Skip to content.
Sections
Personal tools
You are here: Home » Features » Report from the JISC e-Framework for Education and Research workshop

Report from the JISC e-Framework for Education and Research workshop


Last modified 02 Apr, 2007
Published 02 Apr, 2007
A report from the JISC e-Framework for Education and Research workshop which took place on February 12th at the Aston Business School Conference Centre where projects were updated on the latest e-Framework developments.

Introduction

The e-Framework for Education and Research is an international initiative supported by JISC in the UK, DEST in Australia, SURF in the Netherland and New Zealand’s Ministry of Education. Over the last three years The JISC has funded a large number of projects in the e-learning e-science and e-administration domains all guided by the principals of the e-Framework.

This workshop gathered together over 80 invited delegates from these projects to discuss the latest e-Framework developments.

Welcome: Susan Kumari

Susan Kumari the JISC programme manager for the e-Framework introduced the day, and she outlined the overall aim of the day which was to provide;

  • an understanding of the goals of the e-Framework [1]
  • an awareness of the approaches involved
  • the means by which those approaches relate to projects in general and to your project(s) in particular and
  • how to use and contribute to the e-Framework (SUMs and Services).

The day was organised into a number of parallel presentations aimed at projects new to the e-Framework as well as surgeries for projects wanting more practical advice on their work. Seven topics were covered by both presentations and surgeries:

– Domain Modelling

– Service Usage Models

– Process Modelling

– Agile Development

– Managing requirements through service specification

– Using and contributing to the e-Framework

– Scenario based design

Susan also took the opportunity to show participants the new e-Framework animation [2] which explains the purpose of the framework by using a buildings metaphor. Delegates were encouraged to show the animation to their senior managers to explain service oriented approaches and web services.

e-Framework Overview: Bill Olivier

Next up, Bill Olivier, Director of Technical Development at JISC explained the principles behind the e-Framework, and how they fit with the JISC strategy of providing world class leadership. The e-Framework can be viewed as both a co-ordination tool for JISC and its international partners as well as a knowledgebase for projects [1]. From the technology perspective the two guiding principles are user engagement and open standards. Bill reminded the participants that because JISC is funded through top slicing of funding council money for UK institutions, it is required to demonstrate both ‘value for money’ across the sector as well as sustainability.

Institutions face the dilemma that some of their software needs are common with those of industry while some needs are unique to universities. This means that they have to choose between buying commercial systems and developing their own bespoke ones. Bill said in truth neither solution is satisfactory.

The idea behind the e-Framework is that it removes the duplication of functions that exist in many large institutional systems by extracting the common functions and removing the application layer.

Bill discussed IDEO’s rapid prototyping methodology [3] and mapped it onto the JISC project cycle wondering how the process could engage users more in both the selection of problems to be addressed and the prototyping phase. JISC is working with the JISC CETIS, HEA and AUA (Association of University Administrators) communities on how to identify common priority areas across the sector.

The aim of domain mapping and modelling is to identify reusable domain knowledge that is relatively stable. Having a stable domain map leads to a consensus on the key problem areas, so that developers can develop software in the agreed problem spaces.

Bill concluded by outlining how domain maps, workflow (as-is and to-be), case studies, application layers and service usage models (SUMS) fit together to describe a particular domain space in detail [4], [5].

The next presentations were given in parallel sessions.

Domain Modelling: Bill Olivier

Bill began by defining a domain as “a recognisable area of work or activity”. Within an institution the e-Framework identifies five broad domain areas;

– Learning and teaching

– Research

– Libraries

– Administration

– Information Services

And like a set of Russian dolls each of these areas unpacks – so for example sub-domains of learning and teaching might include:

– Course management

– Course preparation

– Student enrolment

– Course delivery

– Assessment

Each of these sub domains will have a set of users, practises and expertise, along with its own vocabulary, professional bodies etc, which cut across different institutions.

Bill referred to Enid Mumford’s work [6] on the common problem of co-designing software with practice. The problem is typified by the following statements:

  • “Users don’t know what is feasible”
  • “Technologists don’t understand the practises”

Bill explained that this was the basis for JISC’s funding of the Reference Model projects in the e-learning space [7]. These projects took different approaches, some focussed on the domain map development, some on workflow, some on how things are now and others on how things could be.

The elements of a domain map include:

– Domain context model

– Stakeholder analysis

– Goal and Function

– A Domain Information model – the information entities

– A Domain system model – systems and the information exchange

– A set of scenarios

Bill rounded up his talk with a useful table metaphor to describe a domain map, the table top having a shallow and broad surface, and “legs” that drill down into priority areas for more detailed models. Ultimately the table would end up with multiple legs for each priority area.

A fuller explanation of domain maps can be found in Bill’s paper: Domain, Process and Service Usage Models [8].

Service Usage Models: Phil Nicholls

Phil is the UK editor of the e-Framework and in his talk he explored how the e-Framework took a service oriented approach and is aiming to offer benefits at all levels from funders and institutions to developers. He looked at how a typical university environment consisted of systems such as Learning Management, library and student records and how these often existed in silos, and could not ‘talk’ to each other, but that workarounds develop which tend to involve humans. There is an increasing need for interoperability but ‘common standards are not common and often not standard’. He discussed how the silo model impacts on funders and leads to duplication of effort, replication of data, and makes it impossible to implement the single sign-on.

