Quality Assurance: Automated Statistical Analysis

A few months back, we wrote a blog article on the history and continuous development of the integrity, agility and automatization of the quality assurance processes at Negotium Technologies. While this project greatly helped us assure the quality of our code, raise the satisfaction level of our customers and the working conditions of our team members, it also brought up a few interrogations, such as:

  • What is the best way to ensure that the quality of our code is constant?
  • How do we help our managers be in the know of the technical debt of a project without having to actually review the code in depth?
  • How do we ensure that our developers raise the quality of their work without having the technical manager review each code line?

Automated Statistical Analysis

The answer to all these questions is actually quite simple.  The solution resides in Automated Statistical Analysis.

The Automated Statistical Analysis enables you to use a set of rules that will verify the structure of the code to detect errors. The Automated Statistical Analysis points out code errors to the developer and provides metrics to the manager for a quick and complete view of the quality of the code. Thanks to the provided analysis, developers can improve the quality of their work and learn first hand from their mistakes making them more efficient et performant in time.

The advanced hypothesis put forward by the Automated Statistical Analysis solution is that if the code is of a good quality and the functional testing has been done thereafter, in theory, it is « impossible » to deliver bugged solutions, or at least, it is much less probable that you would deliver bad quality code to your customers.

The Automated Statistical Analysis with SonarQube

At Negotium, we have chosen « SonarQube », an open management platform, to analyse and measure code quality. The Automated Statistical Analysis program is ran on the branches master and “develop” against the build created following each “commit” sur les “branches master” et « develop ».

Here is a view of what the manager would see :

  • Global annotation of the quality of the code
  • Insight on resolution time

Automated static analysis

Here is a screenshot that shows what the developper sees :

  • Details of the issues found

Automated static analysis

That sums up this very helpful and productive solution. If you would like to obtain further information on how to build and execute the Automated Statistical Analysis program, follow my blog articles. Note that the following articles are however much more technical. Enjoy!

Leave a Reply