Towards a Sharable Digital Library of Reusable Teaching ...



Towards a Sharable Digital Library of Reusable Teaching Resources:

Roles for Rich Metadata

David McArthur, Sarah Giersch, Bill Graves

Eduprise

Charles R. Ward, Richard Dillaman, Russell Herman, Gabriel Lugo,

James Reeves, Ron Vetter

University of North Carolina at Wilmington

Ed Fox, Hussein Suleman

Virginia Tech

Deborah Knox

The College of New Jersey

Scott Owen

Georgia State University

Submitted to the Journal on Educational Resources in Computing

Feb 1, 2001

Contact: dmcarthur@

Abstract

Faculty across the country have created a wealth of digital resources for teaching and are willing to share them; but they lack a repository where they can submit their materials, find related ones, create new content, and exchange them with others. iLumina is a digital library of undergraduate teaching materials for science, mathematics, engineering and technology (SMET) education that will provide such a repository, and related community and user services. The purpose of this paper is to provide an overview of the structure of iLumina, emphasizing the several roles that IMS metadata play in organizing library resources and supporting users services. We argue that iLumina’s metadata will not only help users discover appropriate resources – the typical role for such descriptions – but can also help manage reviews of resources, coordinate the community construction of metadata, and connect library users to one another.

INDEX TERMS

Categories and Subject Descriptors:

Information Systems – Information Storage and Retrieval-Digital Libraries (H.3.7): Collection; Standards; User Issues

Computing Milieux – Computer Uses in Education (K.3.1):

Distance Learning; Collaborative Learning

Keywords:

Distributed Systems, Metadata, XML

The goal: Sharing and reusing online teaching resources

Communities of educators have often created course materials by peer-to-peer exchange of resources that are then combined into larger modules. Such home-grown materials—from lecture notes, to informal demonstrations, to assignments and quizzes—can often be combined in flexible ways with mass-market textbooks to build tailored courses and curricula. One problem with this community-sharing model, though, is that it does not scale-up, largely because traditional educational communities have been geographically separate. The best materials in one region are often unknown to distant faculty who might value them, use them, and improve them. Conversely, mediocre content may endure in one area because geography protects it from competition from outside resources.

These market inefficiencies and scaling failures are among the reasons that publishers now dominate the production and distribution of teaching materials, from primary school to post-graduate studies. But this model, too, has its limitations—not necessarily in terms of scalability, but in terms of the speed of publication and the variety of published materials (see [Borgman 2000], for an insightful discussion of modern publishers and their competitors). To keep margins high, publishers concentrate on big production runs of a relatively few, large-scale products, such as textbooks and software packages. However, in education there is a continuing strong demand for a more diverse set of smaller “learning components” that faculty and instructors can plug together and tailor into personalized courses [Roschelle et al. 1999].

The limitations of the publisher-based model have encouraged a growing number of projects to try to remedy the scalability problems of the community-sharing model. One main approach uses the Internet. Although small-scale educational communities are often physically insular, digital exchanges, designed effectively, can allow faculty to share online instructional resources regardless of location, creating much larger “virtual” communities, and potentially overcoming the market failures of traditional informal course-content exchange communities.

This is the basic premise behind iLumina, a digital library of undergraduate teaching materials in science, mathematics, engineering and technology (SMET) education that is being developed by Eduprise, The University of North Carolina at Wilmington, Georgia State University, Grand Valley State, and Virginia Tech.[1] Our experience suggests that faculty across the country have already created a wealth of digital resources for teaching—often small in scale, such as a well-crafted image—and are willing to share them. But they lack a repository where they can submit their materials, find related ones (without having to reinvent them), create new content (either individually or through collaboration), and, collectively, improve both the quantity and quality of digital teaching resources. iLumina will try to provide such a repository, and related user services.

In this brief paper we provide an overview of iLumina and discuss how it is attempting to foster an online community-sharing model of course-material creation and reuse. The first section sketches the structure of the digital library. Following this we outline the challenges confronting iLumina. Our work is building on several related projects, such as the National Engineering Education Delivery System (NEEDS) and the Connected Curriculum Project (CCP), that have already begun to explore this area. One way we complement these projects is our emphasis on the use of rich metadata to underpin an array of digital library services. We therefore devote most of the paper to an overview of the roles that such metadata can play in meeting the challenges faced by iLumina. For more information on our project, see ilumina-.

The approach: Overview of iLumina

The driving vision of iLumina is to provide a platform of content and services that will dramatically improve the ability of faculty to create and exchange educational materials, across a wide range of science, math and engineering disciplines. These resources will frequently be learning fragments, even single images, rather than large course segments. iLumina therefore must provide tools and services that motivate educators to create new resources from these pieces, as well as offer a diverse collection of materials to the communities that it serves.

