Authors:
Andrej Dyck
;
Andreas Ganser
and
Horst Lichter
Affiliation:
RWTH Aachen University, Germany
Keyword(s):
Model, Recommender, Recommender Systems, User Interface and Interaction, Modeling Support, Survey.
Related
Ontology
Subjects/Areas/Topics:
Frameworks for Model-Driven Development
;
MetaModeling
;
Methodologies, Processes and Platforms
;
Model-Driven Software Development
;
Models
;
Paradigm Trends
;
Software Engineering
Abstract:
Recommender systems for source code artifacts are newly emerging and are now successfully supporting programmers. Their underlying knowledge bases, recommender algorithms, and user interfaces are well studied. Integrated into the development environment, they do a fairly good job in reducing complexity and development time.
In contrast, research in recommender systems for domain modeling is widely missing. As a matter of fact, knowledge bases, studied as model libraries, are only a possible foundation but concerning recommender algorithms and user interface design research needs to be conducted.
Hence, we provide some foundations for graphical user interface design by answering how domain model recommender systems should look and feel like in graphical environments. To do so, we conducted a three-phased survey relating to modeling of UML class diagrams. Most importantly, we found that various user interfaces are required to meet different user needs. Finally, several algorithms are d
esired for diverse knowledge bases and diagram types; hence, leading to a demand for a flexible recommender architecture.
(More)