Quality and Measurement

As any engineering discipline, software specification, development, operation, and maintenance require the application of disciplined and quantifiable approaches. This implies a systematic quality control of all the products as well as of the engineering processes.

As any management discipline, software management also implies systematic quality control of products and processes. Moreover, it implies the integration of quality management of heterogeneous systems within the global quality management process of the organization.

Software quality relies on four levels of concern: (1) quality management, where criteria are defined at the business level according to a customer-centric approach, (2) software process, emphasizing software process modelling, evaluation and improvement, (3) software product, where the focus is put on the definition of quality indicators for the different software products and their relationships, (4) software measurement, where metrics and measurement methods are defined and can lead to improvement.

 

Themes

  • Quality management. At the management level, quality criteria are defined on the basis of customer specification limits. Control of those criteria is performed in order to guarantee the stability of business processes. When processes are stable, opportunities for improvement can be studied through typical quality tools: Pareto analysis, cause-effect diagrams, why-why analysis, etc.   Expertise in software certification has been developed, notably in aeronautics (DoD-178B).
  • Software process assessment & improvement. At the process level, a software process model dedicated to Small & Medium Enterprises has been developed as a specialization of classical software process models like CMMi and ISO-15504.
  • Software product quality models & measurement. At the product and the measurement levels, a generic framework (MoCQA) has been built, providing the derivation of specialized quality models from a multi-standard metamodel.  On this basis, several specific quality models have been designed for different particular contexts when standards models are insufficient, notably the QuaLOSS model for open source systems, with a focus on maintainability and robustness, and the Doca model, for documentation assessment focussing on availability and structuration. A Bayesian model for website evaluation is being investigated with the University of Montreal.
  • Database schema quality evaluation and improvement. A new method based on the identification of defect patterns in database schemas is being developed. It provides a systematic way to measure the quality of a schema precisely and to suggest improvements through transformational techniques.

 

Scientific results

PReCISE has developed (1) OWPL, a model for SME software processes, (2) MoCQA, a generic quality framework to support the generation of tailored quality models, (3) DOCA, a quality model for documentation, (4) QuaLOSS, a quality model for Open Source systems, (5) a tool-supported generic framework and methodology for database schema evaluation and improvement.

 

Industrial results

The OWPL model has been used for several years by dozens of companies. It is distributed by the Cetic (www.cetic.be)  and is used as a major source for the development of ISO standard ISO-29110. The Doca Model is used to evaluate open-source projects while the MoCQA model is used to assist aeronautic industrial partners in certification processes.

 

Resources

  •  Selected Publications
  • Benoît VANDEROSE, Naji HABRA, Flora KAMSEU, Towards a Model-Centric Quality Assessment, in Proceedings of the 20th International Workshop on Software Measurement (IWSM 2010)
  • Raimundas MATULEVICIUS, Naji HABRA, Flora KAMSEU. Validity of the Documentation Availability Model: Experimental Definition of Quality Interpretation, in Proceedings of The 22nd International Conference on Advanced Information Systems Engineering (CAiSE'10)
  • Stéphane VAUCHER, Samuel BOCLAINVILLE, Houari SAHRAOUI, Naji HABRA. Recommending Improvements to Web Applications using Quality-Driven Heuristic Search, in Proceedings of the 10th International Conference on Web Information Systems Engineering, Poznan (Poland), Octobre 2009
  • Naji HABRA, Alain ABRAN, Miguel-Ange LOPEZ MARTIN, Asma SELLAMI. A Framework for the Design and Verification of Software Measurement Methods, in Journal of Systems & Software, volume 81, issue 5, pp. 633-648.
  • Naji HABRA, Simon ALEXANDRE, Jean-Marc DESHARNAIS, Claude LAPORTE, Alain RENAULT. Initiating software process improvement in very small enterprises: Experience with a light assessment tool, in Information & Software Technology , volume 50, issue 7-8, pp. 763-771
  • Methodogies : Quality Models 
  • OWPL : Sofware Process Model for Small & Medium Enterprises
  • QuaLOSS: Quality Model for Open-Source Systems
  • MoCQA : Genric Qulality Approach (Meta-Model + Method + Tool)
  • DOCA : Quality Model for Software Documentation
  • Database schema quality evaluation and improvement framework

Products and services

  • OWPL : Sofware Process Model for Small & Medium Enterprises (product marketed by the Cetic www.cetic.be)

 

Contributing projects

  • MoCQA :A framework for quality-oriented model-driven engineering
  • SAT : Centre for Technological Support for Aeronautical Industry : certification issues 
  • Documentation Quality in Software Engineering : Assesment Model
  • DB-Quality (Database schema quality evaluation and improvement)

 

Former projects

  • Expertise Center in Engineering and Quality of Sotware Systems - Support for certification
  • QualOSS - Quality in Open Source Software
  • Empirical Studies of S.E. Principles
  • Framework for Constructing & Validation Software Measurement

 

Senior members

Researchers

  • Jonathan LEMAITRE
  • Benoit VANDEROSE
  • Flora KAMSEU
  • Hajer AYED