To promote the acquisition of a wide range of content, iLumina will rely on a varied set of contributing partners, whose materials can follow several distinct migration paths into the library. As Figure 1 [A] illustrates, streams of resources can come from existing repositories whose materials have already been peer-reviewed and are set to be cataloged into iLumina. Our partners contributing reviewed materials are the Computer Science Teaching Center (CSTC; ), and the ACM SIGGRAPH Education Committee’s Digital Library (SECDL; education.). iLumina represents an additional distribution channel for such repositories. A second set of providers will be submitting resources that have not undergone independent assessment; they will be accepted only after review, which will be contributed by other members of the iLumina community (Figure 1 [B]). Currently, most of these new resources will come from faculty at UNC Wilmington. Developers from the CCP [math.duke.edu/education/ccp/] and Mathwright [] projects also will contribute. For some of these providers, iLumina will be their first distribution channel. In the future we hope to open up both of these migration paths to a much broader set of content suppliers.

Reviewed resources accepted from UNCW and other iLumina content providers will not be placed directly into iLumina, but rather will reside in additional content repositories, analogous to CSTC and SECDL. As Figure 1 ([C] and [D]) indicates, iLumina itself will actually be a repository not of the content supplied by providers, but only of the metadata associated with these distributed resources. This metadata will be compiled in two phases. First, we require each of the iLumina content repositories to adhere to specifications proposed by the Open Archives Initiative [Van de Sompel and Legoze 2000; Open Archives 2001]. Because repositories comply with these “data provider” conventions, iLumina, as a “service provider” will be able to harvest metadata from the component repositories, as needed. Specifically, metadata exported in the iLumina and Dublin Core formats will be harvested from CSTC and SECDL into the iLumina open archive, and so placed into the library. Metadata for resources supplied by other project partners, such as UNCW, will not need transformation, because they will be created “from scratch”, by metadata tools that comply with the iLumina metadata specification.

Figure1—Overview of the acquisition and creation processes for iLumina resources.

iLumina will expect project partners that own the component repositories to manage their content, ensuring that all resources are preserved and that links remain valid and up-to-date. Component repositories also can offer their own user-services (as do CSTC and SECDL; see Figure 2), in addition to providing data for iLumina. However, iLumina’s main role will be to make the content of the federated repositories available through a centralized web-based interface that includes a collection of user services which rely on a foundation of standardized metadata (Figure 1 [E]). Since the initial version of iLumina will include only free digital content, library users will be able to download digital copies of resources and reuse them without constraints—at least for personal, non-commercial purposes. Users also will be encouraged to employ iLumina authoring to create new modules. For example, a faculty member might compose a titration simulation and graphing applet, found through independent iLumina library researches, into a larger chemistry-course fragment. Finally, as the upward arrows in Figure 1 [F] suggest, the faculty user could submit such newly-created content for inclusion into iLumina, completing the library life-cycle. It is in this sense that we see iLumina as an organic digital library, tying together an expanding community of users and developers who collaboratively construct and share learning resources.

[pic]

Figure2—The CSTC website.

CSTC not only provides resources (metadata) for iLumina, but also offers its content directly to end-users, through a set of search, review and submission services.

Challenges: What will it take to make iLumina succeed?

iLumina must jump many hurdles, both technical and social, before it becomes a valuable source of reusable digital educational materials in science and mathematics. At a minimum, it will need to be designed to accommodate diversity along several dimensions. Materials will cover a wide collection of disciplines; different vocabularies and classification schemes will be needed to describe them in ways that are comprehensible to their various user communities. Resources must span a similarly extensive array of technical formats, from simple pictures to Java applets to complex applications. The “grain-size” of library content, in other words, will range from digital “atoms” to structured components.

In short, iLumina will need to provide relatively traditional user-services that support the selection, description, organization, distribution and preservation of materials. Some of the problems confronting iLumina in these areas are shared by many digital (and non-digital) library ventures. However, other challenges are specific to our goal of serving educators who not only will want to access high-quality learning objects, but also must be able to create new library content by building on existing resources. iLumina will need to describe resources in relatively education-specific terms; for instance, expressing multiple pedagogical purposes and different learning “types” (e.g., tests, simulations, experiments, etc.). At the same time, the library also must include a number of innovative user services, including tools to author and repackage learning objects, to comment on and rate materials, and perhaps to work collaboratively with other educators that belong to the iLumina community.

