COOPERATIVE COURSEWARE AUTHORING SUPPORT Darina Dicheva*, Lora Aroyo** and Alexandra Cristea** *Winston-Salem State University 601 Martin Luther King Jr. Dr, Winston-Salem, N.C. 27110 United States of America
[email protected] **Technische Universiteit Eindhoven P.O. Box 513, 5600 MB Eindhoven The Netherlands {l.m.arroyo, a.i.cristea}@TUE.nl
Abstract We refined our knowledge classification and indexing approach applied in our previously developed system AIMS (Agentbased Information Management System) by introducing ontology-oriented support for cooperative courseware authoring. In order to provide a basis for formal semantics and reasoning in performing generic authoring tasks, we add ontology-based layers in the courseware authoring architecture. Ontological structuring also opens the way for cooperative authoring, (e.g., re-usage), setting the basis for authoring collaboration.
Key Words Ontology, Web-based courseware authoring, cooperative authoring, collaboration, ODL, adaptability
1. Introduction
The World Wide Web has significant impact on many domains. For education, this means information available from all around the world and courses with global dimensions. A Web-based course doesn’t only provide access to local documents, but also has the function of a gateway, allowing access to online educational materials worldwide. These dramatic changes have direct impact on the courseware authoring process. Authoring Web courses means authoring environments implicitly or explicitly open for sharing. On one hand, this requires efficient organization of Web resources, including powerful indexing and knowledge classification, which will allow their efficient use, re-use and sharing. On the other hand, such authoring involves cooperation or even collaboration of course authors and thus requires facilitation or implementation of appropriate supporting tools.
Concerning the efficient organization of learning materials, a promising tendency in present Web-Based Educational Systems (WBES) are concept-oriented (ontology-aware) architectures. Conceptual (ontological) structures, such as concept maps, topic maps and conceptual graphs have a great deal of potential for organizing, processing and visualizing subject domain knowledge in WBES. Concept-based courseware employs conceptual domain presentation to link educational materials to a course structure. This allows for concept-based course sequencing that supports adaptive courseware generation [1,2], concept-based information retrieval, visualization and navigation support that help students to get oriented within a subject domain [3,4], etc. An example of the latter is AIMS [4], an intelligent system aimed at supporting students in retrieving, evaluating and comprehending information when performing learning tasks in a Web-based learning environment. The solution we found to efficiently organize and maintain Web-based resources is based on conceptualisation of the course material. The basis of this conceptualisation is the creation of a subject domain ontology. This solution conforms to the guidelines set by the recent advances in the Semantic Web area (e.g., the layered architecture [5]) and in the ontologies area that provide the ground for applying semantic structures within the educational field. An ontology is a hierarchically structured set of terms for describing a domain that can be used as a skeletal foundation for a knowledge base [6]. In our case ontologies are used as the backbone of domain and course structure definitions and for implementing efficient information search [7,8]. An ontology, however, is not just a collection of terms. Concerning the computational semantics of an ontology Mizoguchi proposes three level ontologies: (1) a structured collection of terms, such as topic hierarchies and tags used for metadata description; (2) formal definitions, relations and constraints added to a level 1 ontology; (3) executable ontology [9]. Most of the current efforts aim at building ontologies at level 1 and 2 of this classification. Linking educational materials to ontological structures, however, requires that the concept-based Web courseware authoring (in systems such as AIMS) differentiates domain-, course- and library authoring, which makes it more complicated and labour intensive than the ‘standard’ courseware authoring. Apparently, this calls for enhancing the authoring support for concept-based Web courseware. Looking from another perspective, Web-based education cannot limit itself to a fixed scheme of one-author with one-audience anymore. Concerning the latter, one of the hot issues discussed by the research community is that instruction has to be custom-designed [10,11,12], to take into consideration the cognitive (learning) styles, the psychological predilections, the motivational aspects, the cultural backgrounds of the targeted students, the environmental influences, etc. However, if we look at authoring under the above requirements, each course would have to be re-authored with, roughly speaking, a multiplicative factor combining all the considerations above. Such authoring is extremely difficult and time-
consuming. Therefore, it needs specialized, modern authoring support. Moreover, one author might never manage to cover all aspects mentioned above. Thus, such requirements naturally lead to the necessity of re-usage, sharing, cooperation and sometimes collaboration among authors [13]. Following the outlined needs of further support for concept-based courseware authoring, we envisage such support to include: automatic or semi-automatic performance of some authoring activities; intelligent author assistance in the form of hints, recommendations, etc.; and support of the activities of different instructors for collaborative building and/or cooperative reuse of domain and course ontologies. There are two key ideas in our approach:
In order to provide enhanced support for authoring concept-based Web courseware we may well use the system’s domain concept map – the ontology, which captures the semantics of the subject domain terminology used by the students when searching (asking questions) for relevant information necessary to perform course related tasks. The same ontology can be used by courseware authors to ask authoring-related questions or by the system to perform (semi-) automatically some authoring activities. Thus, we propose to introduce additional ontology-based layers to the courseware authoring architecture, which would allow for intelligent assistance to the courseware authors.
