Quality Monitor™ reverse code engineering: the view (by yWorks) shows the actual UML design of a small system.
Quality Monitor™ static dead code and dynamic test coverage analyses: the view (by yWorks) shows that all tests worked fine (green), but some classes (boxes) are not covered (blue). Among those there are many ”dead” (light blue) classes, i.e., they cannot be tested.
Research
Softwerk AB/ARiSA is partner in the following research projects:
Effective and Efficient Information Quality Assessment (2012-2015) of technical documentation is a KK foundation funded project. It applies metrics-based quality control to technical documentation, e.g. user manuals and technical specifications. Technical documentation often constitutes the first line of support when users want to learn more about a product or a service. The information provided needs to be correct, relevant, etc. that makes effective information quality assessment mandatory.
Technical documentation has moved from printed booklets to live electronic versions. These are easier to keep up to date and to make available to end-users, and more flexible in accessing information, e.g. by using hyperlinks, search functions, and interactive features. The obvious benefits of electronic technical documentation do enforce a more complex process in terms of producing, maintaining, and publishing technical documentation, which requires efficient information quality assessment all the while it enables the same.
Information quality is often defined as "fitness of use", i.e. the users' satisfaction with the documentation. This makes quality costly and cumbersome to assess. If at all, it is operated manually with checklists and questionnaires. We suggest that (parts of) this assessment can be automated in order to get better economy of scale. We borrow from software engineering, where quality is considered as "conformance to requirements" assessed in accordance to predetermined requirements, which allows for (semi-) automated approaches like measurement and testing.
Our first objective is to develop a novel approach to effectively and efficiently assess information quality, with validated theory, models, methods, tools, and implementations in production settings. Our second objective is to quantify and predict costs and benefits of this approach.
Technical documentation has not only moved closer to software in the way it is produced, maintained, and published. Both documentation and software are part of many technical products and services supporting the same overall requirements. Our third objective is therefore to develop a theory to uniformly and holistically assess quality of software and information artifacts.
From a scientific perspective, the challenge is to define technology for efficient information quality assessment and to show its effectiveness, i.e. correlation to "fitness of use". For validation, we perform experiments aiming to statistically reject the hypothesis that any covariance between our (semi-) automated approach and the state of the art is coincidental. This means that our approach assesses information quality at least as accurate, but more time- and cost-effective.
From a business perspective, it is interesting to adapt this technology to improve technical products and services leading to better prices and user experiences and to control its costs and benefits in real-world situations. Therefore, we collaborate with Sigma-Kudos and Ericsson providing access to real-world content management systems and technical documentation, and expertise in different stakeholders’ perception of quality.
Next Generation Software Engineering (2011-2014) is a KK foundation funded project. The objective of this project was to introduce search-based techniques, i.e. metaheuristic techniques, in an industrial setting. This was mainly conducted by a series of validation steps (internal and external) where the techniques were tested on non-trivial examples. Initially the candidate techniques were selected in collaboration with industry (focusing only on the needs of industry), next a minimum of two internal validation steps took place and, finally, case studies and experiments were performed in industry.
Validation of metrics-based quality control (2005-2008) is a KK foundation funded project. We develop a method for quality assessment integrating industrial standards and automated software measurement (metrics). The research goal is the validation of this method in industry projects. The goal for the partner companies is an improved quality management. In selected development projects, partner companies assess quality characteristics suggested in ISO/IEC 9126 using automated and manual means. We try to validate statistically the significance of the automated approach. The expected result is a service assessing industrially standardized software qualities in an effective (since significant) and efficient (since automated) way.
