The Social Requirements Engineering (SRE) Approach to Developing a Large-scale Personal Learning Environment Infrastructure
Effie Lai-Chong Law, Arunangsu Chatterjee, Dominik Renzel and Ralf Klamma
Department of Computer Science, University of Leicester, UK
Chair of Computer Science 5 - Information Systems, RWTH Aachen University, Germany
EC-TEL 2012, Saarbrücken, Germany
September 21, 2012
2. Personal Learning Environments (PLE)
“a pedagogy-driven infrastructure that facilitates learners to
integrate distributed contents services tools and contacts
contents, services,
based on personal goals and preferences, thereby enabling
them to control their own learning and to connect different
g
learning contexts with the support of communities”
3. Motivation
I know which tools my
Which tools should I learners would need
need...
develop to create But I can‘t develop!
most impact?
4. Motivation
Users (many)
high potential to elicit requirements from domain knowledge [Hipp86]
usually not acquainted with formal specs developer tools & tech jargon
specs,
need intuitive means for stating community-specific requirements
dependent on developers realizing their requirements
Developers (few)
often not acquainted with domain knowledge and community jargon
use formal specs tools & jargon for reasons of effectiveness & efficiency
specs,
used to integration between development tools
interested in requirements prioritization to maximize outcome
Needed: model & software to support negotiation process
ROLE Social Requirements Engineering (SRE)
ROLE Requirements Bazaar
5. Theoretical Premises
Transcriptivity Theory (Jäger, 2002) describing the operational
semantics
ATLAS (Architecture for Transcription, Localization, and
Addressing Systems) - , scalable and interoperable repositories
on top of databases support communities by web service
technologies (Jarke and Klamma, 2006)
Communities of Practice (CoP) (Wenger, 1998)
Actor-network-theory (ANT) (Latour 1999)
Details available via ROLE website > Results > Deliverables > Wp1
deliverables
6. The ROLE SRE Approach – Support for the Long Tail
No Mainstream Web 2.0 RE!
“O
“Overall T N“ naive approach
ll Top-N“: i h
Needs of specialized CoPs neglected
I
Innovation Killer ( l
ti Kill (clones only)
l )
Rather Long-Tail RE
“Community-Aware Top-N“
Special support for niche CoPs
High specialization, but high innovation
10. ROLE Requirements Bazaar – Early Prototypes
Quantitative assistance mechanisms
during negotiation
Hybrid ranking mechanism
Web-based requirements elicitation form
Requirements review
11. Lessons Learned from Early Prototype Evaluations
Evaluations in 2 workshops in 2011
Developers & pedagogical experts (N~20)
Early-stage TEL researchers (N~20)
Issues Identified
Culture & language diversity
Lack of common understanding about PLEs
Inadequate communication
I d t i ti
Distinction between infrastructure and widgets requirements
Lurking
Resolutions Proposed
Documentation for improving understanding of PLE
Easy-to-use templates for requirements elicitation
Mandatory voting
Prioritisation model
12. Intuitive Means for Requirements Elicitation
Comic-like
C i lik annotations
t ti Web 2.0 feedback tools:
on screenshots/Storytelling uservoice.com
getsatisfaction.com
15. Useful Sources for Requirements Ranking
User-to-Service Communication
CoP-aware usage statistics
Identification of successful CoP services
Identification of CoP service usage patterns
User to User Communication
User-to-User
CoP-aware Social Network Analysis
Identification of influential CoP members
de ca o o ue a Co e be s
Identification of CoP member interaction patterns
+
16. ROLE Requirements Bazaar –
Requirements Ranking Infrastructure
URA = User Rating Analysis (baseline)
UMA = User Monitoring Analysis
18. Lessons Learned
Lack of common understanding and awareness about PLEs.
Developers/teachers/researchers acting as “surrogate customer”
which is a potential problem
Inadequate communication among “end-users” and developers
around proposed requirements
Cultural diversity and language barrier – Chinese, German and
English
View Separation important as demonstrated earlier
Expose relevant requirements as per users context
Ensure easily configurable views
Devise mechanism to counter ‘Cold Start’ problems
Mandatory voting
Reward mechanism in terms of visible reputation scores or badges
19. Future Improvements
E
Ease of contribution
f t ib ti
Create templates that users can edit rather than forcing them to
start from scratch
Allow users to enter requirements in their preferred language
Integrate with development infrastructure (e.g JIRA) for enhanced
traceability
t bilit
Prioritisation model
Attempt of extract rationales behind ratings
Utilise SNA and Monitoring data to assist decision support via a
RE dashboard
Requirements dashboard
Users able to view their own activity
View community activity