Prescribing e-Learning Activities

Using Workflow Technologies

Alberto B. Raposo

, Mariano G. Pimentel

, Marco A. Gerosa

Hugo Fuks

, and Carlos J. P. de Lucena

TeCGraf / Dept. of Computer Science / Catholic University of Rio de Janeiro (PUC-Rio)

Lab. of Software Engineering (LES) / Dept. of Computer Science / PUC-Rio

R. Marquês de São Vicente, 225, Rio de Janeiro, Brazil - 22453-900

Abstract. In this paper we propose a new coordination approach for the Au-

laNet Learning Management System. The proposed coordination protocol is

based on a workflow mechanism that could be used by teachers to define some

specific sequence of activities within a course run within AulaNet. This way, at

least part of the social protocol that governs the flow of the course could be en-

forced by software.

1 Introduction

The necessity of coordination mechanisms to regulate interactions in groupware sys-

tems has been the center of a heated discussion. At one side, there are normative

models that try to regulate the collaboration by restricting the interaction between

participants and their tasks. The criticisms on such normative approaches may be

roughly summarized by the fact that their rigidly defined protocols applies only to

very specific scenarios, limiting the flexibility of the collaborative systems. Eventu-

ally, there would be situations not predicted by the specified protocols, restraining the

application of this kind of coordination approach.

At the opposite side, are those advocating that collaborative systems should take

flexibility to the extreme, leaving the coordination burden to the users. The criticism

on this kind of approach is that they augment the coordination workload, since users

must deal with the complexity of articulating their tasks. Moreover, giving the coor-

dination responsibilities for the users does not guarantee that the activities will be

performed according to any prescription.

The discussion makes sense because there are different kinds of collaborative ac-

tivities. There is a large group of activities (tightly integrated collaborative activities)

whose tasks depend on one another to start, to be performed, and to end. They require

sophisticated coordination mechanisms in order to be supported by computer systems.

A coordination mechanism is defined as “a specialized software device, which inter-

acts with a specific software application so as to support articulation work” [11].

On the other hand, coordination does not need to appear explicitly in some kinds

of computer-supported collaborative activities—loosely integrated collaborative ac-

B. Raposo A., G. Pimentel M., A. Gerosa M., Fuks H. and J. P. de Lucena C. (2004).

Prescribing e-Learning Activities Using Workﬂow Technologies.

In Proceedings of the 1st International Workshop on Computer Supported Activity Coordination, pages 71-80

 SciTePress

tivities [9]—such as those realized by means of chats or audio and videoconferences.

These activities are associated with social relations and generally are satisfactorily

coordinated by the standing social protocol, characterized by the absence of any com-

puter-supported coordination mechanism among the activities, trusting users’ abilities

to mediate interactions. The coordination, in these situations, is culturally established

and strongly dependent on mutual awareness (to be aware, in this context, is to ac-

quire information about what is happening and what other people are doing [3]). The

coordination workload and the limited flexibility imposed by rigid coordination ap-

proaches are completely awkward for loosely integrated collaborative activities,

which are more suited for the awareness-based approach.

In spite of this discussion, there is a trend to conciliate both ideas, arguing that

both kinds of activities are “seamlessly meshed and blended in the course of real

world” [12]. This is especially true in e-learning, which is the focus of the paper. At

one side, there is the necessity to “create software supports that enable learners to

follow a range of paths to completing projects” [4]. At the other side, there is the

necessity to have a flexible environment where learners are able to determine their

own paths, according to their knowledge background, time constraints, or access

limitations.

In this context, this paper presents the experience of using AulaNet, a Learning

Management System, to run a course called ITAE (Information Technology Applied

to Education). The course has a well-defined dynamics, and the coordination is cur-

rently based completely on the social protocol. Having this experience as background,

the goal of this work is to explore the possibilities of using workflow technologies

within AulaNet not only to coordinate the course’s dynamics, but also other aspects

of learning activities, trying to conciliate the flexibility of the social protocol with the

normative regulation offered by workflows.