Focus: The roles of learning metadata

The iLumina project will be developing a collection of web-based tools and interfaces to meet some of these challenges. In this brief paper we focus only on the role of metadata, in part because they are the foundation of several key user services. We first review the structure of iLumina’s metadata, and then provide a few examples that illustrate the kinds of user benefits they might provide.

Overview of iLumina Metadata

Metadata descriptions can range from very informal notes, such as readers’ opinions of books on , to the highly authoritative and standardized cataloging descriptions created by publishers and cataloging librarians [Dempsey and Heery 1998; Gilliland-Swetland, 1998]. The metadata structure for iLumina’s digital resources permits descriptions at both ends of this spectrum. We have derived iLumina’s metadata from the IMS metadata specifications[2], one of several areas where the consortium is developing specifications [IMS 2000]. Figure 3 presents an example of iLumina metadata describing a Java simulation from the CSTC repository. Figure 4 shows metadata for a much simpler kind of digital resource, residing in the UNCW repository.

|  |

|- |

|+ |

|- |

|  Applets for Teaching Computer Graphics Concepts |

|  These java applets allow students to explore Bezier curves, 2D transformations operations, lighting models and 3D |

|viewing operations. |

|  bezier_curves transformations lighting_model computer_graphics |

|  |

|+ |

|- |

|  application/x-java |

|  |

|- |

|  Browser |

|  Microsoft Internet Explorer |

|  3.0 |

|  |

|  Java Virtual Machine 1.1 or higher |

|  |

|- |

|  Simulation |

|  Undergraduate Lower-Division |

|  en-US |

|  |

|+ |

|+ |

|+ |

|- |

|  Discipline |

|- |

|  ACM Computing Classification System |

|- |

|  I |

|  Computer Methodologies |

|  |

|- |

|  3 |

|  Computer Graphics |

|  |

|  |

|  |

|  |

Figure 3—iLumina metadata for a Java simulation resource in the CSTC repository.

Metadata are represented in XML, with elements from five of the main categories of metadata – metametadata, lifecycle, rights, relation and annotation – collapsed; and ones from general, technical, educational and classification categories fully expanded.

Both examples illustrate the structured nature of IMS metadata, which is organized at the top level into nine metadata categories—metametadata, general, lifecycle, technical, educational, rights, relation, annotation and classification. Each category focuses on distinct features of an element, from how it is constructed, to what it is about, to who owns the resource and what rights end-users hold. Within each category, elements can be hierarchically organized, permitting resources to be described not only in broad categorical terms, but also in great detail.

Metadata for the resource in Figure 3 were created by mapping CTSC metadata from its native format to the iLumina schema. This mapping was done manually; we are now developing a simple XML-based metadata tool that will automate the process. The record illustrates several education-specific IMS metadata elements that allow resources to be described in terms of its expected grade-level or audience (learningcontext) and its pedagogical type (learningresourcetype), such as simulation, lab, exam, and so on. Such a characterization is distinct from discipline-related classification schemes, also shown in Figure 3. This part of the record shows that CSTC adopts the ACM Classification System to provide subject-level resource descriptions. However, the IMS specification supports all controlled vocabularies typically used for this purpose, such as the Library of Congress Subject Headings or the Universal Decimal Classification. And up to ten distinct taxonomies can be used to describe a given library resource. This allows the federated repositories that provide metadata to iLumina to use the classification schemes they prefer, not ones that iLumina requires.

|  |

|- |

|+ |

|- |

|  Picture of four metals: Aluminum, Cooper, Zinc and Lead respectively |

|  Still image of 4 different metals of varying mass and density |

|  |

|- |

|  Final |

|- |

|  Author |

|- |

|  BEGIN:VCARD N:Ward, Charles R. END:VCARD |

|  |

|- |

|  2000-08-23 |

|  |

|  |

|- |

|  Content Creator |

|- |

|  BEGIN:VCARD N:Smith, Cliff END:VCARD |

|  |

|- |

|  2000-08-23 |

|  |

|  |

|  |

|- |

|  .jpg (image/jpeg) |

|  225000 |

|  |

|- |

|  Operating System |

|  Windows |

|  3.0 |

|  |

|- |

|  Browser |

|  Microsoft Internet Explorer |

|  3.0 |

|  |

|  |

|+ |

|- |

|  No |

|  No |

|  Copyright free for non-commercial use. For commercial use, contact author. |

|  |

|+ |

|+ |

|+ |

|  |

Figure 4—iLumina metadata for a JPEG image resource in the UNCW repository.

Metadata for the general, lifecycle, technical and rights categories are expanded fully, while the other structural categories are hidden in this view.

