|
The following dimensions form the basis for our analysis of the identified TEL interoperability frameworks: layers, viewpoints, roles, quality attributes, and patterns. The table below gives an overview over the characteristics of each framework.
Hardly any of the reviewed initiatives documents or exhibits explicit efforts in considering and evaluating issues of design methodologies, engineering instruments, neither ex ante nor ex post with respect to their concrete design decisions. From an observer’s perspective, methodological or methodical considerations, therefore, seem to be considered irrelevant either to the presentation of research results or even the design process in general. This general lack of attention paid to issues of methodology and methods might equally be seen as proof of an implicit consensus among researchers in TEL with respect to the foundations of a design process in this field. This canon refers predominantely to notions and conceptual work documented for information systems in the context of business organisations. There are, certainly, notable examples that diverge from this general observation: In the context of ELF, Wilson et al. (2004c) take such a meta-perspective and consider design guidances for their frameworking efforts, in particular with respect to concepts of Service-Oriented Architectures.
Table. Comparison of identified TEL interoperability frameworks.
As for the comparison of framework initiatives given in the table above, there are some insights worth commenting briefly:
- Layering is a design and documentation method widely adopted by the approaches under review. However, responsibility-centric or functional layering is the predominant layering strategy adopted which allows for concluding that most initiatives focus on functionality specifics in structurally dissecting their design solutions. An important exception from this apparent tendency is CORDRA with the authors giving an important example for re-usability-centric layering in the field of TEL. CORDRA describes a reference model that is not centred on different re-usability levels with respect to (business) organisations but various community scopes. Apart from the general layering strategy, two approaches, namely LORI and IMS TI, employ indirection layering which resembles characteristics of distributed environments of legacy systems that subject to integration by dedicated wrapper add-on functionality.
- The employment of viewpoints as instrument of requirements engineering is, as expected, common to all approaches, though limited in various respects across the initiatives inspected: An important limitation traceable for most of the documented experiences is the low number of viewpoints devised, i.e. the small number of viewpoint agents considered during the design process. Apart from IBM LI/SIS and IMS TI, there are only single viewpoints considered or at least publicly documented. The actual viewpoint agents considered, especially in the single-viewpoint cases, are developers or implementers. This aggravates the conceptual limitations introduced by the singleness of perspectives actually taken, especially in the context of Technology Enhanced Learning. While IBM LI/SIS complementarily comprises an organisational viewpoint and therefore an aggregated viewpoint agency, IMS TI is the only to introduce learners and teachers as explicit stakeholders and corresponding viewpoints.
- As for the roles conceptualised by frameworking initiatives, the majority emphasises the application level of integration which partly abstracts from or comprises aspects of information integration. However, integration at the presentational level is hardly discussed or conceptually incorporated by the major part of efforts reviewed. Exceptions to the latter are IMS TI and LearnServe.
- Considering quality attributes , it becomes obvious that sets of rather conventional quality types, e.g. performance, security, usability etc., have been identified and applied to inform and evaluate the design solutions that are reported above. More contextual attributes, such as attributes related to the operative environment, as already addressed for business organisations (see e.g. Clements et al., 2002), are entirely missing.
A prominent conclusion that offers itself from the comparison table is the complete absence of pattern-based considerations in the initiatives evaluated. This relates to some extent to the observations made for quality attributes. Though there are important efforts made in devising patterns for the design of learning management systems (i.e. Avgeriou et al., 2004), there are no efforts made for distributed and composite learning systems that iCamp deals with.
– Stefan Sobernig, Fridolin Wild
|
iCamp feed