Segue, WordPress, Moodle, MediaWiki and Concerto as well as other applications now provide a extensible set of services for teaching, learning and research that support the creation of sites for courses, e-portfolios, research, communication, projects, podcasts, media streaming, discussions, blogs, wikis, monographs, personal workspaces, news aggregation, syndication, cataloguing, categorizing, taxonomies and folksonomies and more.  Since most of all the academic technologies in use at Middlebury are open source, the college pays very little in licensing fees for these tools.  Perhaps more of an investment has been made at Middlebury in professional development and staff then at peer institutions.

The Segue v2 has nearly all the functionality of Segue v1 as well as flexibility and extensibility not possible in the first version of Segue.  We’ll want to flesh out remaining missing functionality in Segue v2 so that everyone can migrate and continue to do the same sort of things they have come to expect  New functionality we’ll want to add is some support for dated authorizations, grading, and custom groups.  The extreme flexibility of Segue v2 as well as it granular authorization do offer challenges to its scalability and so there is more optimization that we will need to do.  Concerto has been in use for the last year of so and is meeting the basic needs of its users, but it could also use some improvements.  Finally more work needs to be done on MiddMedia to make it really useful and usable in a variety of systems.

Now that we have a suite of applications (Segue, Concerto, MiddMedia) built on service oriented architectures (SOA) and using open service standards, we would like to develop a set of enterprise services that can be used by a wide range of applications.  The prototype for this sort of development is MiddMedia, our video streaming service.  MiddMedia will be a lightweight media management application built with Harmoni that will provide a web services endpoint for uploading and managing media files.  This will enable us to build simple plugins, modules or widgets for various applications on campus including WordPress, Segue, Concerto and Moodle that will allow users to access MiddMedia files directly from those applications.  This sort of service would complement existing video sharing services such as YouTube, Vimeo and Google, providing Middlebury users more media format flexibility and better access control then these public services currently offer.

One of the primary concerns with using hosted solutions for academic technologies in particular, but also for administrative systems is access control and the management of identity.  Many, though not all, faculty, students and staff are reluctant to publish “curricular” content publicly via sites like WordPress.com and YouTube.  There are concerns of copyright, intellectual property and Family Educational Rights and Privacy Act (FERBA) regulations that require an institution to limit access to content that its community has created as part of that institution’s mission. Another concern is identity management.  Currently, one of the most reliable ways of identifying the author or editor of content at Middlebury is by means of a secure, well-managed link between that content in a given curricular application and a Middlebury user account in Banner, our adminstrative system of reference.

To this end, we would like to develop standard authentication services that can be used by a wide range of applications on campus and possibly be accessible to hosted solutions at some point in the future.  We would design these services to be extensible and support a variety of protocols and authentication providers possibly including OpenID and/or Facebook Connect.  From this we could begin to support single sign-on across many of our applications.  Building on this, we would add services for agents and groups that would allow applications to get a list of all the groups a given user is a member of or to get all the members of a given group.  Closely related would be course management services that would get all the courses for a given student or faculty person or all the courses in a given department, for a given semester… etc.

Finally it would be good to some sort of standardized access to content repositories, whether that content is in a wiki, blog, content or asset management system, media server or online discussion.  Thus some sort of repository service the would provide access to a range of repositories would greatly simple access control, syndication, aggregation, and serve as the foundation for evolving ontologies.  Such services could provide access to assets in a given collection of a digital asset management system such as CONTENTdm or Concerto and make these available for use in other systems such as WordPress, Segue or the college website.

One of the challenges to systems interoperability, data portability and the integration of applications is sharing information across of runtime environments of different programming languages and across network to servers in remote locations.  These challenges can be solved by means of a enterprise service bus, which is one of the primary deliverables in the Mellon grant prospectus we are co-authoring with Jeff Merriman, director of O.K.I. and directors of the UOC CAMPUS project.  Should we receive funding for this proposal, this would allow us to coordinate and inform efforts to build this bus in a way that meets the needs of Middlebury in particular and small liberal arts colleges in general.

The second part of the Mellon grant prospectus we are helping to draft involves the development of a “enterprise services framework.”  This work will be led by the Open University of Catalonia (UOC), building on their CAMPUS project as a prototype for a suite of services that integrate into enterprise systems.  These are the same services we would like to develop at Middlebury and so if the grant proposal is accepted, it is likely we could receive funding to help with this work.  Critical to the viability of these proposals will be forming a strategic alliance with Project Bamboo, the multi-institutional, interdisciplinary, and inter-organizational effort that is also seeking to define services and a service framework.  What we may end up proposing is to develop “core” services upon which Project Bamboo can build the higher level “scholarly” services which is the focus of their initiative.

We don’t envision Harmoni, Segue and Concerto existing as they do for more than a few years.  Harmoni will likely be deconstructed into a set of service endpoints that interface more closely with enterprise systems than they do now and then eventually be swapped out for an educational enterprise services framework something like what is described above.  These core services could then be augmented by any Bamboo “scholarly” services that would be useful to Middlebury.  We expect future curricular applications to eventually have the same sort of granular access control as Segue and Concerto.  Until that time, we expect these systems will continue to have unique value.

We are definitively looking forward to innovating higher up the stack.  We think we have already made signification innovations in UI design that we hope will influence the fields of educational and information technology.  We have been in contact with the Fluid Project and plan to share our work with them.  As well, we expect to learn much from what they have done and hope to be able to swap out some of our UI components with standardized versions that they are developing.