Change search
ReferencesLink to record
Permanent link

Direct link
The Effect of Moving from a Plan-Driven to an Incremental and Agile Development Approach: An Industrial Case Study
Responsible organisation
2010 (English)In: Journal of Empirical Software Engineering, ISSN 1382-3256, E-ISSN 1573-7616, Vol. 15, no 6, 654-693 p.Article in journal (Refereed) Published
Abstract [en]

So far, only few in-depth studies focused on the direct comparison of process models in general, and between plan-driven and incremental/agile approaches in particular. That is, it is not made explicit what the effect is of moving from one model to another model. Furthermore, there is limited evidence on advantages and issues encountered in agile software development, this is specifically true in the context of large-scale development. The objective of the paper is to investigate how the perception of bottlenecks, unnecessary work, and rework (from hereon referred to as issues) changes when migrating from a plan-driven to an incremental software development approach with agile practices (flexible product backlog, face-to-face interaction, and frequent integration), and how commonly perceived these practices are across different systems and development roles. The context in which the objective should be achieved is large-scale development with a market-driven focus. The selection of the context was based on the observation in related work that mostly small software development projects were investigated and that the investigation was focused on one agile model (eXtreme programming). A case study was conducted at a development site of Ericsson AB, located in Sweden in the end of 2007. In total 33 interviews were conducted in order to investigate the perceived change when migrating from plan-driven to incremental and agile software development, the interviews being the primary source of evidence. For triangulation purposes measurements collected by Ericsson were considered, the measurements relating to unnecessary work (amount of discarded requirements) and rework (data on testing efficiency and maintenance effort). Triangulation in this context means that the measurements were used to confirm the perceived changes with an additional data source. In total 64 issues were identified, 24 being of general nature and the remaining 40 being local and therefore unique to individual’s opinions or a specific system. The most common ones were documented and analyzed in detail. The commonality refers to how many persons in different roles and across the systems studied have mentioned the issues for each of the process models. The majority of the most common issues relates to plan-driven development. We also identified common issues remaining for agile after the migration, which were related to testing lead-time, test coverage, software release, and coordination overhead. Improvements were identified as many issues commonly raised for the plan-driven approach were not raised anymore for the incremental and agile approach. It is concluded that the recent introduction (start in 2005 with the study being conducted in the end of 2007) of incremental and agile practices brings added values in comparison to the plan-driven approach, which is evident from the absence of critical issues that are encountered in plan-driven development.

Place, publisher, year, edition, pages
Springer , 2010. Vol. 15, no 6, 654-693 p.
Keyword [en]
agile, incremental, software development, empirical
National Category
Software Engineering
URN: urn:nbn:se:bth-7636DOI: 10.1007/s10664-010-9136-6ISI: 000282100100004Local ID: diva2:835280
Available from: 2012-09-18 Created: 2011-01-03 Last updated: 2015-06-30Bibliographically approved

Open Access in DiVA

fulltext(643 kB)39 downloads
File information
File name FULLTEXT01.pdfFile size 643 kBChecksum SHA-512
Type fulltextMimetype application/pdf

Other links

Publisher's full text

Search in DiVA

By author/editor
Petersen, KaiWohlin, Claes
In the same journal
Journal of Empirical Software Engineering
Software Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 39 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Altmetric score

Total: 108 hits
ReferencesLink to record
Permanent link

Direct link