Cog-Learn: An e-Learning Pattern Language for Web-based Learning Design

http://www.elearnmag.org/subpage.cfm?section=case_studies...

[See page 4 for comments]

By Junia Coutinho Anacleto, Americo Talarico Neto, and Vania Paula de Almeida Neris,
Federal University of Sao Carlos, Brazil

August 4, 2009

Abstract
Designing online learning material is a difficult task for novice teachers who lack experience in their design. Patterns have emerged as means to capture design knowledge in context and offer solutions to designers.

Cog-Learn is a pattern language aimed at supporting the design of learning material for e-learning systems. Here, we describe Co-Learn and discuss the patterns' identification and formalization processes through two case studies in which a set of cognitive strategies was applied with the goal of better organizing the content seen by the student. The purpose is to facilitate the student's interaction with the material's interface and, consequently, improve the learning process.

Introduction
Designing Web-based content for e-learning is a difficult task for novice teachers who lack experience in interaction and learning design for the electronic environment. The results are poorly designed courses and learning contents—for instance, text documents with too much information, which hinder the students' learning [10].

This research, supported by TIDIA-Ae project from FAPESP (process 03/08276-3), aims at designing learning material for Web-based e-learning and considers the different characteristics and knowledge of the multidisciplinary group that interact in such a project. We synthesize the cognitive science's proposals, expressed here as a set of cognitive strategies adopted by Liebman [11], some of them from Ausubel [2], and some concepts used during interaction projects on Web systems such as universal design, participative design, and accessibility. We have documented those practices in patterns to support the design of the learning material.

Considering such patterns, we propose to generate a common vocabulary among the participants of the multidisciplinary group that are responsible for designing the learning contents for e-learning (such as teachers, authors, educators, interface designers, software engineers, and Web designers), separating common qualities of existent designs, identifying successful solutions, and presenting the relevance of such solutions to help teachers better organize the content and thus benefit the students who are going to use it.

Here "teacher" is the professional responsible for designing the e-learning material, while "student" means the user that will interact with the developed Web interface published as learning content.

This article is divided as follows. The first section briefly presents the cognitive strategies theory and the group that we used in the case studies.

Second, we present the patterns and pattern language concepts.

Third, we show the methodology used to conduct the case studies, including the framework that guided the usability evaluations.

Fourth, we show the results from the case studies.

Fifth, we present the e-learning pattern language identified, its details, and some potentiality and restrictions.

Finally, we introduce a pattern-based tool to support the design of learning material, and end with some conclusions.

Cognitive Strategies to Support Teaching
Cognitive strategies are internally organized capacities that students use to guide their attention and manage their learning process. Gagné [8] relates these strategies to the "learn to learn" and "learn to think" concepts. Based on these ideas, researchers such as Beckman [3] and Ausubel [2] have thought about how to help students work with this information, which means using these strategies to facilitate understanding and information retention.

We adopted the following cognitive strategies used by Liebman [11]: organizing, framing, concept maps, metaphors and analogies, rehearsals, and advance organizers.

Another connection between our work and Liebman's [11] is the acknowledgement that teachers can use the cognitive strategies to facilitate the student's learning process. Teachers select and use the strategies when designing the learning material. The goal is to better organize the content through the interface and facilitate the student's learning process.

Patterns to Support the Design of Learning Material for E-Learning
Patterns were used first in the architecture domain [1] to represent success solutions for recurring problems found on this context. A pattern can be understood as an approach to capture and present design knowledge in problem resolution—acting as a knowledge spread tool between the specialist and inexperienced designers, and as a communication tool for the team members.

A pattern usually exists within a pattern language, and it is related to other patterns that offer solutions for other design issues in the same domain, aiming at involving the final user in all the stages of the software development process [5].

A pattern language which supports the learning material design for e-learning must have patterns to guide teachers in how to plan a lecture structure, as well as how to organize its content. The patterns must also help them plan the sequence of students' actions and provide support during the course, besides considering interaction design questions such as navigation and layout [16].

In view of that, we proposed a pattern language aimed at supporting teachers who design learning content and publish it in a learning management system. In this context, the teacher performs the user role when using the pattern language as a tool to establish more efficient communication and participate more actively in multidisciplinary learning design. She or he takes on the designer role when designing, evaluating, and making the learning material available, using the pattern language as a tool, which also conveys knowledge to less experienced teachers.

The patterns were obtained through two case studies whose main goals were 1) to verify if a set of six cognitive strategies [11] increases the usability of the learning material, and 2) if they can be considered solutions for recurring problems in this context and, this way, be written in a design pattern form.

The patterns' form and writing style are based on research by Meszaros and Doble [12], considering the elements name, forces, context, problem, solution, reasoning, examples, and related patterns.

Case Studies
The research method adopted by us is the case study [6] with qualitative analysis based on observation and questionnaires. We choose this method because if the study is not perfect, it is possible to obtain good results using qualitative methods that are based on users and their behavior [15].

The hypotheses used as a starting point to guide the case studies are:

Hypothesis 1. The use of six selected cognitive strategies improves the structure and organization of the content that will be published electronically for the student, and consequently, increases its usability.

