Article Preview
TopIntroduction
Web technologies have changed the way teaching and learning is facilitated using digital means. With the inception of the Web now being more than 20 years ago, many web-based learning technology systems (LTS) have had a long life and have undergone many changes, both affecting content and infrastructure technologies (Palmer & Tulloch, 2001). In Pahl (2003), a facetted factor model to capture factors that cause LTS to change was introduced and a change-aware development method for LTS was suggested. The first objective of this investigation is to add more detailed empirical results to the change model in order to show the relevance of the model. The second objective is to develop these results further into an impact model for change analysis and impact assessment that can support the LTS software development and maintenance process.
We use as our case study an LTS that has been developed, maintained and extended over a period of 20 years (Smeaton, 1991; Smeaton & Crimmins, 1997; Murray, Ryan & Pahl, 2003; Pahl, 2008). During this period some notable changes have been observed, beyond the expected content updates to reflect changes in the subject domain. These include different instructors, an integration of content and infrastructure with the university's learning technology platform, the export of learning objects to content repositories and the shared delivery of content with several universities. A key observation is an unexpectedly high impact of environmental constraints, i.e. aspects that define how the LTS relates to its technical and organizational environment. In addition, new opportunities emerging and reacted to over time (Devedžić, 2006) have had a positive impact on the student learning experience and the effectiveness of system, but also massively changed the LTS itself, having an impact for those managing and delivering content through it.
Ten years on after the publication of an LTS change and evolution discussion (Pahl, 2003), the proposed solution shall be revisited in the light of collected observations over this period. The facets pertaining to the design and use of an LTS were categorised into content, format, infrastructure and pedagogy aspects. We evaluate this change model and discuss its adequacy. Changes do not only happen as a once-off activity, but are part of a lifecycle and reflect changes in the wider context of an LTS. We discuss the impact of changes in technology or teaching and learning research on LTS in everyday use and the long-term costs of these changes. Using a case study analysis, we integrate the lessons we learned developing, maintaining, and using a concrete LTS into a change impact analysis and assessment model. Our contribution is a discussion of change factors in the context of learning technology and infrastructures and how evolution impacts on LTS. Our aim is to highlight the dangers and pitfalls, despite improvements and advancements facilitated through change and evolution. These are captured in an impact model that can serve as an analytic model for LTS stakeholders.
This study confirms concerns in the software engineering community (Easterbrook, Singer, Storey, & Damian, 2008). Software ageing is a known concern for software development characterized by increasing maintenance costs (Taylor, Medvidovic, & Dashofy, 2009). As we will see, we need to look at software-centric ageing as well as contextual ageing. Two concrete forms are architecture erosion and degradation. Architecture erosion is caused by incremental changes of the system. Degradation is the abrupt consequence of some distinct events in the environment of the LTS. In order to facilitate the applicability of the impact model, we introduce suitable metrics for the architectural change concerns.
We first introduce principles of learning technology systems and the LTS we use as a discussion case study, the IDLE system. We use the change factor model to summarize the evolution of IDLE over its lifetime. The introduction of an impact model based on lessons learned follows, which takes into account observations and analyses of the system successes, but also difficulties and costs resulting from software ageing.