Figure 4, describing a simple JPEG from UNCW, shows, among other things, that a number of learning resource-creation roles can be recorded in metadata (see the vcards under centity and contribute elements). Here, just the formal author and the content implementor are distinguished, but the IMS specification supports a wide range of production roles, including publisher, editor, designer and implementer. In addition, the technical metadata elements allow catalogers to express the (mime) format, size and location of the resource, as well as platform requirements that will constrain the use of the resource. This information often helps users determine whether they want to download the resource (is it too big?; will it run on my machine?) and how to make it work if they do so (what plug-ins are necessary?).

The essential roles of iLumina metadata

iLumina metadata will be the basis for several critical user services in our library. Most obviously, iLumina’s extensive metadata element set provides a rich language for describing heterogeneous resources. For example, it enables catalogers to convey not only subject area or discipline (at several levels of detail), but also granularity or aggregation level, as well as the pedagogical type of a resource (in Figure 3, a relatively complex simulation; in Figure 4 a very simple image). Because our metadata are rooted in the de facto IMS standards, iLumina will enable catalogers across our several distributed repositories to use the same standard language to describe very different kinds of resources.

Such uniform descriptions of iLumina resources can, at a minimum, improve the library users’ ability to discover valuable digital resources. First, even though searches and browsing through resources (metadata) will be conducted centrally, through iLumina’s web front-end, users will, in effect, be exploring a federated collection of repositories. Each repository may have its own interface; iLumina users, however, will not need to learn these differing front-ends in detail (unless they choose to). Similarly, by adopting standardized metadata conventions, users will need to learn only one language for describing resources in the federation, relying on iLumina’s automated harvesting and mapping of metadata from the distributed repositories to translate between the different terminologies, and so reduce the cognitive load on users. Moreover, this simplification and standardization of front-ends and metadata terminology should not impose user-costs, in terms of search and discovery performance. Because the iLumina metadata schema is rooted in the IMS standards, metadata can be mapped from CSTC and SECDL, or from any repository that complies with the basic Open Archives (OA) metadata set [Open Archives 2001], without any loss of information. Any search conducted on these repositories, therefore, can be replicated on iLumina (although we cannot guarantee this for repositories that extend the OA metadata set in arbitrary ways).

Using iLumina metadata for more than discovery

Standardized metadata not only support broad and powerful search engines, but can also be the basis for a “second tier” of user services. One important role metadata can play is to express relationships among different learning resources, quite possibly in different repositories. Formally, this could be done using the IMS relation metadata category, which provides a reasonably extensible vocabulary that captures a variety of “family” connections. A notional example, capturing multiple relationships for a particular learning object, might look like this:

  IsBasedOn

  [uri of object]

  This object is a variant of a previous one

  IsPartof

  [uri of object]

  This object is a component for a new one

By itself, this sort of metadata fragment is of limited value. However, it could be the foundation of a tool that partially automates the construction of new learning objects from collections of existing ones, or that simply record the lineage of resources. In other words, it is a first step in reusing resources in novel ways—a critical goal of the iLumina project.

iLumina/IMS metadata can be the basis for several other potentially valuable digital library services. Space permits only a brief itemization of just a few possibilities:

• Managing formal reviews and informal ratings. As Figure 1 indicates, new resources in iLumina will undergo review before acceptance. Such reviews are metadata, of course, and can be included in iLumina records using a new relation element or attribute named iLuminaReviewedBy[3]. This would provide not only the review content, but also a (vcard) link to the reviewer. We are also planning to use the annotation element to allow users of iLumina resources to rate and comment on them—in effect, providing a “non-authoritative” (-style) level of reviews of library materials, to complement formal peer reviews[4].

• Coordinating community construction of metadata. iLumina will rely on IMS-conforming metadata, but such richness comes at a cost: it takes a lot of effort to construct and maintain extensive, accurate and useful metadata. One way to reduce this cost is to distribute it, alleviating metadata bottlenecks. The previous point shows how different iLumina community members can contribute metadata, through relation and annotation elements. We propose to extend this idea by getting members to contribute metadata for the elements they should know best in the production-cycle of digital library content. For example, authors or creators will provide a title and description (see Figures 3 and 4) when a resource is submitted; reviewers can insert peer assessments; catalogers will add discipline-level classifications and perhaps a range of technical information; and elements such as submission date can be set automatically.