Hypothesis 2. The cognitive strategies can be solutions to recurring problems in the Web-based learning design context and, thus, one can then write in a pattern form.

The method used to prove Hypothesis 1 is the accomplishment of two case studies, performing usability evaluations of the interface of two types of learning materials, one using the cognitive strategies and another not using them, in an effort to verify if there are usability issues in the materials designed without the cognitive strategies and if these issues were minimized in the materials designed with cognitive strategies.

The usability evaluations were planned according to the D.E.C.I.D.E. framework [17], which aims to support the planning and accomplishment of a usability evaluation. It was chosen using both an empirical (user tests) and analytical (heuristic evaluation with Web heuristics) evaluation method to identify a greater number of usability issues in learning materials elaborated with and without the cognitive strategies. During the usability evaluations, the focus was the items related to the organization and structure of the content by the interface.

The method used to prove Hypothesis 2 is the observation of where the selected cognitive strategies were applied to solve the recurring problems during the design of the learning material. This observation was planned with the support of a pattern language [12], which captures the best practices of the patterns' identification and writing process.

The patterns of this pattern language provide directives to understand the concepts of patterns and pattern languages, and means of structuring the patterns through the elements that compose the presentation form. This is useful in tasks such as identifying the pattern's subject, as well as the problem solved by the pattern, and indentifying the invariance of its solution, besides directives that must be considered in naming the patterns and making them more comprehensible to readers.

The dynamics used in conducting the case studies are:

1. The teacher creates the content that is made available as learning material (a hyper document), using Web knowledge and their teaching strategies (referred to as intuitive strategies).
2. A specialist in the selected cognitive strategies creates new learning material (a hyper document), starting from the material created by the teacher, containing Liebman's strategies, but keeping the original content elaborated by the teacher.
3. A pattern specialist (HCI and pedagogical patterns) creates a spreadsheet detailing where and what kind of cognitive strategies may be applied in the learning content and the comments made by the cognitive strategy specialist.
4. The material designed by the teacher and the material designed by the specialist in cognitive strategies is evaluated by two distinct groups of specialists in usability evaluation that perform the heuristic evaluation of this material, with the purpose of finding usability issues.
5. After the heuristic evaluation, the material created by the teacher and the material created by the specialist is made available to two distinct student groups during the user tests stage, in which two distinct evaluator groups observe the interaction of those users with the designed interface for the purpose of finding other usability issues. The evaluator groups then pick out testimonies from these users during their interaction with the designed material.

Proceeding with the previously mentioned steps, four sets of learning material were created and evaluated in two case studies.

[Continue to page 2]

 

Cog-Learn: An e-Learning Pattern Language for Web-based Learning Design
 
By Junia Coutinho Anacleto, Americo Talarico Neto, and Vania Paula de Almeida Neris, Federal University of Sao Carlos, Brazil

August 4, 2009

Page 2 of 4

[previous page]

Case Study 1. Case study 1 is composed of Learning Material 1.1 (M1.1), designed and intuitively organized by a teacher, and Learning Material 1.2 (M1.2), designed by a specialist in selected cognitive strategies. Both materials have information about material resources management in health institutions and were studied by students from the nursing department at Universidade Federal de São Carlos, Brazil (UFSCar).

A pattern specialist participated in the design of M1.2, pointing out the interface locations where the cognitive strategies specialist would insert a cognitive strategy, and asking why it was decided to insert such a strategy.

After the conclusion of M1.1 and M1.2, the heuristic evaluation phase began. We ended up with five reports for each material containing usability issues and one general report for each material, which compiled all the issues found by the evaluators according to the heuristic evaluation method.

In the user tests, five evaluators observed the interaction of two student groups that studied M1.1 and M1.2 separately. Each evaluator observed how the students interacted with the material and, after concluding the tests, a joint report was created that collected all the results. At the end of each interaction, each student filled out a questionnaire detailing the degree of satisfaction he or she experienced while using the material.

Case Study 2. In case study 2, two sets of learning material (M2.1 and M2.2) were also generated. M2.1 was designed and intuitively organized by a teacher, while M2.2 was designed by a specialist in the selected cognitive strategies. Both materials were about heuristic evaluation and were studied by computing students at UFSCar.

The pattern specialist participated in the design of M1.2, pointing out the interface locations where the cognitive strategies specialist decided to insert a cognitive strategy in the M1.2 design, and asking the why those inserts were chosen.

In the heuristic evaluation and user tests phases (step 5), the same procedures used in case study 1 were followed.

Results: Pattern Formalization
After completing the two studies, we tabulated and analyzed the results that were collected from the usability evaluation. Using heuristic evaluation, we looked at both the quantity of issues that were recorded, as well as their severity.

From the "Usability Guidelines for the Content Elaboration for Web," we collected the quantity of guidelines that were identified as "present" in the learning materials analyzed. In the user tests, we used the "Software Usability Measurement Inventory" (SUMI) questionnaire to note the time each student took to study the learning materials, the quantity and severity of issues found, average time to find concepts in the material, and user satisfaction.

