Forced to Conform ? Using Common Processes and Standards to Create Effective eLearning

Wo rking on multiple large-scale eLearning projects forces teams to try and s t a n d a rdise processes and pro c e d u res. Tools such as XML allow us to manipulate and exploit content in ways previously impossible. Howe ve r, no academic from any discipline likes to imagine that their content is standard. And terms such as 're u s e ' and 'repurposing' make academics even less comfortable. And perhaps they are right. This article describes a formalised development methodology created by one eLearning development team based at the Un i versity of Oxford, designed as a generic system flexible enough to cope with a wide range of subjects and audiences. T h i s paper will also set this development process in the broader world of academic eLearning development across the disciplines, looking especially at the role of s t a n d a rds to consider future directions and the applicability of any deve l o p m e n t methodology to wider learning development contexts.


I n t r o d u c t i o n
ELearning development has long been divided between the world of training with defined job roles and management techniques and the academic world where d e velopment was first pioneered by enthusiasts and only more recently supported by teams dedicated to the use of technology to support a university education.Wi t h larger scale adoption of eLearning by many universities and an increase in fully online programmes a growing professionalism has entered academic eLearning d e velopment; howe ve r, this has not been a straightforw a rd pro c e s s .Processes and practices from the training world are not always appropriate for academic contexts, whilst those from software development are arguably even less so.Yet at the same time there is pre s s u re for academic teams to deliver content to the standard and cost norms of the better re s o u rced training industry.This sits in sharp contrast to the individualistic focus of much academic endeavo u r, where disciplines are clearly defined and the idea of generic solutions to any aspect of academic life is viewed with d i s t rust.(Becher and Trow l e r, 2001) This article describes a formalised development methodology created by one eLearning development team based at the Un i versity of Oxford, designed as a generic system flexible enough to cope with a wide range of subjects and audiences.Although this process has been successfully implemented in several different contexts this paper will address its application only from the perspective of developing largescale eLearning programmes (i.e.full Masters online).This paper will also set this d e velopment process in the broader world of academic eLearning development acro s s the disciplines, looking especially at the role of standards to consider future d i rections and the applicability of any development methodology to wider learning d e velopment contexts.

. Existing development methodologies for eLearning
The best-described methodologies for the development of large-scale eLearning come f rom the US training industry, with most practice being based on variations of the In s t ructional Systems Design.The most commonly used framew o rk is ADDIE, which stands for Analysis, Design, De velopment, Implementation and Evaluation.It is possible to find numerous articulations of this process in literature and on the we b ( Hall, 1997) (Kruse, 2002).Usually re p resented as a cycle, perhaps the main criticism of ADDIE is the lack of flexibility between the stages as well as a tendency to result in identikit solutions.The main alternative is Rapid Prototyping, which could be described as ADDIE for the impatient: at some point all of the main stages Fo rced to Conform?Using Common Processes and St a n d a rds to Manton et al ( 2004) Create Ef f e c t i ve eLearn i n g a re adopted, but at a far quicker pace and with increased iteration between stages.Focussing on less analysis and more consideration of quick prototypes it gives faster initial results to allow clients to understand output better.Arguably a version of this is what most ad hoc academic eLearning development might be said to subscribe to, albeit unknow i n g l y.
So f t w a re development faces many of the same issues as eLearning development and t h e re are many well articulated project management tools to deal with the complexity of the process, which seem to offer concrete solutions to the eLearning world.Examples include PRINCE2 (Office of Government Commerce, 2003) and "extre m e p roject management" (Yo u rdon, 2001).Howe ve r, even the most flexible of these tools tend to be over complex and rigid to be a realistic solution for academic eLearning deve l o p m e n t .
Whilst it would be unfair and untrue to say that there is no exploitation of pro j e c t management techniques or defined development methodologies in eLearning for Higher Education, project management is rarely mentioned directly in many of the b e t t e r -k n own how-to books such as Weller (2002).Fu rt h e r m o re, much has been accomplished in the field using surprisingly ad hoc appro a c h e s .W h e re an a c k n owledged pro c e d u re is espoused this has tended to call upon variations of the tools described above, with constant mutation to fit the idiosyncratic needs of a c a d e m i a .Perhaps the most well known and well developed deve l o p m e n t methodology in academia is that of the UK's Open Un i versity for print materials, ye t it is clear that even that institution has yet to transfer this to a fully art i c u l a t e d p rocess for technology centred courses.
Un i versities are charged with delivering students high levels of educational experience with a work f o rce more usually employed for their re s e a rch rather than their teaching expert i s e .Disciplines tend to be fiercely independent and will not accept being forced to conform to re q u i rements they inevitably see as externally i m p o s e d .So the question remains -how do we get the benefits of we l l -d e ve l o p e d p roject management methodologies whilst retaining the re q u i rements of a more individualised, academically sensitive approach, especially when academics may not actually be able to articulate (or even know) what they want?The challenge has been to create a development process that saves re i n venting the wheel for eve ry course but which also allows TALL to be flexible enough to create the best eLearning course possible for each scenario encountered, both in terms of audience and discipline.The resultant development process is the culmination of 7 years experience of course development and continues to be improve d .Em e r g i n g f rom the formalisation of ad hoc processes, a pick and choose approach to existing methodologies, and refinement over years of implementation, the process is the end p roduct of close work and reflection primarily between project managers, we b d e velopers and learning technologists.Constant evaluation of real successes and f a i l u res as well as a cre a t i ve approach to the adoption of innovation has resulted in a p rocess that is owned by the whole team.

. Standard development process
The TALL approach to developing online learning has always put the teaching and learning rather than the technology at the centre .At the same time in pure l y practical terms it is often technical, financial and time considerations that drive the most efficient course development pro c e s s .The challenge was to find a way to get the best from all worlds.The TALL development process (Fi g u re 1) gives an ove rv i ew of the total system; howe ver it must be noted that this can change in detail betwe e n

Fo rced to Conform? Using Common Processes and St a n d a rds to
Manton et al ( 2004) Create Ef f e c t i ve eLearn i n g p rojects and over time where appropriate: in fact the ability to customise is a core f e a t u re .

Fi g u re 1: The TALL De velopment Pro c e s s
Journal of In t e r a c t i ve Media in Education, 2004 (14) 1 Note in TALL a learning technologist is an expert on the pedagogical aspects of online learn i n g (analogous to an instructional designer in training terms) not the more technically focussed person a learning technologist often is in many unive r s i t i e s .Education, 2004 (14) Page 6

Journal of In t e r a c t i ve Media in
Key to understanding this flowc h a rt is a clear articulation of roles.For all pro j e c t s t h e re will be a) a TALL team including: a project manager, web developers and a learning technologist 1 among others and b) an academic team (possibly from another institution) usually including: a programme dire c t o r, academic authors, a pro j e c t manager and a programme manager.T h e re is a subset of other roles that can be p e rformed or outsourced by either team.These include quality assurance (QA ) , c o p yediting, copyright clearance and graphic design.The most intense collaboration, especially in the specification stages is between the project managers, the learning technologist, the programme dire c t o r, and academic authors calling upon other ro l e s (such as web development) where necessary to answer specific questions.
What follows is a more detailed examination of key areas of the process (Fi g u re 1), focussing on how it has been designed to take advantage of standardised pro c e d u re s and tools whilst being flexible enough to work realistically in the world of higher e d u c a t i o n .
The specification stage is the foundation of the entire process, laying the basis for a successful pro j e c t .It is here that the whole team considers and agrees the shape of the programme and re c o rds all the decisions.This is facilitated by a simple set of forms to be completed, ensuring that all the areas that need to be articulated for a successful eLearning programme have been discussed and that decisions taken can be accessed and understood by all invo l ved throughout the process.Depending on the s i ze of the programme to be developed the specification documentation can consist of one or many documents, each specifying the course in greater levels of detail.These stages are outlined below.

. 1 Stage 1: Programme specification
The programme specification takes place at course director level and may or may not i n vo l ve the academics who will be authoring and/or delivering the material.T h i s document contains ve ry little that isn't found in programme level documentation for any award-bearing course: despite this it is surprising how often the process is new to teams TALL works with.At this stage the specification is so generic that we use the same document for eve ry programme, collecting standard information.Much is p u rely structural such as Programme title, Programme length, St ru c t u re (number of modules and credits etc), De l i ve ry (period and study hours) and Assessment.Mu c h is content focussed, such as Programme description, Aims, Ob j e c t i ves, and Learning

Fo rced to Conform? Using Common Processes and St a n d a rds to
Manton et al ( 2004) Create Ef f e c t i ve eLearn i n g outcomes (knowledge, understanding, skills).Howe ver we also re q u i re statements about desired pedagogical approach and target audience, arguably the two factors that most affect eventual programme design.
The programme specification does not contain a rationale for developing the p rogramme as it is assumed that this has been completed before TALL becomes i n vo l ved, howe ver if this has not been undertaken, we strongly recommend this stage as a core re q u i re m e n t .

. 2 Stage 2: Module specification
The module specification is the most intensively used document in the whole p rocess; initially used to articulate the learning design, it later acts as the blueprint for the developers and the definitive source of information about the module.Fundamentally it is the module specification that defines what is to be built.
Due to the many roles that this document has to perform it is customised for each individual programme, resulting in the creation of a document optimised for maximum utility and usability.The document and the methodology it encapsulates must be flexible enough for any discipline and pedagogy whilst capturing the information the TALL development team know is re q u i red for any eLearning course f rom a developmental and standards perspective.As such, module specification initially happens on two levels, that of designing the form and that of filling out the form, with much iteration between the two.In a programme with many modules, the format of the document itself will change considerably during the first module, but by later modules will be set in stone.By the end of the process ownership of the specification document belongs as much to the academic team as the deve l o p m e n t t e a m .
At the simplest level the document format is articulated in terminology; 'module' has been used here as a generic term, but if the team want to call their standard unit of study a course, a unit, a section or a topic this is possible.Howe ver it is not just the terminology that can be adapted for individual courses but also more semantically and pedagogically significant information.Thus if a problem based learning model is being pursued one field is likely to be 'the problem', while a collaborative focused p rogramme may have a field for group discussions.This can be summed up in the o b s e rvation that the module specification process is aimed at capturing the learning design of a module as much, if not more, than the course content cove r a g e .C l e a r l y this is not a straightforw a rd process and generally re q u i res close work between a learning technologist and the academic team to establish re q u i rements and suggest h ow they can best be realised through the tools and the enviro n m e n t .
As well as capturing the learning design the module specification also starts to re c o rd much of the practical information re q u i red to build a course online: including both metadata (IMS, 2001) and re s o u rce information (including multimedia and copyright), each of which have major time and cost implications for any eLearning p ro j e c t .
The specification of the learning design in advance of the course content authoring is also core to exploiting the best from the medium of technology delive red learning.
It allows authors to articulate their wishes independently from the medium, but also a l l ows the learning technologist and web developers to interrogate these wishes at an early stage, advising on what is possible or desirable.In addition, it provides a realistic ove rv i ew of the module for a project manager to use as a basis for cost and time calculations.Initial ideas may evo l ve, but with both sides clear about how and why this has happened and what the implications are for the real end pro d u c t .

. 3 Stage 3: Content and resource production
If the specification stage of the project has been undertaken successfully then content and re s o u rce production should be simple, since the learning has been planned and all that is left is to write the words and collect the re s o u rc e s .In e v i t a b l y, this is far f rom the case, and this is the stage most responsible for project slippage.
Authoring within the process is currently controlled using an MS Wo rd template customised with macros (See Fi g u re 2: The authoring template) which can subsequently be conve rted into XML (extensible mark-up language) for content d e ve l o p m e n t .This process offers several adva n t a g e s : • Metadata is collected as the content is authored, improving the standard of the information and its conformity.
• The fields available in the template can be customised to fit the pedagogical model decided at the specification stages, making content more likely to be a p p ropriate for online learning and (re l a t i vely) less likely to fall into the trap of being a text book online; terminology is also adjusted to suit the discipline and thus the authors.
• Authors are able to semantically describe their content, which reduces erro r s b e t ween the authoring and development stages.

Fo rced to Conform? Using Common Processes and St a n d a rds to Manton et al (2004) Create Ef f e c t i ve eLearn i n g
Journal of In t e r a c t i ve Media in Education, 2004 (14) • Significant automation of the content production pro c e s s .
The MS Wo rd template (see Fi g u re 2) is admittedly far from perfect and eve n t u a l l y we would hope to have a W Y S I W YG tool that allows authors to pre v i ew content; h owe ver with currently available tools it offers the best compromise betwe e n functionality and usability.This is ve ry much a work in pro g ress and the template is updated with eve ry pro j e c t .

Fi g u re 2: The authoring template
The success of this stage is largely dependant on the iterative nature of the d e velopment and the close working relationship between the content authors, academic re v i ewers, learning technologists and project managers.

. Stage 4: Web development
The web development stage of course production necessarily remains flexible to accommodate the specific needs of individual projects and is finalised following the p roduction of the course specification documentation.
Web development is a complex process but it can be summarised as follow s : The manifest file describes the stru c t u re of the content and incorporates much of the course metadata.
The programme stru c t u re is determined by the programme specification and, in this instance, adopts the following hierarc h y : Pro g r a m m e > Mo d u l e s > Un i t s > Pa g e s Once a stru c t u re has been determined and agreed it is incorporated into an IMS Content Packaging manifest file "imsmanifest.xml".The manifest, created using the R E LOAD editor (http://www.reload.ac.uk), describes the stru c t u re of the course in a t ree of "item" elements, with corresponding "re s o u rce" elements.Individual pages in the course stru c t u re have re s o u rce elements which identify the XML files that those pages will be created fro m .
Metadata for the programme, module and unit levels are also taken from the p rogramme specification, and inserted into the manifest as IMS Metadata.Me t a d a t a for pages are dealt with in section 4.3 below

. . 2 Convert content to XML/HTML
Content, authored in the MS Wo rd templates, is conve rted in two stages: first to XML following a TALL designed schema and secondly from XML to HTML.T h e details of this process are outlined below :

Conversion from MS Word to TALL XML:
The Op e n Office.orgoffice suite 2 uses a standard XML document format 3 , and has flexible export facilities 4 .Each MS Wo rd document is loaded into Op e n Of f i c e .o r g and exported via an XSL transformation that, in turn, recognises the MS Wo rd styles and generates the corresponding TALL XML mark -u p. Howe ve r, to compensate for omissions in the MS Wo rd template, errors in the transformation and to incorporate editorial changes that are unrelated to conversion the XML usually re q u i res furt h e r hand editing.

Fo rced to Conform? Using Common Processes and St a n d a rds to Manton et al (2004) Create Ef f e c t i ve eLearn i n g
2 Op e n Office.orgoffice suite: http://www.openoffice.org/

h t t p : / / w w w. o a s i s -o p e n . o r g / c o m m i t t e e s / t c _ h o m e . p h p ? w g _ a b b re v = o f f i c e 4 h t t p : / / x m l . o p e n o f f i c e . o r g / f i l t e r s . h t m l
Journal of In t e r a c t i ve Media in Education, 2004 (14)

Manton et al (2004) Create Ef f e c t i ve eLearn i n g
Journal of In t e r a c t i ve Media in Education, 2004 (14) Page 11

Conversion from TALL XML to HTML:
C o n version of the TALL XML is managed using the Apache Ant build tool [5].T h i s enables all conversion tasks for the whole programme to be invoked with one c o m m a n d .Once invoked, Ant examines the IMS manifest file to determine which XML files need to be processed.The XML files identified are transformed into HTML, and copies non-XML re s o u rces such as images, video clips, and PDFs fro m staging directories into the HTML dire c t o r i e s .
This approach supports the generation of multiple instances of the course from the same course content XML.

. . 3 Create content package
Ant is used to manage the building of content packages, again using the IMS manifest file to marshal the pro c e s s .
The ideal would be to generate fully SCORM and IMS conformant content packages.Un f o rt u n a t e l y, to ensure that the packages will deploy corre c t l y, any problems in the LMS implementation of content packaging have to be accommodated.One example of this is that the UKeU's LMS re q u i res that in a programme package (the top h i e r a rchy level), modules (the second level) must be packaged individually.
A fully conformant packaging process is under development in parallel to the more accommodating ve r s i o n .
The manifest created using RELOAD editor intentionally only contains the p rogramme hierarchy and some metadata.Using Ant, a large number of steps re q u i red to build a complete package are automated and invoked with one command.
The content package build process performs the following operations: 1. Splits the programme modules into individual manifests (to accommodate a LMS implementation issue).
2. Modifies re s o u rce elements to refer to HTML files rather than XML files.
3. Processes each XML file, adding dependency and re s o u rce elements to the manifest for images and other re l e vant files.

Manton et al (2004) Create Ef f e c t i ve eLearn i n g
Journal of In t e r a c t i ve Media in Education, 2004 (14) Page 12 4. Extracts metadata from the manifest and XML files, creating separate metadata files (as mandated by the LMS).
5. Re-transforms any XML files where the HTML is out of date.
6. Copies HTML and images, etc. to the packaging are a .
7. Creates zip files of each module.
8. Creates a final programme level zip, which is the completed content p a c k a g e .

. . Upload content to platform
Once created, the package is uploaded to the LMS and deployed using the platform's normal mechanism for handling content packages.
Each of these four development steps will typically be performed one module at a time, as the content becomes available.Periods of re v i ew throughout deve l o p m e n t inform modifications to the process, course content and design.The XML based p rocess allows changes to be made quickly and easily, even significant changes to the whole course.

. 5 Stage 5: Review and publication
Once a unit of study has been processed and loaded to the platform or learning e n v i ronment, the programme director and authors have the opportunity to see their content in its final format.So far, they have been working on Wo rd templates with little formatting or style.This stage of the process gives them a chance to see the full module or unit stru c t u re, how the metadata is presented in the learning enviro n m e n t and the full range of working re s o u rces such as images, audio, video, flash animations, etc.
The programme director and authors compile a list of changes, which at this stage should be reduced to minor errors, and not new additions.In t e r n a l l y, TALL will c a r ry out a click test to make sure all re s o u rces, media and links are all working.T h e web development team subsequently make the necessary changes and reload the package to the platform for the programme director's final re v i ew and sign off for l a u n c h .

Fo rced to Conform? Using Common Processes and St a n d a rds to
Manton et al ( 2004) Create Ef f e c t i ve eLearn i n g Journal of In t e r a c t i ve Media in Education, 2004 (14) Page 13

. D i s c u s s i o n
With courses becoming ever more complex the TALL development process evo l ve d out of a ve ry real need to control the complicated set of interrelated factors that affect the success of an eLearning pro j e c t .The documentation and pro c e s s a rticulated above is designed to give a picture of the project shared by all part n e r s , but it is the process of creating this documentation that is where the strength of the system re s i d e s .At eve ry stage the process is flexible, contingent on the re q u i re m e n t s of an individual programme.Initially what the documentation provides is a checklist to ensure that all the basic issues have been considered in the context of the re a l d r i vers that will shape a pro g r a m m e .
Howe ve r, much of the success or failure of a project is based on eliciting a clear a rticulation of what an academic team wants to achieve in teaching and learning terms, independent of technology, only then taking this information and using the d e velopment process to build a course in the most efficient manner possible.T h e dialogue undertaken with project managers and programme directors is ve ry differe n t f rom that with academic authors, the space within which each role is allowed to shape the programme strictly contro l l e d .Thus another key role of the documentation at the specification stage is as a starting point against which project managers and programme directors can re a c t .Teams who have no opinion when pre s e n t e d with the idea of an eLearning programme suddenly have many when a generic p rogramme is suggested against which they can position themselve s .
Other factors come into play when working with content authors.Thus, in tandem with the documentation, standards such as IMS metadata are presented to authors fait accompli as part of the template.In this way the specification documents and templates guide authors to create a pedagogically sound online learning experience, a novelty for most.Undoubtedly the implied ort h o d oxy of a standard which must be complied to can be ve ry helpful in encouraging desirable behaviours in academics who might otherwise reject being contro l l e d .An interesting aside is that although for a long time TALL encouraged greater freedom among academics to define learning design (not wanting to encroach on their expertise) in programmes where the learning design is defined by the academic team at an early stage and authors have little say by the time their module is developed we have found that they seem to welcome the certainties of a defined pedagogical model.The one concern with this is that they will then go on to assume that all eLearning should be done according to the model with which they are familiar, which is far from the case.

Manton et al (2004) Create Ef f e c t i ve eLearn i n g
Continued development of effective large-scale eLearning courses necessitates we l lunderstood and articulated development methodologies.Undoubtedly the academic context re q u i res more sensitivity to individual re q u i rements than many other eLearning spheres, yet it is safe to say that these are not mutually exc l u s i ve .T h e reality is that howe ver academic partners may choose to articulate their wishes for an online course, the reality of the development comes down to a limited set of tools and processes and the challenge is finding the best way to collect and re c o rd all the information that makes up a successful course.One potential tool for the future is IMS Learning Design (IMS, 2003), designed to encapsulate any pedagogy using the model of people undertaking activities with re s o u rc e s .As yet few are art i c u l a t i n g their course design in these terms, and there are not any widely available tools to d e l i ver learning design courses.Yet this framew o rk does point to the ability of complex learning ideas to be encapsulated in simple machine readable terms in the near future .The TALL development methodology, and accompanying documentation does not yet abstract to this degree, but on a fundamental level is trying to c a p t u re the same information for identical ends.Fu rt h e r m o re this is curre n t l y happening in a way that is actually usable by teams now to output courses in a format that can be read by today's virtual learning environments, IMS Content Pa c k a g e s (IMS, 2003).

. C o n c l u s i o n
The process examined above is far from complete and will undoubtedly change again in the years to come, howe ver any major changes are now more likely to come fro m d e velopments in standards, theory or technology rather than in light of curre n t f a i l u res.Indeed, this process has been adapted applied to several different contexts.
As described above the process can be applied to large-scale project or course d e velopment.Howe ve r, TALL has also applied a modified version of the same pro c e s s to the development of short, personal development courses which only re p resent 100 study hours.Howe ve r, the fundamentals of the process remain unchanged (specific a t i o n > d e ve l o p m e n t > re v i ew > d e l i ve ry ) .
The process has emerged from a need to control complex technical deve l o p m e n t p rojects where the ove r a rching re q u i rement has been to create effective learning.Creating a process that allows the most efficient navigation between these two e x t remes has been fraught, but currently works we l l .New standards such as IMS learning design offer for the first time the possibility of capturing pedagogical re q u i rements in technologically comprehensible ways, but there are not yet the tools to implement this for large-scale course deve l o p m e n t .Howe ve r, as a 'learning design' is the core to the information collected at the specification stages of the TA L L d e velopment process, we would like to undertake further work to explore the IMS learning design standard as a viable semantic framew o rk and development tool for the future .
Whilst it is easy to see technology and the differences within disciplines as the challenges to defined processes in online learning development, what is interesting is that it is actually the elicitation of teaching and learning aims that is the hardest to a c h i e ve.Offering a methodology that elucidates these aims means that academics are not forced to conform, but rather they are persuaded to participate in the d e c o n s t ruction of their teaching with the aim of re c o n s t ructing a pedagogically sound online alternative.Indeed, provision of a strong framew o rk appears to be p o s i t i vely welcomed by most academics TALL works with, as long as it is clearly situated in a negotiated process over which they feel and have some ow n e r s h i p. Journal of In t e r a c t i ve Media in Education, 2004 (14)

R e f e r e n c e s
Be c h e r, T. and Trow l e r, P. (2001) Academic Tribes and Te r r i t o r i e s.Open Un i ve r s i t y Press: Buckingham Hall, B. (1997).Web-based training cookbook.John Wiley & Sons: New Yo rk, NY IMS (2003) IMS Content Pa c k a g i n g h t t p : / / w w w.imsglobal.org/content/packaging/index.cfmvisited 04/12/03 IMS (2001) IMS Learning Re s o u rce Meta-data Sp e c i f i c a t i o n h t t p : / / w w w.imsglobal.org/metadata/index.cfm visited 01/12/03 K ruse (2002) In t roduction to In s t ructional Design and the ADDIE Mo d e l h t t p : / / w w w. e -l e a r n i n g g u ru .c o m / a rt i c l e s / a rt2_1.htmvisited 01/12/03 Office of Government Commerce (2003) Prince 2, http://www.o g c .g ov.u k / p r i n c e / visited 01/12/03 We l l e r, M (2002) De l i vering Learning on the net: the why what and how of online e d u c a t i o n.Routledge Falmer: London Yo u rdon, E. (2001) Managing Hi g h -Intensity In t e rnet Pro j e c t s.Pre n t i c e -Hall: London