Adopting SCORM 1.2 standards in a courseware production environment.

Imagine a time when an entire curriculum exists electronically with lesson information described in such detail that each media file is notated and easily discovered. Consider gathering blocks of instruction from several unrelated repositories and building a unique web-based course without the production time or costs. The newly formed course will run on any content server, track student progress, record grades, and catalog time spent on each lesson without additional programming.

**********

Although currently not a reality, these are the standards the Department of Defense (DoD) advocates in the development of web-based learning technology. Standards may decrease the cost of generating new training by providing reuse mechanisms and by decreasing duplication of training effort. In addition, standards may also increase the quality and availability of web-based training (ADL, 2001). The Sharable Content Object Reference Model (SCORM) released in early 2000 is the standards framework for web-based learning technology.

To standardize an entire industry consisting of proprietary software and hardware vendors is neither quick nor easy. Before the standards can be widely adopted, they must be evaluated, tested, and refined by content providers and end users. Little is known about the actual outcome of adopting the SCORM standard in a production environment. Much of the literature pertaining to SCORM attempts to predict the possible effects on course production (Bohl, Schellhase, Sengler, & Winand, 2002; Cohen, 2002; Shackelford, 2002).

The purpose of this article is to introduce the SCORM framework and its origins, to present an overview of the two main components, and to discuss lessons learned by adopting the SCORM standards in a production environment.

THE ORIGINS OF SCORM

The Department of Defense (DoD) explored standardizing web-based training in 1996, to increase the quality and reduce duplication of training efforts (Shackelford, 2002). In 1997, the DoD established the Advanced Distributed Learning (ADL) initiative to accelerate the development of learning software standards to meet the training needs of the military (ADL, 2003). ADL incorporated existing standards created by the Aviation Industry CBT Committee (AICC), the IMS Global Learning Consortium (IMS), and the Institute for Electrical and Electronics Engineers (IEEE). In 2000, ADL released version 1.0 of the Sharable Content Object Reference Model otherwise known as SCORM (Cohen, 2002).

PURPOSE

The purpose of SCORM is to create content that will run on any conformant learning management system (LMS). In addition, SCORM facilitates the discovery and reusability of content. Moreover, to standardize the way course developers and LMS vendors name learning objects, the Learning Objects Metadata (LOM) model was adopted (ADL, 2003; Bohl, Schellhase, Sengler, & Winand, 2002). Metadata is data about data. Similar in many ways to a library's card catalog file containing data about a book, a metadata file contains information about a learning object. The LOM model provides a standard way of naming content object data. For example, LOM vocabulary includes tags such as title, description, and keywords.

SCORM COMPONENTS

The latest version of SCORM is version 1.2, released in October of 2001. It consists of two components: the content aggregation model and the run-time environment or RTE (ADL, 2001). The two SCORM components enable content to be tagged and packaged to run on any conformant LMS. At the time this article was written, SCORM 1.3 working draft 1 was released with an additional sequencing component.

The Content Aggregation Model

The content aggregation model contains specifications for the content model, metadata, and content packaging. According to the specifications, content is comprised of assets, sharable content objects (SCOs), and the aggregation of instructional units. Assets are electronic representations of media and other pieces of data. For discovery and reuse, assets may be tagged with asset metadata based on the LOM vocabulary.

SCOs represent a collection of assets that communicate back to the LMS. The content model does not specify the minimum or maximum size of a SCO. However, a SCO should be independent of a learning context and reusable according to the framework. The glue that holds SCOs together to form a course is the content package. The content package is similar to a manifest in a shipping document. Like a shipping manifest, the content package contains a description of the items within a course. The content package uses metadata tags to describe the content. Moreover, the metadata tags outline the course information as well as the number and names of each SCO. Finally, the metadata tags provide the physical location of the files.

The Run-Time Environment

The second component of SCORM 1.2 is the Run-Time Environment (RTE). It defines the way content communicates with a conformant LMS (ADL, 2003). The RTE outlines functions such as the launching of the SCOs and the data model elements that define the information being communicated. The RTE also defines the responsibilities of an LMS in a SCORM conformant course. For example, the LMS is responsible for tracking the state of the SCO across sessions, sequencing, and SCO-to-SCO navigation. It does not; on the other hand, define the placement or look of SCO-to-SCO navigation. Moreover, the specification does not allow the LMS to have more than one SCO open at a time, and it requires the SCO to be launched in a child window or child frame of the LMS. Finally, the LMS provides the Application Program Interface (API), the code that permits the exchange of information between the SCOs and the LMS. There are 49 RTE commands defined in the reference, of which 14 are mandatory items to LMS vendors.

To summarize, SCORM 1.2 is the technical architecture that permits the reuse of web-based learning objects. The content aggregation model defines how content is separated into various subcategories. The second part of the aggregation model describes a technique to package courses for reusability on conformant learning management systems. The run-time environment is a technical specification for the interaction of content with the LMS.

APPLYING SCORM IN A PRODUCTION ENVIRONMENT

Project Background

Nebraska Educational Telecommunications (NET), in a partnership with the Nebraska National Guard, began development of web-based military training courses in July of 2000. The web-based training includes topics such as Sniper Training, Self-Aid and Buddy Care, HH60/UH60 Crew Training, Unexploded Ordnance Training, Field Sanitation, Development Counseling, and Sexual Harassment. NET has produced 25 SCORM 1.2 conformant courses for a total of 215 classroom equivalent hours. Three important lessons were learned with the adoption of SCORM:

1. The standards change significantly between versions.

2. Navigation within the framework is not well defined.

3. Troubleshooting is cumbersome.

Lesson One: Evolution of the Standards

SCORM is not a static document. With the release of SCORM 1.1 in January of 2001, few SCORM 1.1 course-authoring tools existed. Therefore, NET developed software tools to create SCO, asset, and course structure format (CSF) metadata files. In October of 2001, ten months after the release of version 1.1, SCORM 1.2 was released. Along with minor metadata changes, the CSF file was depreciated and content packaging was added. The release of SCORM 1.2 ended the usefulness of SCORM 1.1 tools.

At the time of the SCORM 1.2 release more off-the-shelf course-authoring tools existed. Rather than reinvest in a proprietary production tool, NET purchased a SCORM 1.2 content authoring tool. In the end, the authoring software cost NET significant delays in production because of the coding used to create the SCOs. A detailed description will be provided in the discussion of Lesson Three.

Lesson Two: Navigation

SCORM requires that all SCO-to-SCO navigation be handled by the LMS. Unfortunately, the specifics of the navigational system are not defined and instead are left to interpretation by LMS vendors. For example, content may appear within the LMS frameset or in a new window. Consequently, course designers do not know what the student will see on the screen.

Another challenging aspect of SCORM 1.2 navigation was designing a way to move from page to page within a SCO. This intra-SCO navigation process was not addressed in SCORM 1.2. Since most SCOs were larger than one page, a separate navigational system had to be developed.

NET explored two options. The first involved creating one-page SCOs that launched a separate program with an embedded navigational system. For example, a SCO could launch a Macromedia(r) Flash piece with several pages of content within the flash. The second solution was to create a navigational system within the HTML page that communicated back to the LMS until the student exited the SCO. The communication code between the SCO and the LMS was specified within the content authoring software. NET opted for the second solution. The unfortunate result was two navigational systems, the intra-SCO and the SCO-to-SCO systems, on the student's screen at the same time.

Lesson Three: Isolating Problems

Adherence to standards did not guarantee a flawless product. Additionally, isolating errors was cumbersome when dealing with two software vendors and an incomplete technical framework. For example, in January of 2002, dial-up users experienced LMS communications errors on several SCORM 1.2 courses that had passed the SCORM TestSuite provided by ADL. The communication sequence between the content and the LMS was flawed due to timing issues with the LMS's handling of the SCO-to-SCO navigation. The problem only manifested itself with slow Internet connections and when the students navigated quickly from SCO-to-SCO.

At first, the content authoring vendor indicated the course was not conformant due to errors in the intra-SCO navigation code. After some consultation with NET, the authoring vendor reported that a stray media element in the course was causing the problem. Finally, after more testing, the content authoring vendor confirmed that students with dial-up connections might run into problems because the variables communicating the status of the SCO to the LMS were not getting updated. The authoring vendor provided two alternatives: (a) tell the users on a slow connection to allow sufficient time for the page to load before using any of the navigation; or (b) reduce the size of navigation scripts in the course.

With course production halted and no solution from the packaging vendor, NET approached the LMS vendor with the problem. The LMS vendor replied that the code produced by the content authoring vendor would not work because of the way it communicated with the LMS. The LMS vendor offered to fully examine the code for $200.00 per hour, claiming that most problems were not with the LMS but resided with SCORM course developers.

As mentioned in Lesson One, the interpretation of the framework by the content authoring vendor caused delays in production. Therefore, to resolve the issue NET discontinued using the content authoring software and created a new intra-SCO navigation system. New tools were created to build SCORM 1.2 packaging files and SCORM conformant SCOs that communicate with the LMS. Finally, a delay was created in the intra-SCO navigation system to ensure the LMS would wait for the communication from the courseware, a parameter over-looked in the current version of SCORM 1.2.

SUMMARY AND CONCLUSION

In the future, SCORM may provide a way to find and create unique web-based courses that will run on any LMS. Today it is possible to create conformant web-based content that will run on an LMS that recognizes the current SCORM standard. Furthermore, LOM vocabulary may enable the discovery and reusability of web-based learning objects. Conversely, until content repository standards are ratified, the discovery of content objects remains difficult. Moreover, the time and expense to create metadata SCO and asset files does not appear to be justified without content repositories.

NET has successfully produced SCORM 1.2 courses that run on a conformant LMS, partially fulfilling the vision of the ADL. However, producing courses that adhere to an immature standard poses unique problems.

Development time and costs are high because the standard changes significantly between releases. While off-the-shelf packaging tools may help reduce production costs, interpretation of the standards by software vendors may create additional challenges. Because the SCORM standards are open to interpretation, coding techniques that are technically conformant may not work in some instances and thus become extremely cumbersome to troubleshoot.

The decision to produce SCORM conformant web-based courses must be weighed carefully. Conformance to the standard permits reuse among various learning management systems. However, conformance comes at a steep price in time and resources needed to mitigate the limitation of the evolving standards.

REFERENCES

ADL. (2001). Advanced Distributed Learning Sharable Content ObjectReference Model 1.2 [Online]. Available: www.adlnet.org

ADL. (2003). ADL overview [Online]. Available: www.adlnet.org

Bohl, O., Schellhase, J., Sengler, R., & Winand, U. (2002). The sharable content object reference model (SCORM)--A critical review. Paper presented at the International conference on computers in education, Auckland, New Zealand.

Cohen, E. J. (2002). The emerging standards effort in e-learning [Online]. Available: http://www.ltimagazine.com/ltimagazine/author/authorDetail.jsp?id=1303

Shackelford, B. (2002). A SCORM odyssey. T & D, 56(8), 30-35.

BRADLEY S. BARKER, UNIVERSITY OF NEBRASKA, USA

E-MAIL: [e-mail]