Context Management for Ecore Library Recommenders

Being able to adopt and reuse already developed software is a key feature of current software design tools. Especially in Model Driven Development, where UML diagrams and code generators are used, the possibility to re-use those models is an important feature. Due to the mass and complexity of existing models the users of such software design tools need to be supported during the development of new models and advancement of existing models. Optimal support includes recommender systems, which will use context informations from the user’s actual focus. This context informations will not only contain search-contexts like the actual domain, but also paste-contexts like the UML diagram canvas, cursor position and more. A sufficient context management provides these informations in a structured manner, which may then be used by the recommender system to support the user’s decision. This thesis will discuss a concept for context management supporting an Ecore recommender system. The resulting concept will contain the required features, like context information providing in different search- and paste-contexts such as tree views and UML diagram canvasses. To validate the presented concept an implementation into the EMF(Eclipse Modelling Framework) plug-in MOCCa (Model Composition and Combination Vault) tool will be applied.

Project information

Status:

Finished

Thesis for degree:

Dipl. Inform.

Student:

Niklas Franken

Supervisor:
Id:

2012-025