Home

Three Common Properties of Efficient Online Instructional Support Systems

by Sloan-C

I. INTRODUCTION

In this article we describe three common properties of any system that provides efficient online instructional support. The properties are granularity (the size of the learning objects used by the system), discoverability and accessibility. We refer to online instructional support systems as Digital Learning Environments (DLEs). We define efficiency as the opposite of redundancy and waste of resource. This paper is not intended to be a catalog of benefits derived from the incorporation of these properties. Rather, it is intended to identify and describe the properties themselves. The properties are discussed in the context of implementation within the World-Wide Web environment although the concepts likely apply to other networked systems.

II. GRANULARITY

A. Historical Context
The visionary work of Dr. Ted Nelson, who coined both the terms "hypertext" and "hypermedia," continues to be one of the driving forces behind cutting-edge hypermedia developments like the recent advancements in the Web. Nelson [1] has written that in future hypermedia systems, "the operative unit is the VERSION, not the document." That is, static web pages will be replaced by dynamic arrangements of smaller units of information (which he calls "primedia," or primal media).

The first step toward implementing this vision was the web server functionality referred to as the "server-side include" (SSI). SSI gave authors the ability to include external text files in their web pages just as they were currently able to include external images. Using a defined syntax, web page authors could instruct the web server to simply "include" the separate file as a header or footer, for example. This greatly simplified web site maintenance because editing one included file would in effect update the entire site.

The Web is currently taking greater steps in the direction of Nelson's vision as database-driven web sites become increasingly popular. Sites such as barnesandnoble.com use hardly any static or "hard-coded" web pages. The pages you view when you visit the site are dynamic arrangements of primedia from the barnesandnoble.com database.

B. The Goal: High Granularity
In the instructional technology community the word "granularity" describes the types of objects Nelson calls primedia. The EduCause Instructional Management Systems Project (IMS Project) [2] uses "the relative size of the resource" as their working definition of granularity, as stated in their IMS Meta-Data Master Schema. The same document gives the following examples of relative sizes of instructional resources, ordered from low to high granularity: "curriculum, course, unit, topic, lesson, fragment." A large resource has low granularity.

By claiming that granularity is a property of all efficient instructional support systems we do not make the obvious claim that all DLEs contain instructional resources with relative sizes. We mean that the resources within such a system should be as granular, or as relatively small, as possible.

Why do we claim relatively small instructional resources will reduce redundancy and waste of resources? The answer lies in the way instructors use "professionally prepared" resources. Reigeluth and Nelson [3] found that the first thing teachers do with instructional resources is break them down into their component parts and rearrange them in a way that meets their individual instructional needs.

For example, imagine a high-quality reproduction of the Mona Lisa. The individual image would be useful to far more people than an entire art history lecture. The lecture (of low granularity or relatively large size) may only be useful to those teaching art history courses. However, the image alone would be useful to them as well as to anyone who wanted to incorporate the image into any kind instruction.

Instead of combining these component instructional resources in the traditional, monolithic way described by Reigeluth and Nelson, an online instructional support system could provide "anytime, anywhere" access to the component materials themselves--without instructors needing to break them down. This eliminates the first half of the deconstruction / reconstruction process, possibly providing a significant decrease in the time required to effectively use instructional resources.

Highly granular resources can increase the efficiency of online instructional support systems. Because archiving and delivery is so simple with networked technology like the Web, any online instructional support system with efficiency as its goal should be highly granular.

C. Current Issues
Although database-driven web sites can house individual resources, in order to be useful the individual resources must be combined with each other. HTML, the language of web pages, currently allows primitive combination of resources. However, issues such as synchronicity in presentation of the components thus combined are not addressed by HTML. Also, while HTML editors and automated converters have made it easy for instructors to create or convert materials into HTML, "publishing" those materials to the Web is still problematic for most. Organizations like the WorldWideWeb Consortium (W3C) http://www.w3.org/ and the Internet Engineering Task Force (IETF) http://www.ietf.org/) have standards initiatives underway to address these issues. Two of these are discussed below.

