Business Rules Extraction (BRX) with Application Analyzer
Business rules are the micro and macro decisions within your in-house developed applications that ensure varied processes such as customer loan rate calculations, payment authorization procedures and a million other important functions, take place accurately, timely and with the correctly determined outcome. Collectively, they are what your business does.
However, these rules & decisions are deeply embedded in your application systems and over many years they have been amended, deleted, reinstated, changed again and so on; changes on top of changes.
The investment that has been placed in these operational processes has been huge – manpower, cost and time.
So when your applications are no longer as agile as they once were and they don’t want to play ball with the latest technology your using, suddenly they become legacy.
To provide peak performance again, the application may need major rework and enhancements, or even a full ground-up rewrite into a modern technology and architecture.
Identifying business rules is an important precursor to building out a micro-services architecture where that built-in operational intelligence and the core calculation routines within the legacy application will need to be componentized for multiple reuse.
For this to happen, you need to be able to understand and document these business rules in an easily consumable way.
The Challenge
Business rules extraction has been around for a number of years, with different methods that attempt to document the “secret sauce” in your applications. These methods include:
Manual extraction - Analysts reading through code, line-by-line, and making intelligent human decisions as to what part of the logic applies to which rules.
Automated extraction - software mines through the decisions out of your applications and presents them to you in pseudo code.
The manual process is extremely slow, labor intensive and costly.
The fully automated process is comparatively fast but it produces too much information – and you’ll end up with many false positives where not all rules and logic in the source code are connected to business rules.
You will still then need a manual process of sorting through that mass of information and trying to piece information together again.
The Solution: A heuristic and semi-automated process
Working on the principle that certain language statement types are more likely to be involved in business rules than others (for example EVALUATE WHEN is more likely compared to a CLOSE) and that the location where the COBOL variables that are used in these statements are declared also makes it more likely (a DCLGEN variable could be more likely than WORKING STORAGE), our BRX system pre-applies customizable weighting values to only relevant areas of code. This enables Analysts to quickly review only those areas of code that important rather than having to sift through ALL lines of code in your programs.
All that is needed is then for the Analyst to enter free-text descriptions to these areas of code with the tool to document new business rules or to attach existing rules to multiple areas of code, across the application.