A service oriented approach addresses some of these issues and he explored how services can be used to build applications, and how the e-Framework is building a knowledge base which will collect information about services into one place at a number of levels from non-technical descriptions through to details on models and workflows. One of ways in which the knowledge base will develop is though the creation of Service Usage Models or ‘SUMs’ which will describe how a number of services may be used together, and he provided more detail on these, and the components of the e-Framework in his second talk which is described below.

The e-Framework is not a static or prescriptive blueprint and aims to provide components for architectures but not the architectures themselves. For institutions it makes it possible to see what others have done, and can inform strategic thinking, for developers it provides a place to find descriptions of services. One of the benefits of this approach is that the community will contribute to and populate the e-Framework, and it will enable new communities to emerge and will inform future areas of work.

For more on this see Phil’s slides [4], [5].

Process Modelling: Balbir Barn

Balbir Barn from Thames Valley University ran the COVARM reference model which modelled the course validation processes in universities. Balbir talked about how to go about process modelling and set current thinking in the context of business process modelling theory over the last two decades. Business Process Modelling is regarded as the third wave of development and focuses on the principles of process design, enactment and monitoring in a continuous lifecycle. In this system developments are incremental and no information is ever lost. The core modelling concepts are: – the organisational unit

– process

– activity and

– role

Processes can be broken down into the following types:

– core processes

– supporting processes

– process patterns (e.g. handling an insurance claim)

  • event driven processes (eg end of year accounts)
  • cycle driven processes

Balbir went on to give an overview of the two main notations, namely UML and Business Process modelling notation (BPMN). Each has its own elements, strengths and weaknesses, e.g. communication between processes is hard to capture with UML.

In conclusion Balbir suggested that the e-Framework needs a set of guidelines about which modelling methodology to use. However a question was raised at the end about to whether higher education was like business and therefore whether modelling processes that have been developed for business are, in fact, appropriate for higher education? [4], [5].

Agile Development: Scott Wilson

Scott discussed agile development and provided a series of issues, pointers and advice for discussion –

The approach you take can be determined by whether a) you know the user and the process involved, in which case you can work directly with them, or b) it is a new process with unknown users in which case a different approach will be needed. He discussed modelling methodologies and said ‘UML is not a religion’, these tools are needed for communication, it’s therefore important to stick to basics so that people can understand them and they can be a tool to help people think. His suggestions included using existing experience -

  • if market leaders have done a great deal of market research then ‘copy’

– JISC funded projects are small and market research is out of scope for them

  • reuse techniques and common sets of solutions
  • from the experience of Web 2.0 – deal with one problem well and use generic solutions for everything else
  • encourage participation

Design the user experience in the form of Powerpoint or wireframes then ‘fill in the code’, ‘at least the Powerpoint won’t get confused with the product!’ Getting the user experience right is the critical factor and will determine if something gets used,

Agile development philosophy requires that you –

  • release early and often
  • ensure that you have small teams with high levels of regular contact – this is of critical importance
  • stay close to the prototype – don’t drift
  • release regularly and be clear what people can expect
  • focus on small, simple, incremental things that work rather than start complex – ‘complex systems that work always evolve from simple systems that work’ – avoid the ‘inescapable morass of unsolvable problems’

The development cycle of a project never ends [9] [10]

Some issues were raised by the audience –

  • How does this approach reconcile with the JISC projects funding model?
  • Developers don’t like to expose themselves and that may a problem with this approach?
  • This approach maybe good with a small local user community but more problematic with a wider scale
  • This model of early release may not work well with large open source projects
  • Need to brand clearly – ‘this is ‘beta’ ‘
  • May be difficult in an institutional context where individuals have expertise in a particular thing, though there maybe development teams in HE?

The audio files and Scott’s presentation are available on the JISC site[4],[5].

Managing requirements through service specification: Hilary Dexter

Unfortunately we didn’t manage to see Hilary’s talk. This brief summary comes from Neil Fegen of JISC CETIS:

This talk focussed on how to develop use cases and the processes involved. The talk covered the following areas: * Requirements management

  • Use cases and workflows
  • Service development process
  • Mapping use cases to components and services
  • Supporting the developer.

Hilary used a library example to illustrate structuring a use case document, drawing on eight key points:

  • Brief description (purpose of use case)
  • Basic flow of events (what actor and system does - a dialogue)
  • Alternative flows (alternative behaviour)
  • Preconditions (system state prior to use case)
  • Postconditions (state system can be in)
  • Extension points
  • Special requirements
  • Additional information.

Hilary’s talk is available on the JISC web site [4] and the audio file is also available [5].

Using and contributing to the e-framework: Phil Nicholls

The e-Framework was established with the belief that SOA will bring benefits to funders, institutions and developers. Benefits to funders by establishing a rich ‘map of developments across domains’, to institutions by enabling true interoperability between systems and to developers by providing clear descriptions of the problem space and the services it contains.

