190 likes | 302 Views
Sources of Requirements Change. A Goal and Viewpoints-driven Explanation. Johan F. Hoorn. Vrije Universiteit Faculty of Sciences Department of Computer Science Section Information Management & Software Engineering Sub section Human Computer Interaction, Multimedia & Culture. Contents.
E N D
Sources of Requirements Change. A Goal and Viewpoints-driven Explanation Johan F. Hoorn Vrije Universiteit Faculty of Sciences Department of Computer Science Section Information Management & Software Engineering Sub section Human Computer Interaction, Multimedia & Culture
Contents • Status • Problem: Requirements change • The requirements-analysis rift • The goals-to-requirements chiasm • Stakeholder logistics • Conclusions/Discussion • Questions M M I 9 9 0 0 9 Johan F. Hoorn, 2005
Status • Postdoc project: 2001-Aug 2005 • Supervisors:Gerrit van der Veer and Hans van Vliet • Six international publications, three pending • Mens-Machine Interactie • Supervising committee Johan F. Hoorn, 2005
Status (2) • Industries involved Johan F. Hoorn, 2005
Problem • Requirements change Software development-track Change requirements Requirements negotiation System design Requirements elicitation System specification Software implementation $ $$ $$$ $$$$ $$$$$$$$$$$ The later a change occurs, the more costs are involved in redesign Johan F. Hoorn, 2005
The requirements-analysis rift • Stakeholders regard requirements as something of the business • Stakeholders regard goals to achieve with the system as something personal • They fail to see the connection - it’s a viewpoints problem Focus switch Requirements: Business view Goals: Personal view Focus switch Johan F. Hoorn, 2005
How do we know? • Capacity Management System • Police Academy students (novice users) • Scored agreement to requirements and goals from business or personal viewpoint Johan F. Hoorn, 2005
F(1,30)= 10.19, p= .003, ηp2= .25. Parameter coefficient= .91, t= 3.19, p< .004
The goals-to-requirements chiasm • Requirements that the system MUST have change due to situations stake- holders want to AVOID • Requirements that the system WON’T have change due to situations stake- holders want to ACHIEVE Johan F. Hoorn, 2005
How do we know? • Capacity Management System (police officers) • Logistic Warehouse Management System (managers) • Commercial Off-the-Shelf Computers (interaction designers) • Braille Mouse (blind pupils) Johan F. Hoorn, 2005
Type of questions must/won’t support/obstruct goal approach/avoid E-mail ordering increases efficiency E-mail ordering decreases efficiency E-mail ordering increases inefficiency E-mail ordering decreases inefficiency Paper ordering forms increase efficiency Paper ordering forms decrease efficiency Paper ordering forms increase inefficiency Paper ordering forms decrease inefficiency Example items Johan F. Hoorn, 2005
Four replications R2adj= .90, F(5,12)= 30.30, p= .000 R2adj= .70, F(5,12)= 9.01, p= .001 R2change= .16, F(2,11)= 11.88, p= .002 COTS R2adj= .31, F(1,13)= 7.30, p= .018 R2adj= .23, F(1,13)= 5.18, p= .040 R2adj= .65, F(5,8)= 5.84, p= .015 R2adj= .73, F(2,11)= 18.28, p= .000 Johan F. Hoorn, 2005
What if the rift co-occurs with the chiasm? Johan F. Hoorn, 2005
Problem revisited • Requirements change Software development-track Change requests Software implementation System design System use $$$$ $$$$$$$$$$$ $$$$$$$$$$$$$$$$$$$$$ The later a change occurs, the more costs are involved in redesign Johan F. Hoorn, 2005
Stakeholder logistics Performance Effectiveness = .25 R2= .34, F(1,925)= 236.58, p= .000 Efficiency = .35 Usability Satisfaction 16% R2= .16, F(1,926)= 181.85, p= .000 Effort - Survey among 1943 employees - 25 different banking systems Johan F. Hoorn, 2005
Conclusions/Discussion (1) • Beware of the requirements-analysis rift • (changes in viewpoints) - You ask for their goals - You specify requirements to serve these goals - You go back to the work floor - They agree more or less to what you propose - And then while using the system they start complaining that it does not serve them well Johan F. Hoorn, 2005
Conclusions/Discussion (2) • Beware of the goals-to-requirements chiasm • (changes come from crossed relations) • Ask them: • What are the things you want to achieve with the system? • What should the system NOT have to support that? • What are the things you want to avoid with the • system? • What should be ON the system to support that? Johan F. Hoorn, 2005
Conclusions/Discussion (3) • Look at lower-level personal goals • (changes do not come from source concerns) • In most of our studies, personal goals at work were • related to • Effectiveness (e.g., getting targets, less costs, help colleagues) • Efficiency (e.g., being fast and accurate, better planning) • Effort (e.g., less work load, comprehensibility) Johan F. Hoorn, 2005
Requirements change Questions? M M I 9 9 0 0 9 Johan F. Hoorn, 2005