Columns

Rules move to real world

Imagine a world without some of your favorite people -- such as life insurance reps, mortgage bankers and car sales associates. Sad picture, isn't it? Right... Thanks to changes in the way we look at enterprise systems and build software business applications, these somewhat annoying people may soon be merely a memory of yesteryear. Many companies, especially in industries such as finance and insurance, are employing rules-based software systems in an effort to maintain a competitive edge in fast-moving markets while cutting costs and meeting customer needs. Customers and clients will be able to interact efficiently with rules-based software systems to accurately fill out insurance forms, mortgage applications and even select a new automobile. This is being done to a certain degree around the U.S., but expect it to catch on as corporations look to better service customers and easily adapt software systems to meet changing business needs.

What are rules-based systems anyway? They actually have their roots in the older Artificial Intelligence (AI) methodologies that were somewhat popular a few years back. Business rules were embedded within the logic of an application. When business policies changed, it was difficult to reflect those changes in the application that implemented them. It often took a programmer several months to implement a business-driven change within an application. Thus, the limitations of procedural coding.

"We have been doing procedural development for years," said Sandra A. Taylor, director of research at SPG Analyst Services in Natick, Mass. "We keep trying to make it easier, faster, more visual. Vendors are providing wizards and templates and frameworks, but these are still procedural paradigms as far as applications and code are concerned.

"Even Java applets and JavaBeans, once you peel away all of the hype, are still basically a 3GL," Taylor said.

Basically it has been years since IT has changed its thinking about the structure of programs, Taylor said, but everything else has changed. Hardware is faster, communications mechanisms are better, people are beginning to understand the potential of object technology, and more and more pieces of software are becoming componentized.

With the slow but steady maturation of object-oriented technologies, component-based development and the arrival of the Internet as a universal connectivity mechanism, businesses are now utilizing these methodologies to achieve rapid adaptation of software to meet market demands. Adaptive applications are becoming the cost of doing business today. And rules-based systems offer an alternative to procedural programming techniques.

There are two primary varieties of business rules technologies: data-oriented and object-oriented. An example of a data-oriented rules product is Vision Jade from Vision Software of Oakland, Calif. Vision Jade is a declarative rules environment designed to automate the development and deployment of multitier, 100% Java applications in enterprise environments. Instead of writing low-level Java code, developers can use Vision Jade to create business rules that define the application. The Jade software will automatically calculate the dependencies between application rules and transform them into Java code. The business rules are then represented with simple, declarative statements instead of procedural logic. This approach obviously saves time in actual application delivery while lowering costs and eliminating logic errors. A single declarative statement in Jade could replace up to 200 lines of Java code, according to the vendor. The product is data-oriented in that the rules operate on relational data via SQL. Vision Jade applications may be deployed over standard application servers and will support all major data sources. This is important for companies looking to extend new and legacy applications throughout the enterprise while allowing them to easily adapt to changes in business requirements.

An object-oriented rules-based approach is demonstrated in newer products from Neuron Data of Mountain View, Calif. Neuron Data as a company has quickly "morphed" over the past several months, dropping its ties to the crowded IDE market space and focusing its efforts on component-based rules development. Its latest product, Elements 2.0, is an open architecture, component framework which provides a rules-driven development tool suite. The product is designed to work with C, C++ and Java languages, and uses a business rules engine to process rules. The rules are kept separate from the application so they can be shared and changed even while the application is running. The product is constructed with an open architecture so users may select the middleware infrastructure of their choice when linking systems and components.

Applications that most easily benefit from rules-based development are either those that use a complex chain of logic or information to accomplish their ultimate purpose, or those that require rapid updates because of frequent changes in business policies and procedures. Rules-based systems can offer enterprise application developers several immediate benefits in the development cycle: adaptability, time efficiency, flexibility, and last but not least, cost savings. The latter category is often realized sooner than in traditional development environments due to the rules-based approach lending itself to lower training costs, the corresponding faster development cycles, and perhaps most importantly, ongoing and efficient application maintenance.

An example of how this technology works in the real world can be seen at a large insurance company based in New York City. Its member companies write property, casualty, marine, life and financial services insurance policies in approximately 130 countries. The company recently decided to automate its underwriting processes via a rules-driven application. Previously, all underwriting tasks were done manually by insurance underwriters. Now, using rules-based automation software, the company has eliminated manual underwriting tasks and provided the ability to quickly adapt to the ever-changing underwriting rules. Using this rules-based software, which also had to integrate easily with existing programs in-house, the insurance giant has empowered users to adjust their business processes and rules to meet their ever-changing requirements.

SPG Analyst Services expects to see growth in rules-based systems within the medical community. There are several examples of rules-based software systems being utilized for patient intake procedures and preliminary, routine diagnostics. This type of software can also be used in the pharmacy environment to help refine prescription information. As this type of technology matures, the possibilities are limitless.

The growth and success of products such as Vision Jade and Neuron Data's Elements validate SPG Analyst Services' prediction that 1997 was the year that business process rules automation technology would start to come into its own as a key management and information technology trend. Over the past year, an increasing number of information systems (I/S) shops realized the ability to quickly and correctly specify and modify business rules was more important than rapid GUI design and the execution efficiency of the server engine. Although business rules development has existed for decades, the acceptance of the technology into the main stream is fairly recent.

It was only a few years ago that the GUI/RAD phenomena dramatically changed the way developers interfaced with a computer, said SPG Analyst Services' Taylor. With that, the industry radically changed the way it approached the development of front ends. First there were GUIs, then there were browsers. The front end is pretty well nailed down, and the GUI part is easy. The hard part now is how to develop and maintain the business rules.

"Maintenance is probably the harder issue," Taylor said. "There is tremendous energy poured into the initial phases of any development project, but enthusiasm wanes noticeably when it moves over to maintenance mode." She added that with the upswing in interest to develop rules-based systems, there is emerging a corresponding increase in maintenance, management and updating of those rules that drive the business.

Even with these advances, the majority of information technology (IT) is still "hooked into a procedural development paradigm," said Taylor. This is where companies have a chance to make a difference. They can offer a different way to develop business rules. When selecting a rules-based tool, Taylor recommended users consider the following:

3The level of abstraction;

3Simplistic rule specification: declarative and non-procedural;

3Applications can interact dynamically with rules repository;

3Rule changes are automatically reflected in the business process; and

3Rules communicate via messaging.

Taylor's ultimate advice to the IT shop is one of context. Individual companies need to evaluate the technology, not only for its ability to represent the rules of the particular business, but also in light of the existing corporate culture. If there are a majority of C++ proponents on staff, an organization will probably have difficulty moving to a rules-based environment. It is simply a matter of cultural reluctance.

"One thing is clear. The world of corporate development has been turned upside down by emerging technology and the rate at which technology is evolving. Despite this technology frenzy, organizations must stay open to new ideas and evaluate those ideas on their own merit. People that suffered through the death throes of CASE may be too gun shy to look at the new modeling tools," Taylor said. "Organizations that saw the rise and fall of AI may be skeptical with regards to the new generation of rule-based offerings. They should, however, keep their options open, because the companies that can adapt quickly to market changes will be the ones that will gain and maintain a competitive advantage."