Phil explained that true interoperability could only be achieved when “services have well defined interfaces”. The e-Framework site knowledge base does this using the following elements:

  • A non technical description ‘what service x does’
  • A technical description ‘how to write an x service’
  • Combinations ‘if x and y are used together you can….’
  • Details of the actual service
  • Details of the models and workflows.

For clarity the e-Framework has adopted the following terminology:

  • Service Genre – the description of the service
  • Service expression – a standards based and specified way of doing something
  • Service Implementation – the implementation of an expression e.g. a toolkit
  • Service instance – the actual bit of software produced
  • Service Usage Models – describe a number of services and can be technical or narrative
  • CORE SUMs – recurring service combinations e.g. ‘Validate-Enhance-Transform-Operate’

Next, Phil outlined how projects might contribute their work to the e-Framework web site, and what areas different type of projects would fall into. For example a toolkit project might describe a new service, whereas a demonstrator might be able to describe how more than one service could be linked i.e. provide a SUM. Submission template are available on the web site, and Phil estimated that it would take projects 1-2 days to complete a service expression and 3-5 days to create a SUM. Phil will be working with existing projects to help them submit to the web site. The benefit for projects is that it will allow their work to be found by others working on similar problems in the future [4], [5].

The issue of licensing was raised in the questions at the end of Phil’s talk. It is clear that licensing issues have a major impact on projects and therefore need to be resolved at the beginning of a project if the work is to be publicly available at the end.

Scenario- based design: Chris Fowler

Chris Fowler rounded off the day with his keynote on scenario based design. Chris described work his Chimera team at the University of Essex have been doing in developing scenarios. Scenarios have different meanings in different contexts e.g. the military and the theatre. Common elements include

  • A story (narrative)
  • Bounded and scoped
  • Actors and activities
  • Timeframe
  • For a given purpose – communication, analysis and decision making
  • As is or To be

Chris described the SUNA (Scenario Based User Needs Analysis) method which uses a series of workshops to develop scenarios. The first workshop gathers together a team to focus on the problem and the scope of what the scenarios should be about. One individual then writes the scenarios (with the help of a bottle of wine!) and then the group comes together for a second workshop to review the scenarios and identify the user needs and appropriate technologies.

Chris concluded by saying that the key features of good scenarios are that they are; valid, grounded, realistic, consistent, coherent, useful and economical [4],[5].

There was some discussion afterwards about the difference between a use case and a scenario. Which raises the whole question of how models can be used to bridge the gap between modelling what users do or want to do and the generation of precise representations that developers can use.

Conclusions

This workshop was the first to invite projects funded by JISC across the domains of e-learning, e-research and e-admin which could contribute to the e-Framework. It was very ambitious and aimed to convey background information and advice along with surgeries for individual projects.

Throughout the day some interesting, and sometimes challenging, questions and issues arose which demonstrated the importance of having such an event -

Specifically related to the e-Framework –

  • As much of the work is focussed around domains – Sarah Davies (JISC programme manager) wondered to what extent these domains actually existed. Eg. There is a group with interest in the e-assessment space – but what about assessment more widely?
  • Should we be using UML to model everything in the e-Framework space – doesn’t that lead to an attempt to model the world – both impractical and inappropriate?
  • And do different modelling approaches need to be applied in different contexts – ie agile development in the more innovative areas and more formal approaches for institutional processes?
  • How can work taking place across such a wide range of areas and domains be integrated from the point of view of the e-Framework?
  • And in turn - how can this work inform the work of JISC programmes such as the e-Learning Programme?
  • This is so new and developmental – how can the vision and expectations be balanced with realistic timescales?
  • How can awareness be raised and debates be aired, and yet practical guidance be provided for projects?

For us this last point is fundamental. We must remember that in education we are still in the early days of SOA developments and implementation, and it is important to maintain a balance between a concrete framework and being able to explore a range of approaches.

The workshop was followed by a second day which focussed on workflows with presentations from individual projects. The aim was to see if and where there are commonalities between workflows and to see if there are any common points between domain specific workflows. Presentations from the ISIS/ASSIS; RepoMMan, Human Collaborative workflow; ePHPX (escience), COVARM and Kuali project demonstrated the diversity of workflow issues and solutions being used.

Future workshops offering practical advice and guidance to projects will be held later in the year.

References

[1] e-Framework for Education and Research web site

[2] e-Framework for Education and Research animation

[3] IDEOs rapid prototyping methodology

[4] Presentations from the e-Framework workshop

[5] Audio files of the presentations from the e-Framework workshop

[6] Enid Mumfords web site

[7] e-Learning Reference Model Projects

[8] Bill Olivier, Domain, Process and Service Usage Models

[9] Book - Information Architecture for the World Wide Web (Paperback) by Louis Rosenfeld (Author), Peter Morville (Author)

[10] Ambient Findability by Peter Morville (Paperback - 11 Oct 2005)

Acknowledgements

Thanks to Sheila MacNeill, Neil Fegen and Susan Kumari for help with compiling this report.

 

Supported by JISC Supported by CETIS
Powered by Plone