A metamodel based approach
for customizing and assessing
agile methods
Hajer Ayed, Benoît Vanderose, Naji Habra
{ hayed | benoit.vanderose | naji.habra } @ fundp.ac.be
Precise Research center, University of Namur, Belgium
8th International Conference on the Quality of
Information and Communications Technology, Lisbon, Portugal, 3 to 6 September 2012
Overview
Problem statement
Research objectives
Key literature reference
Proposed approach
Agile metamodel overview
Preliminary results
Problem statement
Observations :
– Spread of agile methods
– Difficulty to select the suitable method
– Ready-made methods can hardly take
into account a specific situation
Agile methods customization : A
case study
Objectives :
– Examine the benefits of
agile methods tailoring
– Investigate the usage
and tailoring of XP and
Scrum and how they can
be combined
Problem statement
Need of a general approach to assist
organizations in the customization
process
Need of process measures to assess
the customized agile method
Customized agile methods are often
monolithic : they rarely evolve with
environment change
Research objectives
Investigate a high-level / generic
approach for creating
context-specific agile methods
Focus on the assessment of the agile
method through process and product
measurements
Refinement of the constructed agile
methods thanks to measures
Key litterature reference
Process Modeling
Situational Method Engineering
(SME)
Quality Assessment Modeling
SME
SME
PM
PM
QAM
QAM
Proposed approach (1)
We propose an approach based on the SME paradigm The main components of the approach:
– Agile metamodel
– Agile components repository (stored in the agile repository )
Given the initial agility requirements and the initial
context, the agile methodologist designs a customized method (which is an instantiation of the metamodel) The ongoing agile process is an instantiation of the
designed method
Process and product quality measurement feedbacks enable the constructed agile method evolution
Agile metamodel
overview
Preliminary results
- We considered the real-life experience of Intel Shannon, Ireland (that embraced agile methods, mainly XP and Scrum, to meet the challenge of rapid time to market)
- Instantiation of the proposed approach : At a given time (t)
During the development of the release (R) Agile Team (AT1)
- We measure :
Defect Density (DD) Release Schedule (RS)
System Design Instability (SDI)
- If SDI exceeds a given threshold : Add "Refactoring"
- If high rate of DD and late in RS : Plan a "Reflection Workshop" between the concerned stakeholders (i.e, AT1)
Future Work
Elaboration of a catalogue of reusable interpretation rules that will help provide effective and informed methodological decisions (i.e., inclusion of relevant process and products element)
These rules will be defined through surveys and case studies reviews
Validation of the approach on industrial case studies Tool support for the approach will be provided by