US Technology, a provider of IT services and Business Process Outsourcing (BPO) solutions for Global 2000 enterprises, today announced the launch of its Logic Modeling service to help enterprises detect software defects before they are coded into applications.
Current statistics indicate that more than 50 percent of software defects originate in untested requirements. With the new Logic Modeling service, US Technology offers enterprises the ability to test business requirements and reduce the number of defects by more than half—before the first line of code is written. Once built, logic models are then used to generate high coverage test cases that approach 100 percent functional coverage. This yields extraordinarily low residual defect rates; in many cases zero defects are found in delivered software products.
Limits of Traditional Testing
Traditional testing is more time-consuming and therefore costly when compared to logic model-driven testing. Generally, traditional analysis takes 50 percent longer and produces twice the number of test cases with half the functional coverage. Traditional testing focuses on finding defects in delivered code and:
Does not provide metrics on project health until well past 80 percent of the software development lifecycle
Does not attempt to find defects in requirements, thus does not prevent defects
Cannot effectively measure requirements coverage
Makes it difficult to revise/maintain test suites when the application changes
Is less effective when designed by subject matter novices
Is prone to unnecessary redundancy as analysts attempt to cover all or most input combinations
Achieves at best 60 percent of the functional coverage of logic modeling
Stops to comply with project schedules
How Logic Modeling Surpasses Traditional Testing
When two percent of the defects in an application cause 1,000 times more failures than other defects, thorough testing becomes imperative to minimizing risk. Effectively addressing these limitations—which logic modeling does—can make a big difference to businesses.
“The top reasons why software projects fail are a result of incomplete and/or changing requirements and specifications and the lack of user input. The most sophisticated means of identifying incomplete requirements involves constructing software designs and then completing walkthroughs and testing. However, none of these methods allow organizations to quantify – and thus measure – test coverage,” said Bob Dutile, general manager of strategy and solutions, US Technology. “While code coverage tools can help increase the depth and breadth of testing, they do not measure compliance to what was ordered in the requirements. Traditional test design, when employed vigorously, usually defines twice the number of test cases with only 50 percent of the coverage when compared to logic modeling. Logic modeling reduces the failure rate of software projects.”
Clear and complete requirements are necessary to successfully build a logic model. When requirements fail to meet this standard – which most often is the case – questions (called ambiguities) are raised. Ambiguities identify opportunities to improve requirements documents in order to meet this standard. As an attempt is made to resolve ambiguities, design flaws can emerge that would otherwise manifest as defects if they were not detected at this early stage. Ambiguity resolutions are then incorporated into the requirements document, which makes the development and testing team more productive as scrap and rework are eliminated.
The Benefits of Logic Modeling
Because logic modeling focuses first on requirements, domain novices can be more effective than experienced traditional test analysts. While traditional analysts rely on their subject matter expertise, domain novices must rely on what is documented. This serves to improve documentation dramatically after ambiguity resolution. Additionally, subject matter novices are apt to raise ambiguities that would not occur to the experts.
Originally, service-oriented architecture (SOA) concepts arose in response to demands for better ways to cost-effectively integrate large-scale...Learn More