• Connecting library users. Metadata are mostly viewed as a means to connect library users to the resources they need; however, the previous examples suggest metadata also can help connect users to one another. For instance, a library user might not only be able to gather information about the subject or technical format of a learning resource, but also learn about who has downloaded it, and how they rated it (for users who permit the sharing of such personal information). Similarly, as Figure 4 illustrates, iLumina metadata permit all contributors to a resource to be documented, not just the nominal author. These linkages can provide library users with a whole host of community contacts for future discussions and collaboration. In short, good metadata can provide as much information about the community of users as about the library content itself. In this way, metadata can at least begin to help digital libraries fulfill some of the important (but recently, it seems, very overlooked) social functions of their bricks-and-mortar counterparts.

Conclusions

The kinds of services outlined above are a first step in developing a digital library community that will share and reuse resources as effectively as the best traditional ones have. But it will take careful design and testing to implement these services effectively, and to realize their potential benefits. Many open questions will need to be addressed. For one, while it may be appealing to permit informal user ratings as well as more formal metadata elements, it is not obvious that all this heterogeneous information belongs in a single metadata record, administered in a centralized iLumina repository. Perhaps, for example, the central site should keep only summary rating statistics, while originating sites hold specific ratings and comments, to enhance security and privacy [Miller 1999]. This may not be the best policy, but it is clear that the more broadly we construe information about resources, the more likely it is that this metadata will be as diverse and distributed as the resource repositories themselves.

And what should be the status of non-authoritative user comments and ratings, relative to formal peer reviews? Digital technology democratizes the reviewing process to a degree that the traditional publication pipeline never could. It is tempting, therefore, to keep formal reviews of digital library materials to a minimum, relying on user ratings and online cross-talk to supplant, or reduce, the need for detailed peer reviews. A number of advocates of online publication and eprint servers (see, for example, [Harnad 1995; Odlyzko 1997; The Economist 2000]) have argued that this distributed and informal reviewing procedure is a viable model for the sharing of academic research. On the other hand, thoughtful reviews are part of the traditional quality-assurance chain of publication. If casual user commentary inspires less trust than peer review, the elimination of such formal assessments may weaken that chain—and make it harder for library community members to find high-quality resources, not easier.

Many important questions like these will need to be addressed by experimenting with various production and reviewing alternatives. Still, we believe that rich, structured metadata specifications, such as the ones provided by IMS, will be a linchpin in many of the most creative and valuable answers. In this sense our iLumina project is an attempt to “stress test” the IMS metadata specification, and to assess its ability to provide a foundation for a scalable, community-sharing approach to the exchange and creation of reusable digital resources for science, mathematics, engineering and technology education.

References

Borgman, C. 2000. From Gutenberg to the Global Information Infrastructure. Cambridge, MA: MIT Press.

Dempsey, L., and Heery, P. 1998. Metadata: a current review of practice and issues. Journal of Documentation, 54, 2, 145-172.

Economist, The. 2000. Will journal publishers perish? The Economist, 81-82, May 13.

Gilliland-Swetland, A. 1998. Defining Metadata. In M. Baca, ed. Introduction to Metadata: Pathways to Digital Information. Los Angeles: Getty Information Institute.

Harnad, S. 1995. The PostGutenberg Galaxy: How to get there from here. The Information Society 11, 4, 285-292.

IMS. 2000. The IMS Global Learning Consortium. []

Miller, T. 1999. Improving graduate education through digital library tools. Honors BS Thesis. Virginia Polytechnic Institute and State University.

Odlyzko, A. 1997. Silicon dreams and silicon bricks: the continuing evolution of libraries. Library Trends, 46, 1, 155-167.

Open Archives. 2001. The Open Archives Initiative Protocol for Metadata Harvesting. Document Version 2001-01-21 [].

Recker, M., and Wiley, D. (To appear). A non-authoritative educational metadata ontology for filtering and recommending learning objects. Interactive Learning Environments: Special issue on metadata.

Roschelle, J. DiGiano, C., Koultis, M., Repenning, A., Phillips, J., Jackiw, N., and Suthers, D. 1999. Developing Educational Software Components. Computer, 32, 9, 50-58.

Van de Sompel, H., and Legoze, C. 2000. The Santa Fe Convention of the Open Archives Initiative. D-Lib Magazine, 6, 2, Feb. []

-----------------------

[1] This work, still in a preliminary phase, was funded as part of the NSF DLI-Phase 2, grant #0002935.

[2] For all practical purposes our metadata schema derives equally from the IEEE LOM 3.5 metadata specification [], which differs only in small ways from the IMS specification.

[3] The Digital Library for Earth Systems Education (DLESE) project [], which is also using IMS/IEEE metadata, is considering a similar extension.

[4] The Instructional Architect project (Recker and Wiley, to appear) is planning a similar digital library service; the term “non-authoritative metadata” originates with them.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download