The results collected were documented at great length [14]; a summary of the results follows:

 

• There is a recurrence of some issues in similar places in the learning materials. To resolve these issues, we used solutions based on the selected cognitive strategies.
• Study time (time spent in the interaction between the student and the user interface) was, on average, 21 percent less for the materials designed with the cognitive strategies than for those without them. Based on this, we believe that the applied strategies better organize the material studied, facilitating the students' reading process and the scheme construction.
• The SUMI questionnaire, composed of 50 questions proposed to quantify satisfaction of use, showed that the material with strategies yielded greater satisfaction among students, on average, about 14 percent more.
• We collected meaningful student testimonies, mainly from students who interacted with the materials designed without the strategies. They suggested a better navigational design, improvement of the system's status visibility and text structure, and the inclusion of examples. Moreover, some students who had interacted with the materials designed without the selected cognitive strategies (M1.1 and M2.1) explicitly asked for the use of the strategies which had been selected by the cognitive strategy specialist to be inserted in the material designed with cognitive strategies (M1.2 and M2.2).
• The learning material designed with cognitive strategies has, on average, 35 percent fewer issues detected by the heuristic evaluations and 50 percent fewer issues with a high severity rate than the material without strategies.
• The issues found in the user tests were compared by the researchers responsible for the case studies, and a final list of issues was obtained. The quantity of issues found in case study 1 was reduced by 54 percent, and in case study 2 by 44 percent.

   

  Given the interpretation and the analysis of the collected data from the usability evaluation, we concluded that the cognitive strategies can be a successful solution to the usability issues of the learning material designed for e-learning, and consequently, the improvement of the material's quality. We also concluded that it's possible to formalize the strategies in a pattern form to help teachers design this kind of material.

  Our conclusions were reinforced when the pattern specialist analyzed the tables [14], looking at where issues originated in the materials designed with the selected cognitive strategies. In abstracting those particular concerns, we were looking for generic texts, in other words, problems that did not refer to specific content, in the hopes that they could be documented in a pattern form.

  Here, we present the patterns formalized in this work.

  Pattern Name: Pattern Planning Known in Portuguese as "planejamento."

  Context: The teacher's first task is to elaborate the lecture by structuring the contents and creating a comfortable environment for the students. In this phase, the teacher defines the necessary foundation for conducting a good lecture.

  Forces: The preparation of a complete lecture involves the agreement of a series of concepts and interests, as well as the understanding that different publics have their own abilities and knowledge.

  The teacher generally has familiarity with the subject of the lesson; however, she can forget to mention topics that are important to a student's understanding of a subject. The Planning phase assists the professor in better organizing a lesson and facilitates the knowledge-transference process.

  Problem: How should one plan the knowledge transference from teacher to student?

  Solution: Formalize the problem to be solved by defining the final goal, which is helpful in determining the strategies for a lecture. It's also necessary to define one or more sub-goals of a lecture and to define which aspects the teacher wants to focus on.

  Rationale: A good plan is the result of experience. Teachers should be ready to adapt the pattern plan for subsequent versions of the same lesson, thus incorporating new experiences.

  Examples: Consider the design of learning material about the ozone layer. The teacher establishes the following plan with an objective of learning during the student's contact with the learning material. Having the plan, the teacher elaborates the concept map to organize the ideas and concepts that he or she would like to teach the students.

   

  Plan
  Identification	Lesson: "The Ozone Layer"
  Issue	Biology: Environment
  Prerequisite	To know the concepts of environment, atmosphere, and the conditions for preserving Earth's life forms
  Summary	Presentation of the ozone layer, its composition, localization, and the biological problems destroying it
  Objectives	How is the ozone layer formed? Why is it important? How is the ozone layer being destroyed? What are the biological consequences of this destruction?
  Concept Map	(Refer to image below)
  References	Encyclopedia Britannica

   

  

  Figure 1 A concept map can be used as an index.

  
  Pattern Name: Linkage
  Known in Portuguese as "gancho."

  Context: The teacher has performed the pattern planning of his or her lecture and already has well-consolidated learning goals. So it's necessary to start with the pattern planning, showing the student the main concept of the subject and if this concept is connected with other previously known concepts.

  Forces: Introducing a new concept can make students feel disoriented during the beginning of a virtual lesson, potentially leading to learning difficulties. The teacher can facilitate the student's learning process by encouraging them to remember concepts they have previously mastered and to relate those concepts to the new one about to be presented. And the new concept can be used as an introduction to the next concept that will be learned, preparing students to acquire new knowledge on the basis of what they already know.

  Problem: How can the teacher introduce students to a new topic?

  Solution: Use the "Cognitive Strategy Advance Organizer" [2]. Help students learn, remember, and connect the material they have already studied. New concepts can be potentially meaningful to students. Help them connect the new ideas to their existing knowledge.

  Rationale: We learn new concepts by relating them to skills that we are already familiar with. The Advance Organizer proposed by David Ausubel [2] therefore provides a clue and links our minds to previous knowledge and relates the new information to the old knowledge like a skeleton of the overall concept.

