With billions of dollars changing hands every day, financial trading systems demand extremely high accuracy and reliability. So, how do you improve test process performance in the areas of time to market and efficiency and at the same time reduce failures? Over the last three years, using process and project measurement data as a guide, SIAC has focused on doing exactly that. Steve Boycan highlights the key elements of the process changes that have led to SIAC's current performance: the use of a rigorous requirements engineering process; controlled parallel and iterative work flows; changes to the level of abstraction in test documentation; emphasis on test planning, analysis, and design; causal analysis; and improving the test team's skills.
Ignore or downplay non-functional system testing at your peril. A thorough, well-executed nonfunctional test plan discovers software defects usually overlooked with functional testing. From security, scalability, and usability issues to legal 508 accessibility, recovery processes and more, testing non-functional requirements can mean the difference between success and failure. Julian Harty describes key non-functional test practices and demonstrates the value of each. From the testers’ point of view, functional testing will become less valued as outsourcing, automation, and improved programming techniques combine to make functional testing less a factor. As the risks and costs of failures increase, the need for non-functional testing will continue to grow, and test engineers who have system testing experience will be more valued.
Evaluation of graphical defect trend data can dramatically increase your ability to predict current project quality, schedule milestone compliance, and provide historical data for proper test and development scheduling of later revisions. Jim Olsen will explore some of the complexities in analyzing graphic defect trending in this presentation (winner of the Best Presentation award for ASM'99). Learn ways to determine how much time establishes a trend, when the appropriate time to start taking data occurs, what type of data to track, and how to estimate the amplitude of defect oscillations at the end of the product cycle.