3. Motivation
‣ Semantic interoperability and vocabularies for Linked Data initiatives
‣ Ontology engineering is far from trivial
‣ Requires appropriate methods and tools
‣ Hybrid Ontology Engineering platform in the context of the Open
Semantic Cloud for Brussels (OSCB) project
‣ Evaluate usability and user satisfaction of the hybrid ontology platform
3
4. Background
‣ IBM Post-study System Usability Questionnaire (PSSUQ)
4
Information Quality
System Use
Interface Quality
Overall
5. Background
‣ Hybrid Ontology Engineering
‣ Community promoted to first-class citizen
‣ Social interactions are
‣ formalized
‣ lead to ontology evolution
‣ supported by a glossary
‣ Concepts are described
‣ formally --> with binary fact-types called lexons
e.g., <Research Domain, Call, with, of, Deadline>
‣ informally --> with glosses
5
8. GOSPL Method and Tool
‣ With respect to the previous study
‣ System usefulness scored best
‣ Information quality needed most improvement
‣ Changes made to the system to tackle problems
‣ More (online) documentation
‣ Changes to stylesheets
‣ RSS to keep track of discussions
‣ Additional functionalities
‣ Tool support for managing annotations
‣ A reputation framework
8
9. Usability Test Design
‣ 23 MSc student volunteers part of an ontology engineering course
‣ Objective:
‣ prototype ontology related to research domain
‣ use ontology to annotate
‣ “reverse engineered” existing application
‣ existing research information system
‣ Evaluate formative and summative user satisfaction of the ontology
engineering platform
‣ Compare with results from a previous study with similar experiment
9
10. Usability Test Design
‣ Tasks on the platform
‣ Propose changes to the hybrid ontology
‣ Discuss and vote
‣ Conclude discussions --> ontology evolution
‣ Create and manage communities
‣ Annotate information systems
‣ Participants
‣ 36 students, 9 groups for the ontology engineering experiment
‣ Background in computer science or similar
‣ 23 students participated in the survey (of which 4 complete groups)
10
11. Summative User Satisfaction
‣ Summative evaluation
‣ System usefulness and interface quality remained the same
‣ Information quality and overall satisfaction improved
‣ Formative evaluation
‣ Tracking “hottest” discussions
‣ Correcting mistakes
‣ Voting mechanism (!) and verbalization of constraints (!)
‣ Positive outcome given the problems encountered with application
management and the reputation framework
11
[2] Ioana Ciuciu, Christophe Debruyne:Assessing the User Satisfaction with an Ontology Engineering Tool Based
on Social Processes. OTM Workshops 2012: 242-251
12. Recommendations
‣ Overview of discussions need improvement
‣ (Error-) management of application commitments
‣ Verbalization
‣ Removing voting mechanism
12
13. Conclusions and Future Work
‣ Conclusions
‣ System usefulness performs best
‣ Information quality improved a lot
‣ Overall satisfaction remained stable
‣ Given the fact that new functionalities posed problems, this can be
interpreted as a considerable improvement of the tool
‣ Future work
‣ Third iteration planned in May 2014
‣ Discussion overview + lightweight user participation
13