Design

Digital learning resources can be viewed as consisting of several interdependent layers, shown in the following table.

Five Layers of a Digital Learning Resource

Layer	Description
Context	Language, cultural knowledge, subject knowledge, relations to other learning resources and other factors that are needed to properly interpret the resource.
Pedagogy	How a digital learning resource is used as part of a learning strategy or instructional design.
Structure	How a digital learning resource is structured into assets, information objects, learning objects, etc. and how these are navigated or sequenced.
Content	The information that is contained in a resource and that is intended to affect a change in cognitive state.
Presentation	How a resource is rendered and what visual and auditory elements will be used to render it

Being aware of the effect of each layer on reusability will help guide design choices and reducing interdependence among layers will enhance reusability.

Context

Learning makes use of language, relies on culture, requires prior knowledge and experience and depends on the situation in which it takes place. If the dream of technology assisted learning is to get just the right stuff to the right person at the right time in the right way (Hodgins, 2002), then context is what determines the value of a learning resource.

At the same time, contextual dependencies limit the potential audience of a resource. Inherent contextual dependencies make it harder to use an object in multiple settings and in multiple ways (Robson, 2003; Gibbons, et al., 2003). Thus context is the friend of learning and the enemy of reuse. For this reason, every effort should be made to reduce unnecessary contextual dependencies.

Another way to say this is that the pedagogy, structure, content and presentation of a resource should be as free from dependence on external context as possible. An explanation that cannot be understood without referencing a specific text or external online resource is best replaced by one that may require specific knowledge but that does not require a specific source. An image, example or test question that clearly depends on cultural knowledge for interpretation should, if possible, be replaced by one that does not. A document consisting of a list of assertions may be useful for some instructional methods but will be more reusable if enough scaffolding is provided to make it useful to anyone who reads it. A large resource will be more useful for adoption† and more disaggregated into information objects† and learning objects† for adaptation† if it does not require a lot of inside knowledge to discern the boundaries among presentations of facts, statements of opinion, content intended for a student, remarks made for the benefit of an instructor, exercises, etc. Even for a well-crafted resource, explicitly providing the metadata to identify the substance and nature of components will make it easier to reuse than will relying on context.

Pedagogy

A learning resource is more valuable if it can be used for as many different types of learning as possible. It therefore helps to separate any instructional or learning strategy implicit in a resource from its structure, content and presentation. It should be noted that the intended pedagogical setting and instructional use of a resource are contextual issues, but we single out pedagogical context because we are talking about learning resources, not arbitrary digital content.

As an illustration, consider an educational Web site broken into sections that include some explanatory material, some exploratory material and an online quiz. Suppose that the site is intended for use by middle school children under the supervision of a classroom teacher who will guide the students through the material in a particular order. This is the pedagogical layer.

The intended pedagogical context and instructional design can show up in the choice of graphics and fonts (presentation), in references to the teacher embedded in the Web pages (content), and in the navigational scheme (structure). If material from the site is to be reused or repurposed for use by students learning on their own, or by parents helping their children, or by adult learners, then changes will have to be made to all of these elements. Here are some ways this can be made easier by keeping the pedagogical layer separate.

Reusability will be enhanced if a separate screen is used for navigation and there are no “previous” and “next” buttons or hyperlinks among the sections. This will allow other navigational schemes and instructional designs to be imposed on the same underlying content. For example, someone adapting the middle school site for adult learners might add a pretest and design a system wherein the learner doesn’t see sections passed on the pretest but can see the remaining sections in any order desired. The IMS Simple Sequencing and Learning Design specifications provide standardized ways to separate pedagogy from structure and to implement these types of designs.

If entangling the structure of a resource with its instructional approach is bad for reusability, incorporating an instructional design into the content is even worse. If the premise for all of the content in the middle school Web site is a hands-on experiment that requires materials and supervision only available in a classroom, it will be hard to use any of the content at home. Another problem could arise if the content is targeted at teachers, rather than at learners. Providing a separate teacher’s guide and dividing hands-on experiments into separate sections increases the separation of content from pedagogy.

Finally, the reusability of the site will be greatly enhanced if the presentation elements do not scream “middle school classroom.” Optimally, the presentation layer is kept separate from all others so that repurposing† is easy, but for straightforward reuse it is best if sections of the site can be linked from other resources with other audiences and designs without the presentation and style getting in the way.

Structure

As has already been discussed, many reusers† of learning components only use parts of them. For both adoption and adaptation it is advantageous for a resource to be easily disaggregated at least into learning objects. Content that does not have this property is derogatorily referred to as “monolithic” in the learning technology community. On the technical side, there are excellent technologies (metadata, learning technology content standards and XML) that aid in separating out the structure of a resource. These will be discussed later.

Reusing Structural and Pedagogical Layers

Separating the structure of a resource from its other layers not only increases the reusability of the content but also allows the structure itself to be reused. The same is true of its pedagogical design. In other words, a well-structured and well designed resource can become a template for others.

Structural and design templates are needed to make optimal use of libraries of content assets†, information objects and learning objects. This is reflected in the following quote taken from a Morgan and Keegan white paper (Ruttenbur, et al., 2000) written for the corporate training world:

“By analyzing the learner’s objectives and existing skill level, courses will be assembled on the fly that address exactly what the learner needs to know without wasting time working on areas in which the learner is already proficient or uninterested. This level of personalization will be achieved by using small chunks of information, or learning objects, to assemble a course from the ground up using pre-existing templates. The reusability of these learning objects will make this level of customization feasible in terms of both time and expense.”

Although a different pedagogical approach may be more appropriate in educational settings, the value of learning objects is still predicated on the ability to reuse and assemble them (Reference to be added).

Presentation and Content

Content, and the way it is presented, is the “stuff” of a digital learning resource. Pedagogical designs and structure can also be reused, but much of the focus on reusability is on content.

If content is to be adapted, it should be separated from its presentation. This means it should be possible to easily change font styles, layouts and branding elements such as logos and color schemes , or to render the same content on different devices. If graphical elements and branding are hard-coded into the content, then extra effort is required to remove or replace them. The next section discusses how to achieve the separation of presentation and content that is needed for reusability.

Separating Presentation from Content Using Styles

As an illustration, consider a HTML document and would like to start a new section called “Rebuttal.” Here are two contrasting ways to do this:

Put in a blank line, format the word “Rebuttal” on a single line in a larger bold font, and insert another blank line after it. In HTML this would look like:

 Rebuttal 

Renders in a browser as:

Rebuttal

An alternate method would be to use a style sheet:

Apply a style sheet with the style “section_title” defined. In HTML, this requires either creating a separate document with an entry like:

.section_title {font-size: larger; font-weight: bold;}

and formatting the section title as:

Rebuttal

Rebuttal

The difference between these two approaches is that the information passed to the browser about the word “Rebuttal” is syntactic in the first approach and semantic in the second. In the second approach, additional and separate information (a style sheet) allows the browser to convert the semantics into syntax. This permits the HTML document to be completely reformatted simply by changing the style sheet. Furthermore, applications other than a browser could render the document correctly provided they have a way to interpret the concept of a section title. A device that reads the content might know to pause and say the phrase “new section” before reading the word “Rebuttal.”

Separating Presentation from Structure Using XML

Styles separate presentation from content and their use is applicable to word processors and to mark-up languages (such as TEX) as well as to HTML. But they do not completely solve the problem of identifying the components of the content. For this it is necessary to separating presentation from structure.

A good way to separate presentation from structure is to use eXtensible Markup Language (XML). An XML document can be viewed as a set of containers. The document is the largest container. Every container other than the document is properly contained in a larger one. Everything inside of a container is either content or another container.

Each container in an XML document is identified by an opening and closing pair of tags. Tags provide semantic information about the content in the container and scope the meaning of any containers that are nested within them. Thus an XML document might be divided into sections which, in turn, contain titles, bodies, footnotes and references. The fact that a title is inside of a section makes it the section title, whereas if the title were only inside the document container, it would be the document title.

An XML document is an example of “structured data.” XML is by no means the only way to represent structured data, but it is a standardized and increasingly prevalent way. XML has several advantageous from a reusability standpoint:

An XML document can be disaggregated into semantically meaningful chunks.
The presentation aspects of an XML document can be determined by the nature of the component being presented. In other words, different styles and rendering methods can be assigned to different types of structures. This separates presentation from content in an even more meaningful and flexible way than applying styles directly to the content itself.
In XML, the structure of a document is expressed through a set of tags and the rules for the tags are controlled buy a separate document (either a Document Type Definition (DTD) or Schema). If two products use different labels for the same information, it is easy to do a re-labeling. Thus if one product creates content with “Exercises” and another product uses the term “Practice Questions,” it is relatively easy to write a translator that allows the products to correctly interpret and render each other’s content.

Comment on HTML and Word Processors: HTML is designed for marking up the presentation of a document. Word processors were designed for the same thing. The addition of style sheets in HTML allows a good degree of separation of presentation from content, and someone adept with a product like Microsoft Word™ can do the same. Still, neither HTML nor most world processors were designed to separate presentation from structure. Fortunately, there is now an option to use XHTML, which is an XML dialect that newer browsers will be able to transform into HTML and display. Similarly, commercial world processors will be (or are) able to create XML output. This is not specific to learning content but will increase the potential for reuse.

Design and Granularity

As might be expected, the importance of individual layers depends heavily on the granularity of a learning resource. The following table indicates what layers are most important as a function of granularity:

Design as a Function of Granularity

Granularity	Layer
Content Asset †	The key issue for content assets is separating presentation from content. Contextual dependence should be avoided.
Information Object†	For information objects, separation of content from presentation is important, and it is also important to avoid cross references that entangle the content with the structure, pedagogy and context.
Learning Object†	The considerations for information objects apply to learning objects. Additionally, there is a danger of hard-coding navigational elements and unnecessarily tying the object to a particular pedagogical approach or assumed context.
Learning Component†	As the aggregation level increases, reuse shifts to component reuse. Therefore issues of separating pedagogy, structure and content become more crucial for reuse. Pedagogical approach and contextual dependence become the limiting factors for reusing or repurposing learning components in their entirety.
Learning Environment†	Learning environments may be designed for a specific context and pedagogical approach or may be more general. The more general ones are those that are reusable. For them, it is important to avoid cross-linking of components of the environment.