The second goal is related to supporting cooperative authoring. Collaborative authoring can occur in project-like settings, where the project delegates authoring sub-tasks to a group of authors. This kind of authoring needs synchronization, dialogue support, and coordination of the whole project. In contrast, cooperative authoring mainly involves synchronous re-usage of authoring products, such as course materials, libraries, ontologies, etc. In our current work we focus on the latter, more precisely, on supporting primitive interaction activities in cooperative authoring. In this paper we present an ontology-oriented support for cooperative courseware authoring, extendable for
collaborative authoring. This comes as an elaboration of our approach to knowledge classification and indexing in the previously developed system AIMS (Agent-based Information Management System) aimed at supporting students in retrieving, evaluating, and comprehending information when performing learning tasks in a Web-based learning/training environment. We focus on cooperative authoring, which allows re-usage and sharing of educational material and components, and sets the basis for authoring collaboration. In order to exemplify our method we define a set of generic tasks related to concept-based courseware authoring and present their ontological support by the newly added operational and assistant layers in the AIMS architecture.
2. Courseware Authoring in AIMS
AIMS is a concept-based course support system. The use of ontologies in AIMS was inspired by the need for shareable and reusable knowledge bases among different courses and authors (instructors). From an information point of view, the general information model of a concept-based courseware system such as AIMS consists of subject domain, library, course and user profile objects [4]. We envisage domain concepts as the main linking component for all the other objects. In other words, every document in the library is linked to one or many domain concepts and every course task is linked to several domain concepts. A courseware system can accommodate different subject domains and multiple courses within them. Its information base includes definitions of the subject domains, the courses and the library, (a collection of documents related to these courses and domains). The domain model defines the subject-domain ontology and is represented as a concept map (CM) of domain concepts with labelled links between them. The link types (labels) are based on the generic selection of link types defined in [14]. Two concepts could be related by several link types. Each concept is defined in terms of name, short description, synonym-form list, and classification level. The synonym-form list is a list of weighted synonyms and other word forms, such as plurals, abbreviations, and commonly used forms. The weights are used in determining the relevancy of the retrieved documents to a user query. There are four classification levels (‘category’, ‘sub-category’, ‘topic’, and ‘sub-topic’), which define a hierarchy within the non-hierarchical structure of the domain concept map. Domain concepts are linked to documents and each link is assigned a weight indicating document relevancy to the concept. The library model provides means for maintaining a collection of information items (course materials and domainrelated documents) and their links to different courses and domains. The representation of each document (data source) in the library includes its name, author, year of publication, location (URL or library index), short description and list of weighted keywords, as well as its presentation format and instructional format. The course model defines the structure of a course and includes course topics and course tasks. Course tasks are predefined in a course task library and correspond to course assignments that the student is required to perform. Each course task is represented in terms of goal concepts, which the student must know or learn to successfully complete the corresponding course assignment. It also includes some additional information, such as task description, ‘prerequisite’ tasks, and task status.
Course tasks are related to course topics; thus the topics are automatically associated with the whole collection of domain terms included in all related tasks. As a result, the process of authoring of such concept-based Web courseware should include domain-, course-, and library authoring. Consequently, the AIMS authoring environment includes three main modules: Domain editor, Library editor, and Course editor [4]. The Domain Editor enables the authors to construct domain concept mapping structures by providing them with facilities to add, delete and update domain terms and links between them. The editor also allows the author to create new types of links and to create links between a domain term and existing documents in the AIMS library. The Course Editor enables the author to define a structure of a course within a specific domain by using the domain ontology as framework for structure definition. One course can consist of several topics and each topic can have several tasks. The Library Editor, as most of the library systems, enables the maintenance of information collections. It provides access to all the information related to different courses and domains. The novel feature here is the support of task- and useoriented description of documents. By supporting the authoring activities further we aim at increasing the efficiency with respect to information reuse and cooperation between the course authors.
