design science methodology for a phd project roel wieringa 1is seminar 31st august

IS Seminar 31st August 1

Design science methodology for a PhD project

Roel Wieringa


Goal

• Discuss one of our research projects1. Subject2. Methodological structure

• Relate this to your project


1. Subject

• Project goal: – Extending enterprise architecture design with

goal-oriented requirements engineering• Partners:– BizzDesign (Wilco Engelsman)– Novay (Dick Quartel)– UT (Roel Wieringa)


Work done so far

KAOS and i* and Tropos and causal loop diagrams and ....

ARMOR

Light Armor

Fusion: reasons to add elements

Reduction: reasons to delete elements


Result (roughly)

Stakeholder Desire

Goal

Requirement System property

Goal is desire for which budget is made available

If the goal is a system property, it is called a requirement

Requirements must be satisfied by observed system property

influence decomposition

conflict

Phenomenon Whatever is observable to stakeholders

Desire is phenomenon ideal for a stakeholder Is-a

Is-a

Is-a

Is-a


Status• Concepts are stable for now ... Fixed point in improvement

iterations– Still to look at action goals and state goals– Measure of preferences?

• Relationships require a lot more work– Influence: state→state, action →state, action →action.

• Use + and – of causal loop diagram for state → state?– Decomposition: analytic and synthetic– Conflict: logically impossible, physically impossible, technically impossible,

violation of standards, socially impossible, humanly impossible, too expensive ...

• We will do some empirical work before we return to this


2. Methodological structure

Engineering cycle (a.k.a. regulative cycle):

• Problem investigation• Treatment design• Treatment validation• Implementation (transfer to Bizzdesign)• Implementation evaluation

“Solution”Artefact

TechniqueMethod

....


Methodological structure of PhD thesis

• Problem investigation• Treatment design• Treatment validation• Implementation (transfer to Bizzdesign)• Implementation evaluation


What did we do so far

• Problem investigation: TBD• Treatment design: light Armor• Treatment validation: TBD• Implementation (transfer to Bizzdesign)• Implementation evaluation


Methodological options

• Problem investigation– What we know now: Mging large # of reqs, tracing EA to

reqs, determining impact of req change or EA change are all problematic

– To do: survey • Among BizzDesign consultants? BizzDesign customers?• In-depth interviews?• Questionnaire

• Treatment design: light Armor• Treatment validation: TBD


Initial treatment designInitial problem understanding

Improved treatment design

Improved problem understanding Treatment validation



• Problem investigation: Some survey• Treatment design: light Armor• Treatment validation: TBD



• Problem investigation: Some survey• Treatment design: light Armor• Treatment validation questions– Internal validity: LArmor in BizzDesign consultancy

mitigates identified problems?– Trade-offs: All of LArmor? – Sensitivity: All of BizzDesign concultancy? Only Bizzdesign

consultancy?



• Problem investigation• Treatment design: light Armor• Treatment validation methods

– Simulation: Wilco uses LArmor with toy example– Simulation: Wilco uses LArmor with past project data– Action research: Wilco uses Larmor in a project– Opinion research: Wilco elicits opinion of Bizzdesign consultants about

Larmor concepts & method– Experiment: Wilco studies BizzDesign consultants using Larmor on an

artificial problem, compares to control group using current BizzDesign method

– Case study: Wilco studies BizzDesign consultants using Larmor in a project


Methodological issues

• Generalization– Applicable to similar GORE methods?– Applicable to similar companies/projects?

• Theory– A theory is a dictionary plus a statement with scope of

application; mechanism counterfactually supported• Preferably, statement indicates mechanism• Scope = generalization; always uncertain, possibly limited• Mechanism• Counterfactual

– Theories applied?– Theory that will explain case study observations?


• Organizational design science (Van Aken)– Elaborate problem investigation (diagnosis)– Like a physician

• Artefact design science (Wieringa)– Elaborate treatment design and validation– Like a medicine manufacturer


Discussion

• What about your project?• Written presentation structure?

– Problem – treatment – validation, with varying emphases• Oral presentation structure?

– Table of contents of two items– Context: the stage– Goals: what are you going to do– Why interesting for audience– What is new wrt related work– Substance 1: the version that everyone can understand– Substance 2: brief indication of some technical detail– Please keep your slides readable: minimum point size is 18


Discussion



– Table of contents of two items– Context: the stage– Goals: what are you going to do– Why interesting for audience– What is new wrt related work– Substance 1: the version that everyone can understand– Substance 2: brief indication of some technical detail– Please keep your slides readable: minimum point size is 18


Discussion



– Table of contents of two items– Context: the stage– Goals: what are you going to do– Why interesting for audience– What is new wrt related work– Substance 1: the version that everyone can understand– Substance 2: brief indication of some technical detail

– Please keep your slides readable: minimum point size is 18

design science methodology for a phd project roel wieringa 1is seminar 31st august

Documents