This research study recognizes that providing personalised courses with structured learning objects based on leaners’ background and progress oftentimes requires reshaping learning object representation before placement in the pool of learning resources, in the Cloud. For avoiding the potential problems that could occur while linking the learning objects as discussed in Chapter 3, we have driven the process of adapting the learning objects throughout these three phases:
1. The granularity of learning object is segmented as small as possible, as a chunk object with a content that could stand as an entity by its own,
2. The existing learning objects are enriched with extra annotation for increasing the flexibility of coupling the learning objects with each-other,
3. The context alternatives of that smallest learning object are specified based on the new granularity of the LO.
These three mandatory phases are necessary conditions, because for a LO to be shared and reused the granularity of the content is very important, which is referred as the size of the learning objects. Noor et al. stated that the lower the granularity of the learning object is, it increases the chances to be reused in different context [36]. Whereas from the other side, Shoonenboom [40] has described different scenarios for determining the size of LOs, and the ability to reuse the modules in the personalised way.
After redefining the granularity of the LOs in the first phase, the learning object enrich- ment and the definition of context alternatives respectively, contribute to the flexibility of linking the sequence of Cloud eLearning learning objects. By increasing the flexibility, the
6.3 Modeling the knowledge domain for Cloud eLearning 87 idea advances for combining the content and metadata of the learning repositories for offering loosely coupling of LOs in different sequences in different context [40].
Different context or domains refers to the number of domains in which a particular LO could be used. Wiley [5], expresses the different domains as the ability of reusing the learning object as “inter-contextual use”. Analyzing each of the aforementioned phases, the evolution of learning objects into Cloud eLearning Objects (defined in definition 1.2) has come about naturally. The transformation process of Learning Objects into Cloud eLearning Objects is discussed further in the next section.
6.3.1
The evolution of Learning Objects into Cloud eLearning Learn-
ing Objects (CeLLOs)
As discussed in chapter 3, section 3.2, the pool of learning resources of various format and standards and, in this context, the creation of Cloud knowledge and its representation requires a common standard in order to be able to tailor a personalised learning path and create a coherent course otherwise we might end up with the situation as shown in Figure 6.4.
Fig. 6.4 A sequence of unstructured learning objects
In this respect, the learning objects that are derived from different sources, irrespective of whether these are structured in some standards and stored in some learning object repositories or if they are totally unstructured and untagged, have experienced a transformable process.
Fig. 6.5 Generating CeL Learning Objects from standard Learning Objects
This transformable process generates a new type of Learning Object that is usable, processable and applicable for CeL as shown in Figure 6.5. While this transformable process has created significant challenges, it has also made possible a coherent learning path of tailored learning objects. So, the result of the transformed process generetes Cloud eLearning Learning Objects, known as CeLLO.
Definition 6.1: The CeL Learning Object (CeLLO) as an advancement of learning object is defined as a structured electronic learning resource of a reasonable size and that satisfies an intended learning outcome.
The transformable process of integrating LOs into CeLLO is accompanied with addition of extra features/metadata to all existing learning objects to glue the LOs together in more coherent way. The additional metadata is added to each of the learning materials based on the extra information that are needed for the CeL framework. These additional CeL metadata together with existing metadata in LOs form the so-called CeLLO (Figure 6.5).
Definition 6.2: “Cloud eLearning Metadata - CeLMD” is a metadata approach used to transform the derived Learning Objects (from various sources) into CeL Learning Objects.
6.3 Modeling the knowledge domain for Cloud eLearning 89 As shown in Figure 6.6, the elements of CeL Metadata are subsets of particular elements from existing metadata schemas, such as Dublin and IEEE LOM, with the addition of new elements that are required to achieve CeL aim.
The elements of Cloud eLearning Metadata are as follows: (1) Title, (2) Description, (3) Keyword, (4) Content, (5) Meta-metada, (6) Catalog, (7) Pre/Post requisite, (8) Relationship, (9) Intended Learning Outcomes, (10) Format, (11) Granularity, (12) Cognitive Level, (13) Context, (14) Credibility, (15) Crowd rating CeLLO, (16) Crowd rating set of CeLLO, (17) Date, (18) Language. The Figure 6.6, shows visually which of the elements are derived from Dublin Core and IEEE LOM. The extra added new elements are described seperately in Table 6.1.
Fig. 6.6 CeL Metadata related to IEEE and Dublin core standards
The transformation of unstructured, semi-structured or fully standardised LOs into useful LOs for CeL could be done in a variety of ways ranging from manual, semi-automated or even fully automated which will be discussed as part of future work [34]. However, in this thesis we have proceeded with the manual option for reasons of simplicity and to avoid the associated complex situations which are out of this PhD scope.
The overall flow process of transforming LO into CeLLO is depicted in Figure 6.7, which describes the retrieval of a LO and its adaptation to CeL, as a CeLLO entity. As mentioned, the LOs are derived from different locations, which firstly were checked whether they already support an existing LOM standard. If not, the CeL metadata is applied fully to these LOs. Otherwise, if the LOs support already an existing standard, then the existing elements are inherited from that standard and the new CeL metadata elements are added manually.
Metadata elements such as Title, Description, Keyword, Catalog, Pre/Post requisite, Relationship, Format, Granularity, Context, Credibility, Date, Language, are reused from
Fig. 6.7 The transformation process of a LO to a CeLLO
existing metadata schemas, such as Dublin Core and IEEE. The new elements that are required to achieve the CeL aims, such as intended learning outcomes, crowd rating, cognitive level, meta-metadata, and content are explained in Table 6.1.
In CeL context, a sequence of CeLLOs is generated from a planner (to be discussed in chapter 8) which automatically generates a coherent path, in which CeLLOs have pre- and post- conditions that correspond to what the learner knows and what the learner wants to achieve, respectively. CeLLOs are carefully selected from the pool of available CeL learning objects through a recommender system (discussed in chapter 7) that matches learners’ preferences to suitable learning material. A concrete example is going to be demonstrated in chapter 9.
6.4 Modeling Learner for Cloud eLearning 91