3. Ontology-based Support to Courseware Authoring
We propose a 2D-layer approach (Figure 1) towards concept-based courseware authoring support. This approach allows reusage and sharing and sets the basis for authoring cooperation and collaboration (see Section 5). The Y-axis represents the main information objects in the courseware system, i.e. the vertical- or y-layers define library objects, domains and courses. The X-axis targets system’s support for information object authoring and basically represents the layered architecture implementing system functionality. The proposed horizontal- or x-layers include GUI, Assisting layer, Operation layer and Information layer. The GUI x-layer supports user-system communication. The Information x-layer contains the y-layered description and structuring of the information objects in the courseware system (educational metadata, subject domain ontology, course ontology). The Educational metadata contains the description of the data sources (teaching materials) in terms of metadata. The two new components (x-layers) in the extended AIMS
architecture are the Assisting x-layer and the Operation x-layer. The Operation x-layer handles operations for data manipulation in each information base y-layer. This way it enables the process of modelling data into an ontology and also allows the creation of alternative goal-oriented structures of courses. In this x-layer all functions related to information manipulation, consistency and co-operation are also handled. As a processing layer, it consists of three modules, corresponding to the three y-layers: course engine, domain engine and library engine. All of them include two sets of support operations: consistency check and co-operation support. The consistency modules perform their activities over each of the information base y-layers. They provide functions to facilitate the process of authoring the domain ontology, course ontology and educational metadata in a semi-automatic way. These modules also guarantee the consistency of the educational sources. They deal with tasks such as: handling notions of semantic equivalence and conflict, handling conflict resolution rules, handling equivalence comparison rules, enhancing the resulting ontology and defining additional constraints if necessary. The cooperation support modules offer on one hand a set of operations to check the consistency in alternative (simultaneous) course structure building by different authors and on another – predefined functions (patterns and templates) to facilitate effective reusability of the available course structures developed by different authors. In relation to the reusability support, we pay special attention to the issues associated with merging ontologies, such as: extracting portions of an ontology to be merged with another, identifying which frames are to be extracted from the source ontology, determining if the extracted information has semantic overlaps or conflicts with the target ontology, assisting in merging ontologies, recording the sources of inserted sub-ontologies for later reference and update, selecting patterns, templates in an educational ontology, to present them as predefined objects for other authors. Among the issues of importance when merging two ontologies are these related to the following types of semantic overlaps and conflicts: semantically equivalent concepts but with different names, semantically different concepts but with the same name, semantically equivalent concepts with the same name but different definitions, semantically equivalent concepts linked to different (sometimes conflicting) concepts, etc. While the Operation x-layer actually implements the authoring operations, the Assisting x-layer, which is based on the ontological mapping of the domain, is responsible for helping the author in the process of courseware authoring. For
example, it gives hints to the author of how to create a course structure, how to link a document to the ontology, or how to link a course item to the ontology, etc. According to the computational semantics of an ontology [9], the ontologies we consider here can be situated at level 1 (term collection) and our proposed new model aims at level 3 (executable task ontologies). We still have to address the connection given by level 2 (formal definitions, constrains and axioms). In the following section we discuss the support for some generic tasks in concept-based courseware authoring.
Figure 1. 2D-layer approach towards courseware authoring support.
4. Authoring Tasks
In this section we discuss generic authoring tasks supported by the operation sets of the Operation x-layer and the presentation options for the author provided by the Assisting x-layer, in the sense of the executable task ontologies mentioned above. We are defining a complete set of generic authoring tasks corresponding to all three information y-layers (subject domain, course, and library), which are supported by the domain engine, course engine and library engine respectively. These tasks are defined on the base of a set of atomic operations, which are abbreviated and summarised in Table 1. Due to lack of space, we cannot present in this paper the full ontology of courseware authoring tasks we developed so far. We only show a sample of domain authoring tasks with their operational semantics (realised by the domain engine) and author support (realised by the domain assistant) (see Table 2). The interaction between the domain engine and domain assistant is illustrated in the activity diagram in Fig. 2, which presents system support for the atomic authoring task ‘add concept’ (first task in Table
1).
We
also
briefly
discuss
the
course
and
educational
metadata
(library)
authoring.
