SucheKnowledge based Annotation of Requirements and Reuse of Components
Abstract:
Software engineering nowadays is most often a very complex, inefficient, not clearly schedulable and time consuming task. Most of these facts can be related to the software methods used. These have rarely been upscaled to the needs of modern software development. Although reuse has often been called one key in the solution to those problems no methods for an automatic detection of specific artefacts exist. The development of an automatic method which identifies suitable components based on specific requirements would greatly facilitate the requirements- / software engineer's life. Yet, the standard way is to use Google or any other search engine to find matching components.
This work describes an approach to search automatically for already existing components. One major requirement behind this goal is that the user doesn't have to incorporate a lot of additional work as he would have to in other approaches. This means that he simply enters the normal requirement text into the software, which, ideally, automatically analyzes the text and recognizes all kinds of information needed. These would then be marked with tags. In a next step, the so annotated text is inserted into an ontology, which is then compared to the ontology of one other artefact at a time. In the end, a similarity and satisfaction value are returned. Both of them should help the user to find matching components more efficiently than it would take him if he did the search manually. This should reduce the amount of time needed to finish the project as also improve the overall quality.
Starting with a basic introduction about semantic software engineering, the description of the algorithm and the architecture of the prototype is shown. It is then evaluated on different parameter settings, using different example ontologies.
Downloads:
- 2007-10 - (2007-10.pdf, 1101 KB)