The Synchronized Multimedia Integration Language (SMIL, pronounced "smile") is being developed by the W3C to address issues of synchronicity [4]. Media that are combined using the language are synchronized in their presentation. One could create a SMIL file that instructed the client to "display this image, 7 seconds later roll this audio clip, 6 seconds later, fade to image 2, etc." With this kind of rich, coordinated presentation of materials made possible, instructional designers can create richer learning environments.

An IETF initiative, WebDAV, takes aim at making publishing easier for regular users [5]. The Distributed Authoring and Versioning effort attempts to create an open standard whereby users can publish their documents to a web server and maintain version control in a "one button" fashion. Because it is an open standard, WebDAV should bring this functionality (including utilities similar to Microsoft's proprietary FrontPage extensions) to all popular web servers.

High granularity alone does not increase the efficiency of a DLE. In fact, turning a few large resources into hundreds or even thousands of smaller ones can create quite a problem - that of finding the desired resources. This raises the issue of discoverability.

III. DISCOVERABILITY

A. Historical Context
When search engines first began indexing the Web, users were excited by the ability to perform full-text searches for their topics of interest. One could find just about anything. However, as the Web has continued to explode in size the unscalability of full-text searching has made our lives miserable. (For example, a full-text search for the term "granularity" turns up 49,497 results at AltaVista.) So if full-text searching is not the answer, what kind of search approach is? A can of soup suggests the answer.
Imagine going to the store to pick up a can of chicken soup for a sick friend. Once inside the store you locate the soup aisle and find, to your horror, a row of unlabelled, silver cans shining toward you. How can you be sure to get a can of chicken soup? Only by opening can after can and investigating the contents until you find chicken soup (assuming you know what it looks like).
This kind of search is the approach search engines initially took to indexing the Web-- open each page, examine contents, and record them individually for later searching. But it is evident that I can not find chicken soup this way unless I know what is in chicken soup. It is also evident that if there is another kind of soup with similar ingredients I will probably confuse the two. Finally, it must be evident that there is a better way to shop for soup.

B. Metadata
In reality it is easy to find the soup you are looking for because soup makers produce labels that include information about what is inside the can. You can find chicken soup without opening individual cans, without knowing what chicken soup looks like before hand, and without confusing it with something else. Standardized labels dramatically increase the efficiency with which one can locate a specific kind of soup. This type of standardized labeling system is necessary to make the use of highly granular instructional resources effective as well.
Metadata (or "data about data") is descriptive information about other information -- in this case, descriptive information about an instructional resource. Metadata can be indexed and searched in addition to, or instead of, the full-text contents of a resource. Metadata about an instructional resource might include general information like the title, description, and author's name; technical specifications such as file size or format, and instructional details like difficulty level, interest level, or instructional strategy. Each of these metadata fields provides a greater level of differentiation for each instructional resource, making it increasingly easy to locate [6].

The relationship between the discoverability of an instructional resource and the efficiency of the DLE hosting the resource should be readily apparent. Unfortunately, unless a standard approach can be agreed upon, metadata may only further confuse the process of discovering instructional resources, since incompatible systems may not be able to understand each other's metadata.

C. Current Issues
In early 1998, two of the major initiatives working toward instructional resource metadata standards, namely the European Union-sponsored Alliance of Remote Instruction Authoring and Distribution Networks for Europe (ARIADNE) and the EduCause IMS Project began collaborating on a common metadata specification under the auspices of the Learning Object Metadata Working Group (LOMWG) of the Institute of Electronic and Electrical Engineers' (IEEE) Learning Technology Standards Committee (LTSC). It is anticipated that when the LOMWG standard is finalized ARIADNE, IMS and everyone else will adopt the single, standard way of representing information about instructional resources.

At this point it may seem that the essential properties of an efficient system--reusable instructional resources that are easily locatable--are in place. However, experience has shown that there is at least one more property an efficient system will have, it if is to be widely used. This final property is accessibility.

IV. ACCESSIBILITY

A. Historical Context
Accessibility refers to the ability to get to a resource. Perhaps no one knows how long it was between the writing of the Web's first link and the existence of the Web's first broken link, but it can't have been long. Links that lead nowhere and files too large to download in a reasonable period of time using a modem are both part of the Web's heritage. From the very beginning, accessibility of resources has been an issue of great importance.

Whereas the difficulty with accessibility used to be a simple matter of broken links and huge files, today there are many more variations on this theme: including files that only work with an extra plug-in, files that only execute in a certain operating system, pages that only render properly in a certain browser, and many others. It is a fact of life that many of these incompatibilities are purposefully engineered in the name of market share. Any effective online instructional support system that will use the Web as its delivery medium must sufficiently address these issues so as not to jeopardize accessibility to its resources.

B. Current Issues
An efficient system should automate the process of link maintenance, providing appropriate warnings when previously live links go dead. In fact, a truly efficient system could do away with manually created one-way links altogether and employ automatically generated bi-directional links. Bi-directional links allow users to see what pages are linking to the current page (incoming links), as opposed to only showing the outgoing links to other pages. Some cutting-edge systems like the Hyperwave Information Server are already doing this.

Even though a number of technologies such as DSL and cable modems promise to alleviate bandwidth problems for off-campus students, an effective online instructional support system should still monitor file sizes, provide real-time information about estimated download times, and advise instructors regarding the use of different files in different circumstances. Slow dialup modems will continue to be common for years after faster technologies catch on. An efficient system should also be compatible with streaming technologies or employ an on-the-fly compression scheme like the Mozilla project's HTTP Compression, in which the web server compresses content before sending it to the browser, and the browser decompresses it on-the-fly before rendering it.

Finally, an efficient system should make it easy to set up and control access to resources. Of course the idea of restricting access to instructional resources flies in the face of accessibility as a property of efficient systems, but intellectual property considerations and other interests make this type of facility necessary. For the present, an efficient online instructional support system must deal with issues of security and access restriction to copyrighted resources.

V. CONCLUSION

In this article we have identified and discussed three common properties of efficient online instructional support systems: granularity, discoverability and accessibility. These three properties are interdependent; any two without the other are weak, and any one without the other two is useless.

Hopefully this article will encourage everyone involved - software vendors, instructional designers, and instructors themselves - to begin thinking about some of the critical issues surrounding online instructional support. One of most frequently cited arguments for online instruction is, "it can be done more efficiently." The properties of granularity, discoverability and accessibility are absolutely crucial to such a claim becoming a reality.

VI. REFERENCES

1. Nelson, Ted., OSMIC: Open Standard for Media InterConnection, 1996. Accessed December 17, 1999, http://www.xanadu.com.au/OSMIC/OSMICd1m.html
2. Wason, T., Editor., IMS Metadata Specification, 1999. Accessed March 23, 1999, http://imsproject.org/technical/metadata/library/DID188.html
3. Reigeluth, C. M. & Nelson, L. M., A New Paradigm of ISD? In Branch, R. M. and Minor, B. B., (Editors), Educational Media and Technology Yearbook, pp. 24-35, Englewood, CO, Libraries Unlimited, Inc, 1997.
4. Hoschka, Philipp, Editor, Synchronized Multimedia Integration Language, 1999. Accessed December 17, 1999, http://www.w3.org/TR/REC-smil/
5. Goland, Y., Whitehead, E., Faizi, A., Carter, S., and Jensen, D., HTTP Extensions for Distributed Authoring - WEBDAV, 1999, Accessed December 17, 1999, http://asg.web.cmu.edu/andrew2/rfc/rfc2518.html
6. Wason, T. & Wiley, D. A., Structured Metadata, Journal of Internet Cataloging, 3, 2,