  Examples: In the beginning of a lesson about the ozone layer, an Advance Organizer is presented to explain that the Earth possesses a shield that filters out harmful radiation, stimulating the student to remember the concepts of ultraviolet radiation and atmosphere, and that this filter is composed of ozone gas and has been deteriorating.

  Pattern Name: Knowledge View
  Known in Portuguese as "Estruturação do Conhecimento."

  Context: The teacher has already stimulated the student's prior knowledge. In other words, the student knows how the new concept connects with concepts that she or he has already learned. The teacher must show the main concepts, as well as the content that must be learned.

  Forces: To keep the student active during a virtual lesson, the teacher can show the concept that will be presented, explored, and detailed. Therefore, the students are generally able to better remember the concepts they learn initially and have the capability to know the size of the lesson, its main topics, and the progress.

  The teacher can introduce the important ideas in the beginning of the lesson, even though these ideas cannot be fully explored immediately. Likewise, the students will have a preview of the content.

  Problem: How can the teacher introduce new concepts to the students?

  Solution: Use concept maps to create an index of the content for an e-learning environment. This environment must have one page for each node (concept) of the concept map and an index page that works as a link to the other pages.

  Concept maps are useful for many reasons. They are an observable record of the understanding of one person. They show how the information is relevant. They force students to think about their thinking and knowledge-building processes. Rationale: The concept map theory, proposed by Joseph Novak (1977), is based on the constructivist approach, understanding that individuals "construct their own idiosyncratic understanding of concepts."

  Examples: Figure 1 presents an example of an index for the lesson on the ozone layer based on the cognitive strategy concept maps. The navigational project of the learning material was made in such a way that each Web page of this material represents a set of topics that the teacher wanted to get across.

  [Continue to page 3]

  Cog-Learn: An e-Learning Pattern Language for Web-based Learning Design
  By Junia Coutinho Anacleto, Americo Talarico Neto, and Vania Paula de Almeida Neris, Federal University of Sao Carlos, Brazil

  August 4, 2009

  Page 3 of 4

  [previous page]

  Pattern Name: In Practice
  Known in Portuguese as "Contextualização."

  Context: The students are studying the learning material and would like to know how the learning concepts are connected with the environment where she/he lives and how they can be applied in practice.

  Forces: To keep the students engaged in a virtual lesson, have them visualize how the concept they are learning can be applied. Spark their motivation to make them use such concepts to solve problems.

  Problem: How can teachers exemplify the recently shown concept in the students' environment?

  Solution: Use cognitive strategy rehearsals to select the main information, proposing to facilitate the localization of an item and its identification in the text.

  Rehearsals are activities that help students process material in the short-term memory in such a way that the knowledge can be recovered later. Use cognitive strategy organizing, which in cognitive psychology is known as a partition that suggests the application of taxonomy, similarity and difference, function analysis, advantages and disadvantages.

  Examples: In a lesson about the ozone layer, the teacher uses the cognitive strategies: rehearsals to select the main information, proposing to facilitate the localization of an item and its identification in the text, and organizing to identify the causes and effects of moderate and extreme exposure to the sun. Such concepts are useful for the student, and the presentation form strengthens its importance.

  Pattern Name: Top-Down
  Known in Portuguese as "hierarquização."

  Context: The topics in a lecture are divided into fragments, which are introduced in an order that makes it easier for students to solve a problem.

  Forces: It's necessary that students know all the topics that they are going to study before learning each concept individually. Since the brain learns better when it can associate new subjects with preexisting knowledge, the more associations the better. Extensive topics can be broken down into many learning pieces and a lot of concepts. And all the concepts can be further divided into sub-items. The problem is that students often get an overload of information when all these components are put onto a text document. Students need to know beforehand which concepts are important before learning about them in detail.

  Problem: How to introduce a concept that has a large number of sub-items?

  Solution: First, sketch out the basics of the concept, and then gradually break it down in detail.

  Rationale: Significant learning is a continuous process in which more meaning is acquired when new connections are established among the concepts.

  Examples: In a lesson about the ozone layer, first show a summary about the origin of ozone, ultraviolet light, and the stratosphere localization. In this way, student get an overview before the details.

  Pattern Name: Means the Same
  Known in Portuguese as "correlação."

  Context: It's difficult to learn things beyond the range of our own experience. Complex metaphors can help.

  Forces: Students need a powerful and consistent strategy for wrapping their minds around a complex topic. The strategy must relate the topic being taught to the context of the students' daily lives. Too many details can easily make a student feel lost and unable to see how topics are related—particularly true when the concepts are strange or new.

  Problem: How to make the student see how the topic is connected to the main goal of the lecture and to understand how the concepts are connected.

  Solution: The proposed solution is the use of Cognitive Strategy Cognitive Metaphor and Analogies, which provide the students a faster way to think about the topic.

  Examples: In a lesson about the ozone layer, the teacher creates an analogy between the ozone layer and a shield that protects the Earth from the sun's harmful radiation. This metaphor conveys the same information in a different way to the student's memory.

  Pattern Name: Knowledge Retention
  Known in Portuguese as "sedimentaçã."

  Context: The student has studied a reasonable quantity of the learning material; now there is a need to retain this information while she or he prepares to acquire new knowledge.

  Forces: The brain can only concentrate on one topic for a limited period. After a certain point, students do not learn efficiently. To retain information, students needs to connect new information with what they already know, as well as apply it to real-life scenarios.

  Problem: How to keep newly acquired knowledge continuously working in the student's short-term memory, while she prepares to learn more.

  Solution: The proposed solution is to integrate the new material with the previous material, showing comparisons, relating new and old ideas, like considerations, tables, conclusions, and exercises.

  Examples: In a lesson about the ozone layer, encourage knowledge retention with exercises, with the objective of keeping the recently acquired knowledge in the student's long-term memory. For example, show the students figures of molecules and ask them to identify the ozone molecule.

  E-Learning Pattern Language Identified
  The case studies demonstrate that the applied cognitive strategies improve the usability of e-learning hyper documents. The studies also allowed us to identify seven patterns. Furthermore, during the learning material design that composes the case studies, some problems were identified that the cognitive strategies do not solve. These problems were related to the sequence of a virtual lecture (progress), as well as aspects related to interaction (navigation among topics), layout (color, interface item position), and to the lack of functionality (search) in the pages of the learning material. The identification of these problems motivated us to write this pattern language.

  Some pattern languages were studied, like pedagogic pattern languages [4, 7, 16] and HCI patterns [13, 19, 20]. We selected some patterns from these pattern languages that were related to the identified patterns, looking for the intended pattern language to support e-learning (Cog-Learn). These patterns include practices and knowledge that were documented in a pattern form and address a larger number of issues that the teacher could suddenly identify during the design of the learning material for e-learning.

  The Cog-Learn (the proposed pattern language for e-learning support) groups are:

  1. pedagogical patterns, which discuss pattern planning topics related to course pattern planning and lecture sequence and are based in practices identified in attendance lectures,
  2. HCI patterns, from Web design practices that approach the interaction and layout of the learning material,
  3. and hybrid pedagogical-HCI patterns, acquired from practices of inserting cognitive strategies in the context of design Web learning material.
     Because these patterns include practices from pedagogy that, when organized, improve the usability of the learning material's displayed content, they are called hybrid. An example of a hybrid pattern is the knowledge view, which proposes the pattern planning of the lesson using a concept map and later as a concept index. The concept map comes from the pedagogy, while the way that it is presented and its interaction plan are approached by HCI.

     The e-learning pattern language was elaborated from the pattern language theory [1] that visually connects the patterns in graph form. To facilitate the pattern language understanding, the identified patterns and the selected patterns were organized according to the symbols to sequential organization, grouping, and specialization used [7], as already shown in Figure 1.

     

     Figure 2. This is the scheme used to organize the patterns [7].

     

     Figure 3. The Cog-Learn, its groups, and abstraction levels are shown.

     By organizing the patterns identified in this work according to the previous description, we were able to represent the patterns in three categories, organized in two abstract levels, as shown in Figure 2. It's important to note that the pattern planning and interaction groups are composed, respectively, of pedagogical and HCI patterns, both selected in the literature. The content exhibition group is composed of patterns identified in this work (highlighted) and of pedagogical patterns selected from literature. All abstraction levels of this pattern language group the related patterns by filled arrows (black), which represent a specialization between them, and by empty arrows (red), which represent a relation of sequence between the patterns.

     In the following section, we present all the patterns that compose the Cog-Learn Pattern Language. It starts with the planning pattern. Each pattern includes a "related patterns" section, which enables the reader to move toward other patterns according to the kind of problem that occurs when the teacher designs the learning material.

     The E-Learning Pattern Language Detailed
     The teacher's first task, shown at the first abstraction level defined in this pattern language to support e-learning (Figure 3), is the pattern planning of a lecture with the definition of the desired learning results and the lecture objectives (it starts at the pattern planning pattern [18]).

     

     Figure 4. The pattern planning group of the pattern language is shown.

     The pattern planning pattern, which appears in the highest hierarchical level, when used, can result in other contexts that are treated by patterns in a hierarchical level below it. That is, during the lecture pattern planning the teacher may want to adapt the lecture to the students' abilities (discussed by the adapt to participant pattern [7]) and also give them the possibility of deciding on the way to conduct the lecture to be presented (discussed by the let them decide pattern [7]).

     After pattern planning knowledge transfer to the student, the teacher must observe the pattern language sequence, indicated with the open arrow that goes from the pattern planning pattern toward the active student [16], which shows the importance of active learning and finding a better way to keep the student participating in the learning process.

     The active student pattern results in new contexts, expressed in the patterns that show information about different activities susceptible for incorporation into a Web-designed e-learning environment. For instance, it is the explorative learning [4], which presents the importance of providing the student with a certain variety of resources and learning tools, and the collaborative learning [4], which discusses the possibility of sharing knowledge, discussing, and communicating. The virtual lectures pattern [4] discusses the importance of structuring a virtual lecture so that the student does not feel lost, and the searching [19] shows the importance of publishing a search tool to facilitate the study of the learning material. Both of them are resources that can be used by teachers to accomplish explorative learning.

     Next, observing the sequence defined by the open arrow that leads the virtual lectures pattern toward the linkage pattern [18], the teacher explores the second abstraction level of the pattern language (Figure 4), which treats the content exhibition, that is, a part of a lecture, a whole lecture, or even a complete course, depending on the subject's complexity and on the time necessary to present the knowledge.

     

     Figure 5. Content Exhibition group of the pattern language is shown.

     The second abstraction level is composed of two pattern groups; one treats the content exhibition, while the other discusses interaction topics, as presented next. These groups of patterns must be used sequentially to obtain a better result during the design of the learning material.

     The first pattern of the content exhibition group is concerned with the beginning of a new subject, which can mean the beginning of a new lecture. The teacher uses the linkage pattern to start what was planned before in the first abstraction level of this pattern language, and to relate the new material to concepts already known by the students. Notice that this pattern is not in a subgroup, which means the pattern language can be extended in the future by introducing patterns that specialize the current pattern, solving specific issues, for instance how to present a new lecture to the student.

     Continuing the sequence, still in the second abstraction level of the pattern language, the teacher is taken to the first pattern subgroup that treats the learning material, that is, it solves the common issue of how to introduce new concepts to the students, expressed by the knowledge view pattern [18], as well as more specific issues—like how to apply a concept recently shown in the student domain presented by the "in practice" pattern [18], or how to introduce a concept that has a great number of sub-items presented by the top down pattern [18].

     During the course from the first to the second subgroup of patterns, the teacher sees the breaks pattern [7], which shows the importance of taking regular breaks during the explication of a concept so that the brain can process the new information. Considering e-learning, breaks can be contents explicitly introduced in the learning material that show the end of a concept to start a new one or to start a new knowledge-retention stage, presented by the knowledge retention pattern [18], inside the third subgroup, in which the teacher shows students the comparisons, conclusions, tables, and exercises (discussed by the retention types pattern) so they can establish relations between what was presented and what they already knows—maintaining the new information for a longer time in the short-term memory, thus facilitating the assimilation and the storage of the knowledge acquired in the long-term memory.

     After the knowledge-retention stage, a cycle between the first and second patterns subgroups can be created, which means presenting a concept, establishing a break, establishing relationships or practicing what was learned, review after a break (review after break [7]), and starting the presentation of a new concept. At the end of the cycle, which may have many or no iterations, the teacher ends the lecture with summaries, references, etc. according to the bring the show to an end pattern.

     According to the case studies, during the design of the learning material, the teacher may come across the following issues related to the interaction between the student and the material (see Figure 6):

     • How to design the environment in which the student will be.
     • How much information should be shown in each interface?
     • How to prevent a student from getting lost in reading the content in different pages.

       Those issues can be solved using a set of HCI patterns that were combined to the patterns presented before (Figure 5).

       

       Figure 6. The Interaction Group of the pattern language is shown.

       These patterns can be considered a second-abstraction-level pattern language and can be used in parallel with the second patterns group, as shown in Figure 2. Thus, they were not included in the pattern language as a third abstraction level and as a continuation path (sequence) related to a specific pattern, but an alternative path that can be used during the learning material design to be published by the teacher.

       A fact that demonstrates this parallelism is the "bring the show to an end" pattern, listed in the second patterns group (Figure 4) because it discusses how to end a virtual lecture. As shown in Figure 2, after the teacher decides to use the virtual lectures pattern, he might come across the problem of how to set the environment for a lecture and decide that everything is ready to display the new material.

       The solution for this problem is described in the "set the stage" pattern [19] that starts the appropriate part of the pattern language e-learning to solve HCI problems. After setting the lecture's initial environment, the teacher follows the paths proposed by the pattern language and starts the design for content presentation through the navigable spaces pattern [19], which explores how to present the content in a way that allows the student to take it in at his/her own pace, in a comprehensible way, and one that keeps him interested. A specialization for this pattern is allowing the student to easily navigate through the interface and quickly become familiar with it (subject described by the repeated framework pattern [19]).

       Later, the teacher uses the "displaying page content" pattern [19] to decide how to use multimedia (texts, sounds, videos, and animations) to represent the information and worries about issues related to colors and graphics [13] that are used in the content. Next, the teacher may use the central working surface pattern [19] to get a central area in the interface to show the content, which will be the object of study. This content could be presented in an orderly fashion and in a way that makes sense to the student. These subjects are shown by the "overview beside detail" and "small group of related things" patterns [19].

       It can be necessary to set aside a space for the student to interact with this material. This topic is treated by the subgroup started by the control panel pattern specialized through actions such as progress, step by step, go back, and go forward [19].

       Note that it is not necessary to use all the patterns during the learning material design.

       [Continue to page 4]

Cog-Learn: An e-Learning Pattern Language for Web-based Learning Design
By Junia Coutinho Anacleto, Americo Talarico Neto, and Vania Paula de Almeida Neris, Federal University of Sao Carlos, Brazil

August 4, 2009
Page 4 of 4

[previous page]

Potentiality and Restrictions of the Proposed Pattern Language
We emphasize the possibility of inserting new patterns or substituting some existent pattern for another in the proposed pattern language. In the first abstraction level (Figure 3), which treats the pattern planning tasks, a new specialization can be introduced under the pattern active student, which treats other types of learning—collaborative learning, learning by doing, and learning by observing. Notice that collaborative learning was introduced in this pattern language, but it was not explained in detail because it is not within the scope of this work.

The data collected in the case studies was insufficient to think of more specific patterns for this pattern language. In the knowledge retention pattern [18], the learning material must try to integrate the new material with information previously presented through comparisons that reference new and old ideas, considerations, tables, conclusions, and exercises. The case studies provide sufficient data only to specify the knowledge retention pattern through the knowledge retention types. In future research the knowledge retention types pattern should include a detailed description about the motivation of using each one of knowledge retention items (considerations, tables, conclusions and exercises) and show examples.

The groups and subgroups of patterns also can be replaced or extended according to the need of each teacher. Group 3, which includes HCI patterns (Figure 5), could be replaced by another collection of patterns of the same context or it can be extended by patterns that discuss particular details, for instance, the colors issue treated superficially in this work due to the lack of patterns in the literature that address this theme.

The pattern language to support e-learning is not totally formalized, but that is not necessarily a problem, once it is known that one of the main pattern characteristics is the evolution from projects obtained in practice.

For this reason, we believe that the relationship and the detection of the lack of a pattern in this language also can be obtained from practice, as accomplished in the case studies.

Cognitor: A Pattern-Based Tool to Support the Design of Learning Material
Cognitor is a tool developed to assist teachers in their task of designing and editing high-quality instructional material for e-learning. The instructional material produced by Cognitor can be reused in other contexts of learning, therefore it is designed following the learning object concepts, and can be executed in many e-learning platforms, and therefore it can be exported in SCORM and HTML.

These two characteristics are basic for competition in the market of distance courses production and advantageous for the student and for the teacher. The advantage for students is that when the learning objects are well chosen, it can help their learning process. It can benefit teachers because they have available a great amount of learning objects. Thus they can plan their lessons making use of these objects, obtaining better flexibility to adapt to the rhythm and the interests of the students.

Beyond the previously mentioned advantages, Cognitor also offers aid by means of the computational representation of the Cog-Learn, expressed as functionalities to the teacher. These functionalities let teachers design and edit their instructional materials, using the success solutions for recurring problems that have been identified, widely used in real projects, and finally registered by specialists in the format of a common language of easy access for the ones involved in the e-learning project.

The instructional material is designed considering the implicit questions of the Cog-Learn patterns such as: the design of the lesson structure, as well as of its content; the project to elaborate the sequence of actions; the aid during the accomplishment of the course with the stimulation of the student's cognitive strategies; issues on interaction design (usability, accessibility, navigation and layout); and finally, considering topics of portability, content reuse and performance control by means of the development of contents according to SCORM.

These characteristics had provided the idea to use the Cog-Learn patterns as a framework to support the design and the edition of instructional material for e-learning, considering the pedagogical questions and the questions relative to the student's interaction with the interface.

The idea of expressing the Cog-Learn as a framework comes automatically, as well as in other domains, like software engineering, where the design patterns demonstrated to be so useful that many of them were included as a framework of tools that support the coding in certain programming languages, after standing out in the arsenal of tools and techniques for software development [9].

Cognitor (Figure 7) also offers the functionalities of design and reuse of learning objects by means of: an HTML publisher, a media aggregator, and the support of the pedagogical and HCI Cog-Learn pattern.

Figure 7. Cognitor's main interface is shown.

To create learning material, teachers may choose one document organization previously defined, create a new document organization, or use a pattern to help him in this task of planning the material. Afterward, the teacher can edit each page using the media insertion or the interaction design area. The interaction design area provides functionalities about either pedagogical or usability design issues like course design, content structure, navigation, layout, and learning activities.

One example of a pattern that is included as functionality of Cognitor is the knowledge view pattern [18], which supports the automatic generation of the content structure of the learning material by using the concept map theory. The teacher need only provide the Cognitor with the concepts, the links between them, and the names of these links. Cognitor builds the content structure and creates the pages and the links between them. So Cognitor provides useful information about the links between the related concepts previously defined by the teacher on each page generated with the pattern's aid.

After the content insertion, the designed material can be exported in HTML for execution in a Web browser, or in a SCORM format, with learning objects and metadata, for execution in a learning management system (LMS).

Cognitor also offers an innovative technique that considers the common sense knowledge provided by the Brazilian Common Sense project [Open Mind Common Sense Brazil (OMCS-Brazil)]. The aim of this project is to provide computers with knowledge and to promote natural interaction between the humans and the computer.

We decided to develop a common sense module to help the teacher when using the knowledge view pattern, which proposes the planning of the lesson using a conceptual map. When identifying concepts, teachers can query the common sense knowledge base to find related concepts, and include it in the conceptual map. This is made using a Wizard that guides the teacher.

Figure 8 shows instructional material designed using Cognitor and delivered in a Web LMS platform. The characteristics of this material are 1) it's composed by Assests and SCO's, described in an XML manifest in conformance with SCORM, and 2) its interface is implicitly designed using the design patterns of the Cog-Learn pattern language.

Some of the options offered by the Cognitor and the Cog-Learn pattern language were discussed here, but other ideas, as well as new patterns, will also be analyzed and incorporated in the other versions of the prototype tool, focusing on the quality and usability of designed material.

Figure 8. Cognitor also designs Web instructional material.

Conclusions
In this paper we have presented Cog-Learn, a pattern language for e-learning support which contains HCI patterns, pedagogical patterns, and the ones based on cognitive strategies, aiming at supporting the teacher in the task of designing Web-based learning material.

Such pattern language comes from the study of cognitive strategies application used by Liebman [11] and extended in a way to support teachers in the task of designing high-quality learning material for e-learning. The case studies provided the selection of patterns from the literature and the writing of new ones, presented here. They also allowed us to verify that cognitive strategies increase the usability of learning material for e-learning and, consequently, its quality.

Furthermore, we also present the Cognitor, a computer-based tool, and more specifically, a pattern-based editor that incorporates the Cog-Learn pattern language to support teachers in their task of designing learning material that promotes active learning, reducing knowledge-acquisition complexity.

In the future we intend to perform usability evaluations of the Cognitor tool in order to find out if it could be successfully used to promote a quick and efficient integration between the different professionals involved in the teaching and learning process in e-learning environments.

It is important that researchers and developers involved with e-learning realize that e-learning environments must be adaptable to teachers and students of varied fields, and that pedagogical support in the creation of material is needed, as well as targeting questions about the quality of the learning material designed. It is expected that this work will contribute to this objective.

References
1. Alexander, C. et al. A Pattern Language. Oxford University Press, New York, 1977.

2. Ausubel, D. P. Educational Psychology: A Cognitive View. New York: Holt, Rinehart and Winston, 1968.

3. Beckman, P. Strategy Instruction. ERIC Clearinghouse on Disabilities and Gifted Education Arlington. Educational Resources Information Center, 2002.
http://www.ericfacility.net/databases/ERIC_Digests/ed4743....

4. Bergin, J. "A Pattern Language for Course Development in Computer Science." Pace University, 2002.
http://csis.pace.edu/~bergin/patterns/coursepatternlangua....

5. Borchers, J. A Pattern Approach to Interaction Design. John Wiley & Sons, 2001.

6. Fidel, R. "Qualitative Methods in Information Retrieval Research." Library and Information Science Research, 15(3), 1993, pp. 219-247.

7. Fricke, A. and Völter, M. "Seminars—A Pedagogical Pattern Language about teaching seminars effectively." Proceedings of EuroPLoP, Germany, July 10, 2000.

8. Gagné, R. M. The Conditions of Learning, 3rd Edition. Holt McDougal, 1974. 9. Gamma, E., Helm, R., Johnson, R. and Vlissides, J. Design Patterns: Elements of Reusable Object-Oriented Software. Addison-Wesley, 1995.

10. Kessler, G., Rosenblad, K., and Shepard S. "The Web can be suitable for learning," Computer, 32(2), 1999, pp. 114-115.

11. Liebman, J. "Teaching Operations Research: Lessons from Cognitive Psychology." Interfaces, 28(2), April 1998, pp. 104-110.

12. Meszaros G. and Doble J. "Metapatterns: A Pattern Language for Pattern Writing," in Proceedings of the Conference on Pattern Language of Programming, Illinois, Sept. 4-6, 1996.

13. Montero, F., Lozano, M., Gonzáles, P. and Ramos, I. "A first approach to design web sites by using patterns," in Proceedings of VikingPLoP Conference, 2002.

14. Neris, V.P.A, Talarico Neto, A, Silva, J.C.A, and Mascarenhas, S.H.Z. "Hyper Documents with Quality for Distance Learning: Cognitive Strategies to Help Teachers in the Navigational Project and Content Organization," in Proceedings of the 11th Brazilian Symposium on Multimedia and the Web, Poços de Caldas, 2005.

15. Nielsen, J. Usability Engineering. Academic Press, Cambridge, 1994.

16. The Pedagogical Patterns Project, Retrieved 2001: http://www.pedagogicalpatterns.org.

17. Preece, J., Rogers, Y., and Sharp, E. Interaction Design: Beyond Human-Computer Interaction. John Wiley & Sons, 2002.

18. Talarico, N. A.; Silva, J.C.A.; Almeida, V.P. "Padrões para Apoiar o Projeto de Material Instrucional para EAD," in Latin American Conference on Pattern Languages of Programming-SugarLoafPLoP, 2005. /in Portuguese/.

19. Tidwell, J. "User Interface Patterns and Techniques": http://time-tripper.com/uipatterns. 2003.

20. van Welie, M. Patterns in Interaction Design, 2003: http://www.welie.com.

21. West, C. K., Farmer, J. A., and Wolff, P. M. Instructional Design: Implications from Cognitive Science. Allyn and Bacon, 1991.