(Figure 1, from a presentation by Bill Olivier, Nov 2005)
(Figure from presentation by Bill Olivier, Nov 2005)
Reference models are part of a bigger conceptual framework. What are their relations to specifications and standards, to service oriented frameworks, to services and other concepts that are now coming into vogue? Again, let us follow the traces of the e-Framework partners.
In July 2004 DEST (Australia), JISC-CETIS (UK) and Industry Canada presented a paper on "Service-Oriented Frameworks: Modelling the infrastructure for the next generation of e-learning systems" (Wilson et al., 2004). The paper aimed at providing a common set of concepts and terms that could be used in conversations about e-learning infrastructure, and to present a case for developing and maintaining service-oriented e-learning technical frameworks. Basically the authors proposed four concepts to facilitate the conversation, see the figure below.
(Figure 3 of a simple relationship model, from Wilson et.al., 2004)
To start in the soft end, framework provides a vocabulary and grammar and it is left to individual organisations to write the stories. According to Wilson et al. "a framework creates a broad vocabulary that is used to model recurring concepts and integration environments. (...) A Reference Model is a selection of Services defined in one or more Frameworks. (... ) A Design specifies an Artifact, such as a piece of software, and is a collection of specific technologies applied to either a Reference Model or the Framework. (...) An Artifact is a realisation of a Design."
You might derive multiple Reference Models from a single Framework, and a Reference Model might be derived from multiple Frameworks.
At first glance this model might seem a little too abstract to spark heated discussions about different directions of technology development for learning, education and training. The primary objective of the model is not to provide the language for the conversation, but to point out the conditions for having such a conversation at all. And then you need a language. This is not as self-evident as it may sound.
Technologies very often present themselves as “black boxes” – they "just work" and make the political, social, cultural and other filtering of knowledge opaque (Hoel, 2005). You almost need some kind of "reverse engineering" to tease out the values and interests that are inscribed into the designs of technology (Latour, 1996, 1999; Hanseth & Monteiro, 1998). This is to some extent due to the fact that we lack a common vocabulary to carry the conversation across different communities, eg. educational users and technologists.
Going back to the definition of a framework Wilson et al. states that a framework is equivalent to the concept of a pattern in the software community (2004). They cite Brad Appleton who says
“The goal of patterns [Frameworks] within the software community is to create a body of literature to help software developers resolve recurring problems encountered throughout all of software development. Patterns [Frameworks] help create a shared language for communicating insight and
experience about these problems and their solutions. Formally codifying these solutions and their relationships lets us successfully capture the body of knowledge which defines our understanding of good architectures that meet the needs of their users. The primary focus is not so much on technology as it is on creating a culture to document and support sound ...design.” (Appleton, 2000)
From our point of view the need for "creating a culture" is the crux here. This understanding was strengthened during the aforementioned JISC CETIS conference on Reference Models. By focusing on Reference Models and Services as the key concepts of the e-Framework (Olivier et al., 2005) it is made clear that this is a community effort where the development process is intertwined with Governance and Stewardship processes, according to this diagram explaining how the parts of the e-Framework fit together:
(Figure 4. Diagram of how the different parts of the e-Framework fit together, Oliver et al., 2005)
The JISC Community in the UK are now working on a number of Reference Models using this template (note: http://cetis.ac.uk:8080/elearningprogram/refmodelsITT ):
1) Domain Scope and Aim of Reference Model
A clear definition of the domain-area of the reference model and the aims and scope of the Reference Model.
2) Use Cases and Business Processes for the Domain-Area
This section defines the domain area by breaking it into use-cases and business processes.
3) Identification of the Services Relevant to the Domain-Area
4) Factoring of Services
Each service in the domain’s service-profile should be factored using the following hierarchy:
• Definition of functional scope
• Abstract model of data and behaviour
• Data representation specification (i.e. XML binding)
• Application Programming Interface (API) Specification
• Web service specification (i.e. WSDL)
Implementation of Service Oriented Architectures – a multi-layered process
Service Oriented Architectures are true examples of disruptive technologies (note: Christian Brothers University defines disruptive technology as: "Technologies that enable the breaking of long-held business rules that inhibit organizations from making radical business changes" http://www.cbu.edu/~lschmitt/I351/glossary.htm ). As such one should not expect too much from goal setting exercises like the ones that come from the EU Commission and the Norwegian Ministry of Government Administration and Reform. Institutions resist change, and they will find all kinds of excuses to carry on their software development processes as before (ref. the example given above from the Norwegian Directorate of Education and Training).
