From Maisqual Wiki
This PhD work is accomplished in the SequeL team from the INRIA Lille, with Philippe Preux as the thesis director. The SequeL (for Sequential Learning) works on data mining algorithms and learning systems.
SQuORING Technologies, publisher of SQuORE, funds the 3 years needed to achieve this work. SQuORING Technologies has a long-term experience of software processes and products quality analysis, measurement and improvement.
The thesis work is accomplished by Boris Baldassari as the student. Boris is a Software Configuration Management expert who worked for Telelogic, IBM and his own company, Dellos Consulting, before working for SQuORING Technologies.
The different actors may be contacted at the following numbers and address.
- Squoring Technologies:
- SQuORING Technologies SAS
- 76 allées Jean Jaurès
- 31000 Toulouse
- Phone : +33 5 81 34 63 97
You can also use the contact form on the Squoring website.
- Boris Baldassari has contact information on his wiki user page.
- Christophe Peron (SQuORING Product Manager)
- Phone: +33 5 31 98 59 55
- Philippe Preux (PhD director) has contact information on the SequeL page.
The Maisqual PhD project
Much of software development nowadays relies on intuition, gut feeling and experience of senior managers for the day-to-day decisions and overall vision. When to publish a release, what module to re-engineer, how much time to implement a correction or new feature, what good practice and measures should be used to achieve specific requirements, are some of the questions that have no easy answers.
This often results in a lack of control on some or all of the required characteristics of the software system been built: short-range decisions, lack of visibility on results or improvements for the stakeholders, or even failures to achieve time and/or quality objectives. Numerous norms and standards have been created to help this, but they are still difficult to understand, apply, and measure.
The aim of this PhD work is to provide pragmatic help to all actors of the development process to improve the quality of the process and products. This is achieved by giving an accurate vision and understanding of the whole development process and providing the right decision at the right time, for the daily questions as well as the global development strategy.
How it works
Basically the main idea is to apply automatic learning and data mining advanced algorithms on software engineering, to learn from a set of good and bad projects and retrieve what did work (or not), and what should be done (or not) to achieve specific quality characteristics.
For a more in-depth understanding, this is done in a three-steps process:
- The engine builds its analysis model from the experience gathered in thousands of projects, representing open-source and corporate developments, waterfall and agile processes, desktop and real-time software.
- The engine then correlates causes and consequences through advanced data-mining technologies, extract what works and what don't, and build its decision model from this enhanced knowledge database.
- Then for a given project history, the engine analyses the characteristics of the current development and computes the shortest path from one state of quality to another chosen state: improve reliability, maintainability, etc. by applying the patterns discovered in its previous analysis and proposing what works for that purpose.
If you are interested in joining Maisqual, have a look at the dedicated page: Joining Maisqual.
More information on the Maisqual project and actors can be found here: