The main objective of this research and development project is to design and evaluate new ways of information and knowledge acquisition and its further processing in environment of a software house. Looking at software development, and specifically at enterprise application development (EAD), which has historically been a closed in-house process, we observe that the used information space is rich in interconnected information artifacts (e.g., documentation, source code, developer blogs or help forums). Many useful methods devised for the Web can be used to search, browse and navigate within this web-like structure. We believe that by leveraging experience from rapidly evolving field of information processing on the Web and employing web-based methods in the EAD domain one could improve software quality and development efficiency, e.g., support the creation of better code, improve progress visibility or help developers be more efficient. This can be facilitated by providing users (both developers and managers) with the right information in the right context.
We model the content and actor-related artifacts that affect and support day-to-day work of software developers in these categories:
- Source codes, consisting of events in source control repositories and their relations to bug reports, customer change requests, persons, projects.
- Knowledge documents that the developers work with (e.g., best practices, guidelines, notes stored in a company's intranet, blogs and bookmarks).
- Developer interactions, consisting of meta-data about email and instant messaging, and emergent interactions in the source code.
We opted for a lightweight representation of semantics of various domain artifacts as well as of users' relations towards these domain artifacts: information tags – metadata that describe an aspect of an artifact. The information tags are an extension to the basic concept of a tag as a simple keyword or term assigned to an artifact; in that an information tag is any metadata that adds additional value to the artifact itself; e.g., explicit and implicit feedback generated by developers working on source code. In our case, the information tags contain explicit and implicit feedback generated by software developers working on the project.
The research is directed to these actual areas:
- models of information and knowledge (models of application domain, user models, context models and their combinations),
- representation and storage of models holding metadata,
- knowledge acquisition (acquisition of user model/environment model via comprehensive monitoring and data processing solutions),
- knowledge generation and presentation (recommendations, machine learning).
co-funded by the European Regional Develoment Fund
Investigators: Gratex International, a.s.; Faculty of Informatics and Information Technologies - Slovak University of Technology
Project location: Galvaniho 17/C, 821 04 Bratislava | Ilkovčova 3, 842 16, Bratislava, Slovak Republic
ITMS: 26240220039, March/2011 – January/2015