The text for my "system perspective" contribution to the bid:

From the system perspective, personal learning environments require, first and foremost, integrating well with the 'rest'of the web. Modern 'web2.0' developments have created a rich ecology of web services, data, and user-controlled interaction with content - a natural environment for personalised learning, opened by simple interoperability standards to both deep integrations and casual 'mashups'. But there is important work still to be done to adapt, present and enhance web2.0 so that ordinary learners can make effective use of these opportunities. New interfaces and applications are important, but it is also vital to fill gaps in the services ecology to make the system work for group activities by novice users.

Recent excitement about the possibility of PLEs has largely been inspired by recent developments in 'web2.0'. Early adopters of RSS, weblogs, social-bookmarking, etc have noted that their own expert use of these technologies and services could be seen as an ad-hoc personal system for professional development. Several authors have proposed that powerful PLEs can be assembled almost ad-hoc, for instance by providing learners with omnibus applications for aggregating content feeds and participating in weblogs (Anderson 2005, 2006; Farmer (2006). CETIS (the Centre for Educational Technology Interoperability Standards) has promoted a project to prototype this kind of functionality (Liber, 2005).

The present consortium is well-placed to pursue similar development ideas: KnowNet and IER have previously collaborated in the development of KNotes, an open-source collaborative blogging and discussion system, and have become expert in using web2.0 techniques and services. KNotes provides a rich platform within which to rapidly prototype personal learning features, and KNotes' support for web-service APIs and microformat standards is constantly improving.

It is important to understand, however, that work on interfaces and aggregators is not going to be enough to create environments which ordinary learners can make good use of. Gaps in the services ecology must also be addressed. Practical experiments with web2.0 techniques in realistic learning contexts are vital in revealing these gaps.

For instance, group-wise sharing of activities is currently very awkward. A canonical example is trying to share the semantics of tags in social bookmarking: trying to meet up in 'tagspace' for learning purposes turns out to be extremely tricky. The development of new types of services, applications and interfaces to mediate group activities and shared semantics is a key area in need of research and development.

A further limitation of most web2.0 services is that they require considerable expertise to make more-than-casual use of them, without which it is not possible to realise the benefits of use. As with groupwise sharing, there is great scope for new types of services and interfaces to mediate between novice users and the full functionality of web2.0 without compromising the basic design principles. This is another key area for research and development.

Finally, a rich area for research is exploring the combination of the 'small pieces loosely joined' approach with more sophisticated computational algorithms (cf. below). It should be noted that such combinations have always been at the heart of the best examples of web2.0: Google's sophisticated indexing and ranking system has been a crucial enabler of the emergence of blogging, while the user-appeal of blogging has been a key enabler of 'Googling' as an effective way to find appropriate content. In the case of the proposed research, we envision simple interfaces for producing small chunks of 'microcontent' in context that can produce the kind of large datasets of rich content required to make effective use of computational algorithms (Sinha, 2005). The combined complex-simple approach could also be useful in addressing the problem of generating rich metadata. Users are notoriously averse to entering structured metadata, but by making it very easy for users to tag and annotate web resources while engaged in real knowledge-work contexts, services like del.icio.us have evidenced the possibilities for metadata creation. Furthermore, there is much scope for making user-generated tagging in learning activities even more powerful by combining it with approaches which automatically extract metadata about the learners' activities.

References:

• Anderson, T (2005), Educational Social Overlay Networks. Weblog entry in 'Virtual Canuck', November 28, 2005.
• Anderson, T (2006), PLE's versus LMS: Are PLEs ready for Prime time? Weblog entry in 'Virtual Canuck', January 9, 2006.
• Farmer, J (2006), Blogs as personal learning environments. HigherEd BlogCon 2006 [screencast].
• Liber, O (2005), PLE Project Summary.
• Sinha, R (2005), Collaborative Filtering strikes back (this time with tags). Weblog entry in 'Rashmi Sinha - everyday cognition on the web', October 27, 2005.