Abstract:

In learning environments team collaboration face the same challenges as in business scenarios and without proper management team collaboration can become more an obstacle than a solution, Workflow Learning is the adaptation of WfMS (Workflow Management Systems) from the business domain in the learning domain (DePree et al. 2011).

The aim of this research is the design, implementation and evaluation of a Software Reference Architecture for WfLMS (Workflow Learning Management Systems) with Mobile, Cloud and Collaborative functionalities. The Reference Architecture was the result of applying the WfMS Reference Model in conjunction with Mobile, Cloud and Collaborative Architectural pattern solutions.

A Reference Model explains in a simplified way, how to resolve a problem in a specific domain. According with Albin (2003) a Reference Model decompose a problem in a functional way describing the components its functionalities and the connections between the components. Architectural Patterns are well known solutions for solving architectural software problems and are domain independent (Angelov et al. 2012).

The Design Science Research methodology was applied in this research as the principal goal of the research is the same as in the methodology, the creation of artifacts for a practical purpose. This methodology should be use for address unsolved problems with an unique and innovative way or for trying to enhance a solved problem in a more effective or efficient way (Hevner et al. 2004). Nicolaou & Geerts (2011) mention that the four possible artifacts that can result from the application of this methodology are: concepts, models, methods and instantiations.

The implementation of the Reference Architecture resulted in a mobile WfLMS for the iOS platform. A Delphi Study with a group of 12 experts was carried out in order to evaluate if the Mobile, Cloud and Collaborative architectural pattern solutions applied in the Reference Architecture enhanced the functionalities of a WfLMS. The guidelines for how to design a Delphi study proposed by (Worrell et al. 2013) were followed.

14 factors were identified in the pattern solutions that were implemented, the experts ranked each one of the factors with a 5 point Likert scale in a third round questionnaire. 12 of the factors achieved more than 75% of agreement with a strong level of consensus.

The statistical analysis method was based on (McGinn et al. 2012) where consensus on a factor was considered strong when 75% or more of the experts reached an agreement, moderate consensus when the agreement was between 60% and 74% of the experts, and the absence of consensus when less than 60% of the experts agree. To calculate the level of consensus McGinn et al. calculated percentile scores (10th and 25th) where the 10th percentile indicated the lowest number on the Likert scale upon which at least 90% of the participants agreed and 25th percentile scores indicated the lowest number on the Likert scale upon which at least 75% participants agreed. The statistical dispersion function known as interquartile range was used for measuring the strength of the consensus, a result close to 0 indicates a strong group consensus and close to 2 indicates the responses of the experts are disperse.

Keywords:

m-learning, workflow learning, collaborative, cloud, reference architecture, mobile.