The UK led e-Framework endeavour is an example of applying a domain approach to meet the goals set in the eGovernment strategies of the European Union. The work is carried out almost without reference to these pan European visions. The situation i Norway is quite contrary, we have European visions and national goals but very little domain activities (this holds for the educational sector and to some extent for the cultural sector). A central principle in the eNorway strategy is that the different public sectors have full responsibility for their own ICT and service development – and they control their own pace.
It is obvious that the coordinating authorities should extend their support toolbox to speed up the adoption of service oriented architectures. We would think that the conceptual framework developed as a cooperative effort by UK and Australian experts might be a source of inspiration. However, if we have the Norwegian situation in mind, we think would need more than a conceptual understanding of a Service Oriented Framework.
The e-Framework presupposes that there is a community of developers that could be mobilised to build the framework through an open conversation coordinated by JISC, DEST (note: DEST is the Australian Department of Education, Science and Training. ) and eventually other national authorities. There is no such coordinating and funding agencies in Norway. The community of developers is less organised, and do not have a clear vision as to which services that are needed in learning, education and training. We would therefore need more incitements to get the SOA process to roll.
Openness is a key concept in SOA development (open as in Open Source, Open Standards, Open User Interfaces etc.) However, these days government institutions are more and more run according to a private corporations ethos with tight control on information, no exposure of business processes etc. It is therefore a need to (re)define a layered business model of public sector software development within a SOA (note: An example will illustrate the problem we are grabbling with. In 2004 seven work groups under the Norwegian Digital Library Programme were commissioned to come up with recommendations for a comprehensive framework for a digital library in one year. The groups delivered their reports in time. During the twelve months of 2005 the professional conversation in the library community was put on hold, waiting for the recommendations to be processed within the formal authorities. 95 percent of the content of the reports was of strict academic nature and might have stimulated a necessary professional conversation. (Hoel & Kolstad, 2006) ). The IDABC interoperability model does not give guidance to how authorities and institutions should organise their work processes. Semantic, organisational and technical interoperability presupposes to a large extent open processes between systems. However, once you put a governmental institution into the equation you will experience that this kind of systems have a mixture of open and closed processes.
In figure xxx we have identified four distinct layers that fall into two categories. In one way it is Academia meeting Government, which is in fact the case when Ministries of Education are commissioning software for learning, education and training. The problem arises when the red line between the decision making and administrative processes is crossed and the light is dimmed for necessary knowledge processes.
Figure 5. Process layers withing governmental institutions.
We have given examples of projects where knowledge building activities have been dealt with from a decision making logic where control of information has been the leitmotif. For a country with scarce access to experts and a weak professional discourse one year's halt in a field as agile as ICT and learning technologies could be devastating.
With Service Oriented Architectures being a new software paradigm we could foresee that a small country like Norway could be set back if we are not able to discern between the different types of processes. This is because much of the software development has been left to the market, managed by the vendors alone or in a closed relationship between public authorities and a small number of software and consultancy companies. It is not possible to put SOA out for tender as such. Someone has to do some orchestration to come up with the right factorisation of the services that should be developed. And that orchestra rehearsals have to be done in public as a knowledge building exercise, to be able to pick up the right tones and subdue the false ones.
Conclusion
If the EU policies on open standards, service oriented architectures and open sources shall have any impact on the development of technologies for learning, education and training the Ministries of Education have to rethink the way they handle their knowledge building and decision making processes.
Up till now we (to a large extent the vendors) have built individual systems, e.g. Student Management Systems, Learning Management Systems, Test Administration Systems etc. When these systems start to crackle into services the Ministries suddenly are faced with the need to some orchestration. Some of the services are vital to society (e.g. security, privacy, data protection, exam systems etc.) and cannot just be left to "the market" to develop without an open and common architecture.
A common architecture for learning, education and training must be build upon up to date knowledge through research and development activities in close connection with communities of practice throughout the educational sector. These knowledge building processes could collide with the public management culture of educational authorities with the result of open processes being closed for some periods.
To follow up EU and national goals on SOA the government bodies should be very careful to define which processes should be open and which should be closed when they for instance commission reports, put out bids for tender etc. One paragraphs with recommendations in a full report to the Ministry on a technical issue is enough to keep it from the public eye for months. Instead the Ministry should ask for a public technical report and a separate letter of recommendation, to be able to contribute to a knowledge building processes.
To lift the administrative veil from knowledge processes is not enough to harness a service oriented architecture. The e-Framework exercise to forge adequate concepts to grasp the essence of the bricked wall (note: The collection of domain services and common services in the e-Framework is often referred to as "the Wall", see http://www.elframework.org/learning_domain_services and http://www.elframework.org/common_services ) they are building is worth a study for Ministries not participating in this effort. What should be noted is that several of the concepts (e.g. Framework, Reference Model, Design etc.) have a dual nature. They are precise technical concepts that is used in formal models, often expressed in UML diagrammes. On the other side, they are soft concepts that are meant to bridge the user community and the technical community by facilitating the professional conversation between stakeholders. The very fact that communication and dialogue is built into these concepts that are essential to understand what service oriented architectures are, makes it even more important to facilitate development of SOA services as open processes. Closed SOA development is a contradiction of terms.
References
Appleton, B. (2000): Patterns and Software: Essential Concepts and Terminology, available at http://www.cmcrossroads.com/bradapp/docs/index.html#PatternsNutshell, accessed Feb 20, 2006
EU Commission (2006): Interoperability for Pan-European eGovernment Services, COMMUNICATION FROM THE COMMISSION TO THE COUNCIL AND THE EUROPEAN PARLIAMENT, COM(2006) 45 final, available at http://europa.eu.int/idabc/servlets/Doc?id=24117 - accessed Feb 20, 2006
Hoel, T. (2005) Model filtering - how to prevent black-boxing of discourses about learning and teaching"
http://www.hoel.nu/publications/Tore_Hoel_model_filtering_2005-11-28.pdf, accessed Feb 20, 2006
Hoel, T. & Kolstad, A. (ed) (2006) Digitalisering og tilgjengeliggjøring for kultur og læring, Report in Norwegian from a project exploring the challenges of co-operation between the library and educational sectors, available at http://www.abm-utvikling.no/publisert/ovrige/digit_tilgjeng_rapport_web.pdf, accessed Feb 24, 2006
Hanseth, O. and Monteiro, E. (1998, forthcoming) Understanding Information Infrastructure online at http://www.ifi.uio.no/~oleha/Publications/bok.html, accessed November 5th 2005
IDABC (2004): The European Interoperability Framework for pan-European eGovernment Services, ISBN 92-894-8389-X, available at http://europa.eu.int/idabc/servlets/Doc?id=19528, accessed Feb 20, 2006
Latour, B. (1996) ARAMIS or The love of technology, Harvard University Press
Latour, B (1999) On recalling ANT in Law, J. & Hassard, J. (eds), Actor Network Theory and
After, Blackwell and Sociological Review, Oxford, 1999 (ISBN 0 631 21194 2).
MOD (2005a): The use of Open Standards and Open Source in Government, in Norwegian only (Bruk av åpne IT-standarder og åpen kildekode i offentlig sektor, Forslag fra arbeidsgruppe nedsatt av
Moderniseringsdepartementet etter anbefaling fra Koordineringsorganet for eForvaltning, Avgitt 21.06.2005)
Mod (2005b): The eNorway 2009 – the digital leap , available at http://www.odin.dep.no/filarkiv/254956/eNorway_2009.pdf, accessed Feb 20, 2006
Olivier, B. (2005) Why refererence models. Presentation at the JISC CETIS 2005 Conference. Available as Powerpoint presentation at http://www.e-framework.org/events/conference/ , accessed Feb 22, 2006
Olivier, B., Roberts, T. & Blinco, K (2005): The e-Framework for Education and Research: An Overview, A Paper prepared on behalf of DEST (Australia), JISC-CETIS (UK) ,Version R1, July 2005
Star, S.L. & Griesmer, J.R. (1989): Institutional ecology, Translational and Boundary Objects. Social Studies of Sciences. 19, 3
Wilson. S, Blinco, K. Rehak, D (2004) Service-Oriented Frameworks: Modelling the infrastructure for the next generation of e-learning systems, available at http://www.jisc.ac.uk/uploaded_documents/AltilabServiceOrientedFrameworks.pdf , accessed Feb 20, 2006