2 AulaNet and its Use in a Distance Course

In this section we present the AulaNet environment and the dynamics of a course that

uses it. The course dynamics suggests the use of workflows (section 3) to coordinate

the activities that the learners must realize through AulaNet services (section 4).

2.1 The AulaNet Environment

AulaNet is being developed since June 1997 at the Software Engineering Laboratory

at the Catholic University of Rio [7]. It is freeware, and is available for download in

Portuguese, English, and Spanish versions (http://www.eduweb.com.br). In order to

access the different services available at the course, the learner uses a menu repre-

sented by remote control unity (Fig. 1).

The AulaNet is based on a groupware approach. To work in group, people need to

exchange information (i.e., to communicate), to organize themselves (to coordinate)

and to act together in a shared space (to cooperate). Awareness, which is the act of

acquiring information through the senses, also plays a central role in this model.

These three concepts—communication, coordination and cooperation—are the main

elements of the 3C model for collaborative work (Fig. 2a) [7]. Based on this model,

groupware applications are classified according to their support to each of these con-

cepts, being located in a triangular space presented in Fig. 2b [2].

Conference

Remote

Control

Debate

Fig. 1. Services accessed by the remote control

arranges

that are managed by

demands

mediates

Communication

Coordination

Cooperation

Group

Awareness

generates commitments

fosters

mediates

fosters fosters

tasks for

Communication

conferencing

systems

message

systems

workflow

Coordination

Cooperation

Message to Participant

Contact with Teachers

Lesson Plan

Follow-Up Reports

Bibliography

Webliography

Documentation

Tasks

Discussion List

Co nference

Debate

Exams

electronic

meeting rooms

shared information

space

group editors

inetlligent

agents

Co-authorship

Notices

Fig. 2. a) The 3C Collaboration Model; b) Classifying applications according to the 3C model

The communication services of AulaNet allow the message exchange among the

participants of a class. The Discussion List and the Conference services enable the

asynchronous message exchange among participants. The Contact with Teachers

service offers private and asynchronous communication between a learner and the

teachers. The Debate service is a chat tool for synchronous communication among all

the participants. The Message to Participants service is an instant messenger that

establishes the synchronous message exchange between two participants.

The coordination services of AulaNet were designed for teachers. The Lesson Plan

enables the teacher to organize the course contents. The Exam enables the creation of

multiple-choice questions for an evaluation of the learners. The Follow-up Reports

service presents information about learners’ participation within the course activities.

The cooperation services define different spaces for information exchange. The

Bibliography and Webliography services offer the teacher the possibility to create

references for documents. In the Documentation service, didactic content can be

disposed dissociated from the Lesson Plan. Finally, the Task service defines an activ-

ity in which learners must generate content that will be available to the other learners.

The creation of a course with AulaNet consists in the selection and configuration

of the services that are going to be used during the course, which may be done in

advance or on the fly, during the course.

2.2 Dynamics of a Course using AulaNet

The ITAE course has been taught since 1998 as a regular course of the Computer

Science Department at PUC-Rio and is entirely taught via the Internet within the

AulaNet environment. The main objective of the course is to train educators to use

the new information technologies for teaching and learning in a collaborative way [6].

The ITAE is organized as a set of collaborative activities. Table 1 presents the pro-

gram for the first semester of 2003. This table defines the activities, their dates and

the AulaNet services used in these activities.

Table 1. Program of ITAE for the first semester, 2003

Phase Activity AulaNet Services

Introduction

13/03 (12AM) – Inaugural session

13/03-24/03 – Learners self introduction

None (classroom)

Discussion List

Study and discus-

sion of the

course’s topics

14/03-20/03 – Introduction to AulaNet and ITAE

21/03-27/03 – Groupware and Digital Communication

28/03-03/04 – Web based Instruction (WBI) and Classroom

04/04-10/04 – Learningware

11/04-24/04 – Facilitating WBI and Learning Concepts

25/04-08/05 – Teaching, Learning and Implementing WBI

09/05-15/05 – Interactive Multimedia and WBI Design

16/05-22/05 – New Trends in WBI

Conference

and

Debate

Content genera-

tion and publica-

tion

23/05-16/06 – Content generation about a specific topic

17/06-23/06 – Collaborative evaluation of the contents

07/07 – Publication of the revised contents

Task

Conference

Task

Closing 14/07 – Final degrees publication Discussion List

The Introduction comprises a face-to-face session in the classroom (the only one)

where the course’s dynamics and guidelines are presented. Then, learners introduce

themselves to each other using the Discussion List service. This first message is used

to tell their expectations, how they learn about the course, etc.

