1. An Overview of Business Rules in the Semantic Web |
Home | Business Rules | Rule Based Systems | Rule Engines | Semantic Web | Links | Under the Hood |
The Basic Idea
Humble Beginnings
From Powerhouse of Business Transaction Processing to ...
.. Personal Knowledge Server ?
... and the Semantic Web ? |
Note: This is an old version. See the updated Overview of Business Rules on the Drupal Home site. Eventually the most recent version will be republished back to this page as static HTML. From humble beginnings in the 1990s, business rules technology evolved into commercial-grade platforms capable of driving big and complex systems for Fortune 500 companies. Some technical infrastructures cost millions of dollars to implement and deploy. Business rules technology has become essential to the functioning of core business processes. Every day, many millions of business transactions pass through rule-based customer ordering systems, ATM and credit card verification systems and an endless variety of business applications. Rules are major components of the core business processes. I think that a technological threshold has been reached and that rules technology is no longer the the sole domain of large-scale business processes. It can be made easy enough and powerful enough to support simple problem solving by ordinary, non-technical individuals. With the advent of inexpensive web hosting services and web-logging technology, a combination of workstation and server side functions can implement a sort of personal knowledge server. The critical factor is to create an the interface that is simply enough for everyday, non-techie to understand and use. Furthermore, I believe that the technology can adapted to a vision of the Semantic Web. However, developing a rule systems technology for a Semantic Web may require significant modifications and extensions to existing concepts and tools. In fact, it is not clear what the term 'Semantic Web' means and the challenge of resolving competing visions and definitions is taken up in a later section covering the Semantic Web. |
|
Classical Definition of Business Rules
Declarative
Atomic
Distinct
Natural Language
Business Oriented
Business Owned |
That invaluable tool, the Wikipedia, has the cleanest and concise definition of the term 'business rule' I've encountered. The Wikipedia defines business rule as: Declarative: A business rule is a statement of truth about an organization. It is an attempt to describe the operations of an organization, not an attempt to prescribe how an organization should operate. This is why business rules are said to be discovered or observed and not created. There are many definitions of the term "business rule". Some focus on the high-level idea of business "policies", delineating how the company will do business. Business policies are distinct from procedural knowledge about business operations and often seem to be arbitrary, such as a policy to send a customer debt to a collections agency after 60 days. Why not 75 days or 45 days ? It is purely a policy. |
|
Other Definitions of Business Rules
The Process View
"Localized" Business Rule
Bottom Up Approach to Business Rules |
Haley has an interesting take on the definition of business rule.
Note that the definition treats procedures as a legitimate expression of knowledge about a business process of some sort. In the field of Business Process Modeling, the sequence of operations in a business processes can become something like a business rule in itself. It is an import distinction and the correct view, a process-based or rule-based point of view, depends largely on the category and level of interaction of the business rule under consideration. Business rules tend to appear at the enterprise level. Processes tend to use business rules in a 'localized' way, for instance by geographic location or by market segment. A certain set of business rules can play a role in some aspect or 'facet' of a particular business process. They might even be categorized along several dimensions simultaneously, for instance by geographic location and by market segment at the same time. The demands of large globalized business applications can generate massive numbers of "localized" rules by combinatorial explosion of 'sub-types', sometimes many thousands of individual categories of business rules. In those situations, a bottom-up, process-based approach to business rules analysis can be very effective. |
|
A Tentative Definition of Rules for a Semantic Web
Is there such a thing as Web Rules ?
Owned by Originators
Knowledge Focused
Unique
Unambiguous
Declarative
Well-formed |
Note first that there is a big difference between The Semantic Web to A Semantic Web. In this version of the Semantic Web, knowledge shared by individuals. A web would be owned by its participants or subscribers rather the general public, something to be actively joined, perhaps more like the basic model of a Wiki than an openly accessible public site. If there were a version of business rules called "semantic web rules", what would its definition look like ? Certainly, the definition should include a very basic set of characteristics pertaining to the Semantic Web. It would also need to stay as close as possible to the classic definition of business rules or there is a danger that the entire meaning and "discipline" of the methodology could be lost.
The last three requirements for business rules may raise an important issue for using business rule technology in the Semantic Web. Note that the last three characteristics listed above ( that is, clear, declarative and well-formed ) may restrict the form that web rules can take. To some extent, they may also limit the usefulness of classic business rules as a vehicle for defining things in the relatively unstructured environment of the Semantic Web. |
|
Are the Definitions Realistic for Real World Problems ?
Dealing with Incomplete and Contradictory Information
Well-Formed or Just Adequately-Formed ? |
The central question is, how realistic are these requirements outside of the limited, well-defined domain of business transactions ? In well-defined domains such as making bank deposits or processing customer orders, a rule can be either "completely true or completely false", but how often is that true in real life. Another question is, to what extent "clear" and "well-formed" rule can be utilized by individuals in confronting the incomplete and contradictory information of every day life, such as those requiring someone to spend three hours on the phone attempting to correct the consequences of erratic business policies and information systems. Some web rules will not be "well-formed" enough to be expressed within the confines of a restricted and specialized computer-friendly rules language. They may be difficult to express and not every rule 'owner' in the web community will be interested in learning a new language to express their rules. More likely, they will think in terms of everyday language and they may not be willing or able to translate their knowledge into a formal rule language. What we are seeking is rules that are not necessarily "well-formed" but are "adequately-formed" enough for the more rough and tumble world of the Web, where the quality of the rules themselves are likely to be less controlled and less reliable than in a classical big company environment. |
|
Real
World Rules for the Semantic Web
Let's say that we've discovered some "web rules". Then what ? |
The
formulation of rules is not an exact science for either a business analyst or
a web surfer . Clearly, there are two distinct types of a rules at work,
well-formed rules which can be exactly stated 'before the fact' in a unique and
unambiguous form and adequately-formed rules discovered 'after the fact'
by struggling
with more or less real world information problems on the Web.
The most immediate goal is to capture knowledge about the Semantic Web in whatever form it present itself, even if encountered in a contradictory form. It is assumed, in a tentative way, that discovery of web rules will be very similar to rule discovery in a business context, even if the content of the rules is very different. The basic "laws" of analysis and discovery ( via logic induction, for example ) will be be valid within a semantic web the same as any classical business rule domain. In fact, the relatively simply, people-oriented user interfaces developed by 20-year-old business rules technology may even have something to teach 20-year-old Semantic Web developers. ;-) After discovering relevant business rules, the rule analyst need to transform their natural language descriptions into "well-formed" statements that can be used to satisfy the demands of a knowledge-intensive task. And that requires a business rules methodology. |
|
<<< |