*** Draft ***

Report of the
SCORM/NSDL Technical Meeting
Alexandria, Virginia
May 18, 2004

4. Persistence and Reuse

Persistence and reuse are two aspects of the same problem, how to manage educational resources so that they are used repeatedly. Reuse has a focus on expanding the community of people who use a resources. Persistence emphasizes longevity.

4.1 Reuse

The NSDL Reuseable Learning project is developing a set of reusable design guidelines. These will help content developers enable the use and reuse of their content and will help repositories support reuse. The following discussion highlights reusability issues and their implications.

Note: There are differences between “reuse” and “repurposing” or (in other terms) between “adoption” and “adaptation,” as well as differences in the needs of authors, learners and collections. Some issues only apply to certain types of reuse. Please see www.reusablelearning.org for a more complete discussion.

Issues

Reusable Design: Ultimately reusability depends on the structural and instructional design of content. However, there are several lower level issues that have the most potential affect on CORDRA and that create the most immediate barriers to reuse. The challenges below are the ones that must be overcome before design makes a difference.

Rights Management: Traditionally, content producers and publishers a have been connected to consumers and users by a linear chain with terms and conditions of use at each link governed by business relationships and contracts. Now repositories, digital libraries, (learning) content management systems, and authoring / assembly environments serve as nodes on a distributed network where content is aggregated, disaggregated, repurposed, and re-distributed multiple times with relative anonymity. Rights and terms of use must therefore be explicitly associated with content and must persist as content is distributed, reused and repurposed.

Interoperability: Interoperability encompasses technical issues addressed by specifications and standards such as SCORM. These specification and standards are particularly applicable to content at the granularity of a “learning object” (in the Learnativity model) or “SCO” (in SCORM terminology) that address a single instructional objective and that are navigationally self-contained and as contextually self-contained as is feasible given the subject matter. One overarching challenge is finding the right standards to address interoperability for objects that are both more and less granular.

An ongoing problem is that there is entirely too much emphasis on specifications and standards in the conversation surrounding learning and learning technology. Interoperability should come from using tools that incorporate standards without making an issue of it. Projects that have leadership potential must address this (presumably by setting good examples) and deal with two specific challenges: (a) the use of different specifications by different communities, in particular by the educational community, and (b) the need for a roadmap that moves learning technology to a Service Oriented Architecture compatible with all significant existing interoperability efforts.

Metadata: The discussions on May 18 brought forth many fundamental questions about metadata and identifiers, including FRBR issues and provenance and preservation questions. From a reusability perspective, the important thing is to have (good) metadata in the first place. Reusability requires giving users access to meaningful and accurate educational and technical assertions about objects, as well as to rights and terms of use.

For objects that are “eternal” (where, following Bill Arms’ suggestion, eternity might be defined as lasting 10 years), the effort required to create such metadata seems worthwhile, noting that the cost of professionally creating and recording metadata is significant. For truly ephemeral objects, metadata may not be needed and in any case will not be generated if it costs time or money. From a reusability perspective, the most important objects might be those that they are intended to be used for periods of months or years but that they are also intended to be modified. These might be called “evolving.” Examples include content that lends itself to disaggregation into reusable parts. For evolving objects, automated metadata generation is needed to reduce the cost of metadata and issues of identity, versioning, and relationships must be resolved.

User Behavior: It is all too easy to create guidelines, technology and structures that are theoretically sound but that ignore what people actually do. For example, the digital library community understands the value of validated content and metadata-enriched discovery, but people go to Google. It makes more sense to enable users to discover validated resources through Google rather than to force users to go to a separate interface. More generally, the community of experts should always remind itself that users will come up with new and unforeseen ways to use and reuse content and that its job is to enable rather than to enforce reuse practices.

[Section 4.1 is based on notes by Robby Robson.]

4.2 Persistence

The discussion concentrated on 10-year persistence. What actions can be taken to increase the likelihood that materials used in education and training today will continue to be available for 10 years? When an education package such as a course uses a learning object, the user of the package wants to be confident that the resource is available exactly as the builder of the package intended, even though it has not received consistent maintenance.

The following observations were made during the discussion:

• Persistence of the organization(s) that manage the resource and make it available is of fundamental importance. The risk of losing resources is reduced if they are managed by organizations that have incentives and procedures for preserving materials and make them available (e.g., publishers or libraries).

• The existence of several copies, maintained by separate organizations, enhances persistence. For this reason, among others, clarity of rights and permissions is important.

• Technical risk is lowered by following commonly used technology and commonly used options within the technology chosen. (For example, PDF is widely used, but it is possible to create PDF files that depend on machine-specific features. A file in PDF-A format uses only generic features.)

• Identification of resources and version control are important, but the administrative procedures are more important than the technology used.

[Section 4.2 is based on notes by Bill Arms.]