Table 1. Atomic operation definitions (adapted from [2]). Atomic operation ‘Add’ ‘Del’, ‘Edit’ ‘U’ ‘L’ ‘V’ ‘Chk’
Range
Description
{To, Ta, Co, Li, Doc}, where To ∈ {course topics}, Ta ∈ {course tasks}, Co ∈ {domain concepts}, Li ∈ {domain links}, Doc ∈ {library documents}. as above
Adds an object to the corresponding structure, i.e. To, Ta, Co, etc.
as above {CM, CS, EML}, where CM=Concept Map, CS=Course structure, EML=Educational Metadata Library. {DirLC, RelC, RelCo, RelTa, RelDoc}, where DirLCo = Directly linked concepts, RelC = Related courses, RelCo = Related concepts, RelTa = Related tasks, RelDoc = Related documents. {Graph, Text}, where ‘Graph’ is a graphical and ‘Text’ a textual view. {Ta, To, Co, Li, Doc, RelCo , RelTa , RelDoc , DirLC}
Deletes an object from the corresponding structure. Edits an object settings. Updates the corresponding information structure, i.e. CM, CS or EML. Lists the objects of the set(s). Gives alternative views of the engine results to the author. Checks the existence of objects within the set(s).
4.1 Domain
The interaction stream between the domain assistant and domain engine is triggered by such common authoring tasks as createNewDomain, editExistingDomain, copyExistingDomain to a new one and mergeDomains. Since the domain ontology is represented as a concept map, its authoring is in terms of manipulating concept maps [15], for example, creating and modifying CMs, comparing CMs, merging CMs, extracting subsets of CM, etc. Thus, domain authoring tasks involve a set of basic concept-maintenance related tasks such as: Add(Ci,CM), Del(Ci,CM), Edit(Ci,CM) to add/delete/edit concept Ci in a concept map CM; Add(Li,C1,C2,CM), Del(Li,C1,C2,CM), Edit(Li,C1,C2,CM) to create/delete/edit a link of type Li between concepts C1 and C2 in concept map CM; Add(Li,T), Del(Li,T), Edit(Li,T) to create/delete/edit link Li of a type T. At a coarser granularity level, authoring tasks include:Del([DirLinks], Ci, CM) to remove all direct links to a concept Ci in a concept map CM; Del([Path], C1, C2) to remove all segments of a path between two concepts; Create(CM, Domain), Edit(CM, Domain) to edit/create a concept map (the ontological structure of the domain); Link(Ci, CM, Domain, Doc) to make links between the domain concepts and the library documents. These authoring tasks trigger a set of operations performed by the domain engine over the domain ontological structure. These operations ensure data consistency by including domain specific checks for conflicts. For instance, for the authoring task Add (Ci, CM) the domain engine performs Chk (Ci, CM, exist) to check whether the concept Ci is already in the map, then updates the CM with U (Ci, CM) and performs Add (Ci, weight). Finally it provides information on the task results
to the domain assistant for analysis and presentation to the author. In this case the information depends on whether the concept was found in the CM and can be included: (a) L (Ci, synonyms), (b) L (DirLCi) and (c) notification that the new concept Co has been added to the CM. This information is used by the domain assistant to present the task results to the author in an appropriate format and view(s) in order to customize and support his/her authoring most efficiently. In this case the domain assistant allows the author to choose from the alternative operations V (Text, DirLCi), V (Graph, DirLCi) as well as from the alternative views for the synonyms V (Text, Ci, synonyms), V (Graph, Ci, synonyms). There are also a number of composite actions such as: DeleteLinks applied to all direct links between concepts in the domain CM, or DeletePath applied to all the concepts on a direct path between C1 and C2 of the domain CM. These actions can be implemented with a repetitive call to atomic operations, in this case Del(Ci,CM) and Del(Lj,CM). Table 2. Domain authoring tasks. Task Add (C, CM)
Add (L, C1, C2, CM)
Del (L, C1, C2, CM)
Del (C, CM)