The ITAE course is organized by topics. Every week during the Study and Discus-

sion phase, the learners should study all contents regarding the weekly topic. From

12AM Monday to 2PM Wednesday all learners have to send their messages to the

conference seminar where that week’s topic is being discussed. On Thursdays from

12AM to 1PM all learners join the weekly debate to round up that week’s topic study.

After the Study and Discussion phase, learners are subdivided into small groups to

generate content based on a specific topic determined by the mediators. The content

prototype is submitted to the other group’s appreciation using the Task service. The

course closes with the publication of the final degrees in the Discussion List. This

degree is calculated based on the quality and quantity of the messages sent by the

participant to the conferences and to the debates, together with the general apprecia-

tion of the content generated by its group.

Currently AulaNet has no workflow mechanism to manage the phases of the

courses. In the ITAE case, a Learners’ Companion Guide is made available at the

beginning of each semester defining dates and activities. The coordination of these

activities is done by mediators, enforcing the social protocol sending messages to the

Discussion List and Message to Participants services. Therefore, AulaNet lacks a

workflow service to formalize at least partly of what is defined in the Guide.

3 Workflow Technologies

The Workflow Management Coalition defines workflow as “a process definition that

consists of a network of activities and their relationships, criteria to indicate the start

and termination of a process, and information about the individual activities, such as

participants, associated IT applications and data, etc.” [13]. Basically, a workflow

defines procedures composed of a sequence of activities, establishing their relations,

the resources needed for the execution of each activity, and criteria for activities ter-

mination. Moreover, a workflow may also define roles or groups that indicate who is

responsible for the execution of each activity. Two important contributions of work-

flow technologies are the separation of the business process logic from the implemen-

tation of activities and the connection of independent activities, allowing the migra-

tion “from islands of automation to system support for the overall business process”

[8].

“Conventional” workflow systems fit in the normative coordination approaches,

being more adequate to tightly integrated collaborative activities. However, it is rec-

ognized that, in order to be useful for the majority of real world procedures, they

should be flexible [5].

In order to refine the interdependency relations among activities and to enhance

the flexibility of the workflow representation, we use some dependency operators [9]:

enables, forces, blocks and unblocks. The enables operator represents what is called a

passive interpretation of the relation. In this case, “activity A enables activity B”

means that activity A is a prerequisite to the execution of B, but the execution of A

does not imply the execution of B. This operator is more relaxed than forces, which

represents the opposite situation. If “activity A forces activity B”, then activity B

must be realized after activity A. The blocks operator indicates that the execution of

an activity blocks further executions of another activity.

In the following section it is shown how the workflow approach is used to model

ITAE dynamics, serving as a means to implement a computational support for the

course coordination. Especial attention will be given to the trade off between the

rigidity of a workflow and the flexibility of the current social protocol. Moreover,

some other possibilities of using workflow technologies within AulaNet are also

explored.

4 Coordinating Activities in AulaNet

In section 2.1 it was shown that AulaNet offers coordination services. Although these

services are related to coordination, they actually stand for awareness instruments for

teachers and learners, which will use the information provided by these services to

guide their activities. For example, the Lesson Plan is used to organize the course

contents and offers the learners an indication of the order in which the contents

should be navigated. However, this does not guarantee that the learner will follow the

course sequencing, or even that any content will be opened. Therefore, one may con-

clude that the current coordination services of AulaNet are instruments to support the

social protocol, which is actually the coordination means currently available within

AulaNet.

Nevertheless, a workflow service may be used within AulaNet to coordinate some

of the educational activities. The first step in any workflow implementation is to

model the process that needs to be coordinated. For the AulaNet context, the ITAE

course is a good starting point, because the course process is well defined and docu-

mented in the Learners’ Companion Guide, as presented in section 2.2.

Fig. 3 shows the workflow of the ITAE course from a learner point-of-view. It

starts with the course and members introduction. Then, there are 8 content studies,

each of them composed of content readings, seminar and synchronous debate. Fi-

nally, there are the content production, which is also subdivided into three sub-

activities, and the course finalization, with the degree assignment.

Course

introduction

Members

introduction

Readings 1

Sem in ar 1

Debate 1

Topic 1

Readings 2

Sem in a r 2

Debate 2

Topic 2

Readings 8

Sem in ar 8

Debate 8

Topic 8

Prototype

submission

Prototype

evaluation

Prototype

re-submission

Degree

Assignment

Content study

Content production

Course finalization

Course initiation

Fig. 3. Workflow of ITAE activities: Overall view

The overall view of the course workflow indicates that it is necessary to have a hi-

erarchical representation of activities. In Fig. 3, for example, the composed activities

represented inside the boxes are internally subdivided into more detailed sub-

activities. For example, the content production by the learners has three sub-activities,

namely, prototype submission, its evaluation by other groups of learners, and proto-

type re-submission. Getting into an even lower abstraction level, it is necessary to

decompose some activities until atomic ones and detail the kinds of interdependency

that exist among these activities.

When working with “conventional” workflows, the non-execution of a previous

activity by the person who is supposed to do that implies that the workflow stops in a

certain state. In the ITAE course, there is an additional timing factor that determines

when an activity will be declared finished. For example, if a learner does not appear

in the face-to-face class, her workflow begins in the second activity. This unusual

characteristic in workflow settings is essential to keep the course dynamics and indi-

cates that the use of workflows in e-learning systems requires especial flexibility

features.

Returning to the course workflow, Fig. 4a presents the expanded workflow for the

weekly seminar activity. This activity involves three roles. The mediator (teacher)

selects the learner that will play the seminar leader in that week and initializes the

seminar session. The leader must then submit the seminar message to the conference

and propose a number of questions to be discussed. The discussion takes place by

means of message submissions by the learners to the conference. Each of these mes-

sages is evaluated by the mediator, who may also finalize the session.

Se le ct

seminar leader

Se m i n ar

submission

Questions

submission

Message

submission

Message

evaluation

Conference

session

finalization

Conference

session

initiation

Readings

Debate

Mediator Seminar Leader Learner

enables

forces

enables

forces

enables

blocks

Select

debate leader

Sum m ar i ze

the seminar

Present

a question

Commenting

on the question

Declare

debate session

initiated

Se m in ar

Mediator Debate Moderator Learner

enables

forces

Vot e on

a contribution

Free discussion

on the selected

contribution

Draw your

conclusions

Declare

debate session

finalized

Message

evaluation

enables

Readings

forces

blocks

Fig. 4. a) Expanded workflow of a seminar. b) Expanded workflow of a debate

In Fig. 4a there are also indications of the kind of interdependency among activi-

ties, expressed by enables, forces, and blocks operators. An example of enables rela-

tion takes place between the leader’s question submission and the learners’ message

submission. Learners are not able to submit messages before the leader’s questions,

but these questions do not oblige each learner to submit messages. Actually, the non-

participation of a learner may have a negative impact on her degree, but this does not

harm the procedural flow of the seminar. Other activities are connected with a

stronger relation. For example, the session initialization by the mediator forces the

leader to submit her seminar, followed by the questions, otherwise the seminar would

fail. The blocks operator is also used, indicating that learners cannot submit messages

after the mediator finalizes the seminar session. This doesn’t imply that the mediator

won’t be able to continue evaluating the remaining questions. The mediator must

evaluate all questions; some of them may be evaluated after the seminar finalization.

The examples of Figs. 3 and 4a show that the teacher should model the course ac-

tivities, specifying the services, learners and sub-activities involved. An activity may

comprise several sub-activities, different participant roles, and also different services.

This is not an easy task, but it is important not only to enable the course coordination,

but also to offer a pictorial view of the course dynamics, facilitating its re-design.

Once the course is modeled, it will be straightforward to modify its dynamics.

Another important feature that will be obtained with the use of a workflow based

coordination approach within AulaNet is the possibility to create different workflow

paths for different learners or groups over a single course. For example, a more ad-

vanced class may skip introductory contents, while novices should study them.

4.1 Case Study: Coordinating the Debate

On ITAE debates, one participant is selected to play the role of moderator, with the

job of coordinating other participants and presenting topics to be discussed about the

subject that was studied during the previous activities. On recent classes of ITAE, a

well-defined sequence of activities has been used on debates (Fig. 4b). The mediator

declares the debate session initiated. Then, the moderator summarizes the seminar and

formulates the first question. Next, each participant sends a message commenting on

the question. After all participants sent their comments, they choose which one must

be discussed. Then, everybody got involved in a free discussion about the selected

comment. Finally, they draw their conclusions. This cycle—question, comments,

vote, free discussion and conclusions—is repeated for each question presented on the

seminar. Then, mediators finalize the debate and evaluate learners’ participation.

To apply this dynamics, “Mediated Chat 2.0” tool was developed [10], implement-

ing some group conversation techniques: circular contribution, where the participants

are organized in a queue and each participant sends a message when it is her time in

the line (used on “Commenting on the question”); single contribution, where each

participant must send a single message and the message can be sent at any time with-

out order (used on “Vote on a contribution”); and free contribution, where any par-

ticipant can send a message at any time (used on “Free discussion” and “Conclu-

sions”).

The Mediated Chat 2.0 tool was a first step in introducing workflow like coordina-

tion techniques within an AulaNet service. This experience was important not only to

show the potential of such technologies to prescribe user activities, but also to con-

firm that the new coordination paradigm generates unexpected situations that requires

flexibility of the coordination mechanisms.

Frequently, on traditional chat tools a participant breaks up text to be sent into sev-

eral messages—this resource generally is used to increase the dialogical potential and

reduce the chat response time. Therefore, in the beginning on the use of Mediated

Chat tool, some messages were sent with incomplete text, and the sender was not able

to send another message completing the text, on account of the fact that the single

contribution or circular contribution conversation techniques were in place. This

feature of the tool, although generating some problems at the beginning of the first

debate, was quickly understood and the participants developed strategies to conform

to the well-defined conversation protocols imposed by the tool.

Some problems also occurred while using the circular contribution conversation

technique. Some participants reported difficulty in identifying their position within

the queue, even though the tool makes this information available to the user—what

reinforces that the users must have clear awareness elements that enable them to

quickly know the current workflow state.

Another problem happened when the person in the top of the queue takes a long

time to write her message, causing the entire group to wait. In this situation, the me-

diator has to call her attention to complete the message and unblock the next contri-

bution. This situation shows the importance of the social protocol even when the

dynamics of the activity is well defined.

Finally, another observed fact is that the last participants to contribute tend to sub-

mit long messages, because they have more time to elaborate their texts. This results

in atypical elaborated messages in a chat conversation. This fact indicates that the use

of workflow techniques, although organizes the activities sequencing, may uncharac-

terize or make inadequate the realization of an activity. This reinforces that workflow

like coordination mechanisms requires a careful introduction in the real world.

It is important to note that some of the aforementioned problems happened because

the chat interaction is a loosely coupled collaborative activity. Other activities, like

the Lesson Plan, are more structured, been more suited to a workflow based ap-

proach.

5 Conclusion

The coordination approach presented in this paper is a first effort to “embed” part of

the social protocol that governs the interaction between mediators and learners within

AulaNet. This workflow-based approach basically offers a pre-defined task sequenc-

ing, which aims to help learners to process the knowledge available on the course.

Another potential use of workflow technology within AulaNet is as an abstraction to

represent the course dynamics. The synthesized representation of the activities flow

helps the coordinator to re-design the course. Moreover, it makes it possible to define

different sequences for different learners or groups of learners within the same

course.

Another coordination approach, which is also being implemented within AulaNet,

is an activities monitoring mechanism proposed by SCORM (Sharable Content Ob-

ject Reference Model) [1]. While the workflow approach aims to enforce the proper

way of fulfilling the courses activities, the monitoring approach, as the name says,

aims to check if some milestone was reached. The former is based on events, pro-

active, while the latter is passive, based on content navigation.

Finally, although the use of workflow technology somehow restricts the paths to

be followed, it is important to stress that there is still room for the unstructured coor-

dination based on the social protocol. As previously stated, both unstructured aware-

ness based coordination and rigid workflow like coordination protocols have their

importance in an e-learning environment. The challenge is to offer coordination

mechanisms that enable a guided knowledge acquisition without reducing the learn-

ers’ possibilities within the e-learning environment.

Acknowledgements

AulaNet project is partially financed by Fundação Padre Leonel Franca and by the

Ministry of Science and Technology through its Program of Excellence Nuclei

(PRONEX) grant nº 76.97.1029.00 (3366), and also through its Multi-Agent Systems

for Software Engineering Project (ESSMA) grant nº 552068/2002-0. It is also fi-

nanced by individual grants awarded by CNPq to: C. Lucena nº 300031/92-0, H. Fuks

nº 303055/02-2, A. Raposo nº 305015/02-8 and M. Gerosa nº 140103/02-3. M. Pi-

mentel received a grant from CAPES.

References

1. ADL – Advanced Distributed Learning. Sharable Content Object Reference Model

(SCORM). http://www.adlnet.org

2. Borghoff, U.M., and Schlichter, J.H. Computer-Supported Cooperative Work: Introduction

to Distributed Applications. Springer, USA, 2000.

3. Brinck, T., and McDaniel, S. E. Awareness in Collaborative Systems, Workshop Report,

SIGCHI Bulletin. 1997.

4. Cuthbert, A. J. Designs for Collaborative Learning Environments: Can Specialization En-

courage Knowledge Integration? CSCL’99 – Computer Supported Cooperative Learning,

1999, 117-126.

5. Deiters, W., Goesmann, T., and Löffeler, T. Flexibility in workflow management dimensions

and solutions. International Journal of Computer Systems Science & Engineering 15(5),

2000, 303-313.

6. Fuks, H., Gerosa, M.A. and Lucena, C.J.P. Using a Groupware Technology to Implement

Cooperative Learning via the Internet - A case study. Electronic Proceedings of HICSS'35 –

35th Annual HAWAII Int. Conf. on System Sciences on the Big Island of Hawaii. 2002.

7. Fuks, H., Gerosa, M.A. and Lucena, C.J.P. The Development and Application of Distance

Learning on the Internet. Open Learning - The Journal of Open and Distance Learning,

17(1), 2002, 23-38.

8. Klingemann, J. Goal-Based Execution of Flexible Workflows. GMD Res. Series 12. 2001.

9. Raposo, A. B., and Fuks, H. Defining Task Interdependencies and Coordination Mechanisms

for Collaborative Systems. In Blay- Fornarino, M. et al. (Eds.), Cooperative Systems De-

sign. Frontiers in Artificial Intelligence and Applications, 74, 88-103. IOS Press, 2002.

10. Rezende, J. L. Aplicando Técnicas de Comunicação para a Facilitação de Debates no

Ambiente AulaNet. M. Sc. Dissertation. Computer Science Department of the Catholic

University of Rio de Janeiro, Brazil. March 2003.

11. Schmidt, K., and Simone, C. Coordination mechanisms: Towards a conceptual foundation

of CSCW systems design. Computer Supported Cooperative Work, 5(2-3), 1996, 155-200.

12. Schmidt, K., and Simone, C. Mind the gap! Towards a unified view of CSCW. COOP 2000

- 4th International Conference on the Design of Cooperative Systems, 2000, 205-221.

13. Workflow Management Coalition. Workflow standard – Terminology & glossary.

Technical Report WFMC-TC-1011, Version 2.0. 1996.