After identification of weaknesses (through assessment or application deployment), decision was taken to improve the Verification & Validation (V&V) process and especially for high interactive, client-server applications developed with Object Oriented techniques. The target is to get at least the same level of maturity as for requirements and design process for which a huge effort has been invested in the past. This means:
decrease the number of remaining bugs when the software is delivered, that should be reduced by 20%,
avoid the risk of regressing from one version to the next one,
optimize the effort dedicated to V & V (20% less is objective) and reduce time to deliver the product with the same level of quality.
This improvement will be based on three aspects : the definition of the Verification & Validation process itself, the technology to support and monitor the process and the training to facilitate teams' adoption. AVAL should address these three aspects for which measures will be installed in order to track the results. The AVAL software process improvement experiment (PIE) is divided into 3 phases:
a status and overall framework definition phase: set quantitative improvement goals for the V & V process, by implementing ami method, both for defining improvement goals and for measurement plan production.
a set up and formalizing phase: through strategies and scenarios developed from the project specification, we will prepare how practitioners of the baseline project will implement standard V&V techniques and tools on the baseline project.
a final synthesis phase : cost/efficiency analysis, lessons learned, introduction of successful practices in other 0BJECTIF TECHNOLOGIE projects.
The baseline application, based on a client-server architecture, manipulates objects with geographical and technical attributes and daily information. This product has been developed using current technologies, especially object-oriented method for specification and design. 0BJECTIF TECHNOLOGIE employs 30 high-level engineers, and up to 10 of them were involved in this project.
EXPECTED IMPACT AND EXPERIENCE
We first expect short term lessons from our improvement effort: measurement framework for progress monitoring and cost/benefit analysis and V&V process definition according to standard recommendations. Procedures to plan, measure, document and control the validation process and results will be established. 0BJECTIF TECHNOLOGIE must - at the end of the current Software Process Improvement experimentation - be able to repeat and control a well documented validation process. The commercial impact of such improvement will be in several directions:
- increased confidence in the baseline project and Verification & Validation process and culture,
- opportunity to show and reuse our know-how and practical results (ami measurement collection),
- measures related to tool implementation and usage costs, and suggested tool functions evolution and improvements.