we plan to use the current RuleML 0.85, with it's Object Oriented capabilities, to capture policies that can be encoded as derivation, constrained or reaction rules. For the long-term, we plan to extend RuleML towards incorporating deontic expressive features such as logics to capture rights, obligations and empowerments as aspects of policy rules. With this new concept, we will define a new category or class of rules in the RuleML hierarchy. Here are some examples of deontic rules:
There is a particular form of permission rule assigning exclusive rights having the following form:
- Only an agent of type A has the right to perform actions of type a.
Examples:
- Branch managers are permitted to drive cars to the service station
- Office clerks are prohibited to drive cars to the service station
Duty assignment rules have the following form:
- Agents of type A have the duty to fulfill commitments (or to react to events, or to monitor claims) of a certain type.
Examples:
- Customer service clerks have the duty to react to reservation requests
- Drivers have the duty to fulfill commitments to provide suitable cars to customers at specified locations
- Financial accounting clerks have the duty to monitor claims against customers to pay their invoices on time
The general structure of an institutional power assignment rule, where a is a type of ritualized action for creating institutional facts of type F, is:
- If agent i is of type A, and specified pre-conditions for the arguments x of a hold at time t, and i performs a(x) at time t, then F(x) holds at the immediate successor time point t' of t.
As pointed out in [8, 9], the concept of institutional power has to be distinguished from the concept of permission. An agent may be empowered to create some institutional fact (or to perform some institutional action) without having the permission to do so.
Examples of empowerment rules are:
- A priest is empowered to marry a couple.
- A department manager may be empowered to purchase office equipment, but, under certain circumstances, he may not have the permission to do so (e.g., there could be a temporary budget constraint).
We propose, as a starting point, the following set of usage scenario as candidates for investigation. We give these below in roughly descending order of priority to investigate: