Despite the prospect of a 4am start, the ELF Developers' Forum in Glasgow promised a great opportunity to update other JISC projects on XCRI's aims and progress to date. Jointly-hosted with the CETIS Assessment SIG, Sam (Easterby-Smith) and Rowin (Young)'s carefully-organised two day event was themed and streamed to provide something for all participants. XCRI's interests were best served by the ELF Developers' Forum and an impromptu curriculum modelling workshop on the afternoon of the first day.

In the morning session, Sam explained how the shift in emphasis from ELF to e-Framework reflected the growing potential of a service-oriented approach across the education sector, in e-Learning, e-Research and e-Admin. The middleware from Middle Earth gags might now be history, but it was apparent that a lot of current enterprise integration work is better classed as e-Admin than e-Learning, particularly work in the area of person, group, membership, grades and curriculum services.

Sam's talk was followed by Warwick Bailey, founding Director of Icodeon Ltd, who reported on his investigatory/prototyping work using BPEL (Business Process Execution Language) for orchestrating assessment web services within the University of Hull's ASSIS project. The ASSIS project aimed to provide an adaptive assessment service by orchestrating assessment, sequencing and content web services. Warwick outlined a 9 step process for creating the composite "adaptive assessment" web service:

1. Define partners: composite ASSIS service + Content Packing service (CPRun) + Simple Sequencing Service (SSRun) + Assessment Service (QTIRun)
2. Declare partners in XML
3. Define workflow (UML Sequence diagram proved useful)
4. Write down steps
5. Declare process using BPEL activity constructs
6. Add business logic
7. Use XPath to pull out required data from web service calls for use in business logic (eg in switch statements)
8. Validate XML process definition
9. Deploy XML to web service orchestration service

Wilbert Kraan provided a briefing on the IMS General Web Services initiative: an attempt to promote greater interoperability between web services in the education sector by controlling versions of WSDL, SOAP, UDDI, XML, HTTP(S) in use; defining standard error handling approaches and extensions; and providing guidelines (and some supporting Poseidon-based tools) for converting UML into WSDL. This could inform the later stages of the XCRI project when efforts move towards a prototype Curriculum Services WSDL.

After lunch, Balbir Barn outlined the aims and progress to date of the COVARM project: an initiative, led by Thames Valley University, to develop a Reference Model for the course validation process. UML models had been created from in-depth studies of the workflow and procedures of four different institutions. Analysis was now underway to produce a canonical model that captured the essence of the four. Balbir's project was keen to integrate with XCRI's curriculum work, and COVARM's UML diagrams marked information flows that were intended to be understood through XCRI's contribution.

I then presented a brief summary of XCRI's aims and progress, which mentioned in particular:

• our collective review of Norway's CDM and our intention to contribute a UK perspective to connected work on a Europe-wide standard
• the XML brainstorming that had taken place with Scott Wilson (CETIS), Sean Mehan (UHI), Ben Ryan (KaiNao) and Pete Johnson (UKOLN), which had resulted in a close working relationship between XCRI and the Pathways4Progression Regional Pilot and a draft XML schema
• the case studies and tools for capturing and transforming curriculum content for multiple audiences that had been kindly supplied by Tavis Reddick (Fife) and Paul Walk (London Metropolitan University)
• the survey work that was currently underway to examine the content and structure of UK online prospectus entries

In parallel to the assessment codebash, Ben Ryan (Pathway4Progression), Balbir Barn (COVARM), Lorna Campbell (CETIS), Wilbert Kraan (CETIS), Christina Smart (CETIS) and I decided to share ideas around the area of curriculum modelling. Discussion topics included:

• Lessons from the LOM Core regarding vocabularies and mechanisms for managing equivalence between them
• The challenge of APEL (Accreditation of Prior Experiential Learning)
• Requirements for modelling award credits: the need to include scheme, level and value, and the possibility that a particular curriculum component may be accredited against multiple schemes: for instance, CATS and ECTS
• The need to incentivise curriculum metadata capture by providing value-adding features such as automating the production of matrices showing stage and programme learning outcomes coverage by modules, and consistency checks that relevant learning outcomes are covered by the proposed assessment strategy
• The reality of composition rules for complex courses (Wilbert kindly supplied some first-hand experience of some particularly challenging scenarios in which a valid selection of units depended on the owning body of the offering as well as pre- and co-requisites)

• relationship constraints (parent, child and sibling)
• ownership constraints
• prior study constraints (must have studied X)
• prior achievement constraints (must have gained more than 40% in X)

In the best XCRI tradition, the day's intense activities ended in a "summit", this time at Rab Ha's - a fine choice from Rowin and a great chance to discover some spooky coincidences and interesting points of connection between the different projects. (The "Dangermouse" conversation alone made it worth the 4am start and the very, very long day, and the phrase "roll your own" will never have the same meaning again!)

Key Lessons

• XCRI is a best classified as an e-Admin project within the new e-Framework
• BPEL has potential for Reference Model projects, but component web services in the curriculum area are much less mature than component assessment services
• The IMS General Web Services spec is worth consulting when drafting the Curriculum Services WSDL
• Close dialogue with the COVARM and Pathway4Progression projects is essential and links with FREMA need to be pursued
• Complex course specifications provide an excellent way to test the expressive power of the emerging curriculum information model, and contributions should be sought from partners to develop the model further