Domain Assistant suggest options for the author: - add or delete domain engine results give alternative presentation: - V (Text, RelC, Relevance %) - V (Graph, Domain Ontology, Matched concepts) ‘you are here’ notify the other authors of adding the C to CM suggest options for the author: - add or delete domain engine results notify if one of the concepts is missing from the CM notify if the link exists and offer editing option notify about an indirect linking path between the C1 and C2 and give altern. Present. Of DE results: − V (Text, RelC, Relevance %) − V (Graph, Domain Ontology, Highlighted paths) notify the other authors of adding the L to CM give alternative presentation of all NonDirLinks: − V (Text, L, Relevance %) − V (Graph, Domain Ontology, Highlighted paths) • Give user the option to either graphically or textually cut all segments or sub-sets of them • notify the other authors of removing the L from CM • Notify the author if C has DirLinks • Give alternative views with the option to delete them or cancel the delete action: − V (Text, DirLinks, Relevance %) − V (Graph, Domain Ontology, Highlighted paths) • notify the other authors of removing the C from CM
Domain Engine Chk (C, CM, ∃ ) = true perform keyword search on C within DO (domain ontology) U (CM, C)
Result L (synonyms, C, CM) L (DirRelC, C, CM) L (NonDirLinks, C, CM)
Chk (C1, CM, ∃) = true, Chk (C2 CM, ∃) = true Chk (L, CM, ∃) = true Chk (C1-x-C2, CM, ∃) = true (indirectly linked with path ‘x’) U (CM, L)
L (x, C1-C2, CM)
Chk (C1-x-C2, ∃ for L in CM) = true Del (L) and U (CM)
L (x, C1-C2, CM)
Chk (L, ∃ for C in CM) = true Del (C, CM) and U (CM)
L (DirRelC, C, CM) L (DirLinks, C, CM)
Figure 2. UML activity diagram for the atomic domain authoring task ‘add concept’: Add (Ci,CM).
4.2 Course
The interaction process between the course assistant and course engine is triggered by a set of common authoring tasks, such as Create(CSi,Dj) to create new course structure CSi in domain Dj; Edit(CSi,Dj) to edit existing course structure CSi in domain Dj; Del(CSi,Dj) to delete existing course structure CSi from domain Dj; Copy([CSk,Di],[CSm,Dj]) to copy existing course structure CSk in domain Di to CSm in domain Dj, where i could be equal to j. Each of them involves a set of basic coursemaintenance related tasks, such as Add(To,CSi), Del(To,CSi), Edit(To,CSi) to add/delete/edit topic To of course structure CSi; Add(To,Ta,CSi), Del(To,Ta,CSi), Edit(To,Ta,CSi) to add/delete/edit task Ta in topic To of course structure CSi; Add(Cj,To,CSi),
Del(Cj,To,CSi), Edit(Cj,To,CSi) to add/delete/edit concept Cj in topic To of course structure CSi; Link(Doc,To,CSi), Link(Doc,Ta,CSi), Del(Doc,To,CSi), Del(Doc,Ta,CSi) to link/delete document Doc to topic To or task Ta in course structure CSi. The above tasks trigger a set of operations performed by the course engine over the existing course structures. The operations ensure data consistency by performing domain specific checks for conflicts. For instance, when the authoring task Add (To, CSi) is performed the course engine performs keyword search (in both the domain and course ontologies) on the entered topic expression. The course assistant provides the option of manual editing of the results. Then, it presents the author with alternative views on the course engine results: (1) textual list of results ranked according to their relevance to the search query, (2) graphical representation of the course tree with the matched concepts highlighted and (3) graphical representation of the domain ontology with the matched concepts indicated (‘you are here’ indication). The course engine also ensures storage of the results for further reuse. Other possible course authoring tasks related to document library and education metadata include Link(Doc, To, CSi) and Link(Doc, Ta, CSi) to link document Doc to topic To or task Ta of course structure CSi; Del(Doc, To, CSi) and Del(Doc, To, CSi) to delete document Doc from topic To or task Ta of course structure CSi. There are a number of composite actions such as DeleteAll applicable for course topics and tasks in a course structure or for concepts -in course topics, which are implemented with a repetitive call to atomic operations, such as Del(To,CSi) and L([To],CSi).
4.3 Library
A set of common authoring tasks triggers the interaction stream between the library assistant and library engine. The authoring of library involves maintaining a list of educational items and their metadata. This includes maintaining the links between the library documents and the terms of the subject domain, on one hand, and the topics and tasks of the course structure, on the other. Each document in the library also needs authoring with respect to keywords, their weight, instructional and presentation format and general library data, such as author, title, etc. Thus, the library authoring involve a set of basic tasks such as Create(Libi), Edit(Libi), Del(Libi) to create/edit/delete library Libi; Add(Doci,Libj), Edit(Doci,Libj), Del(Doci,Libj) to add/delete/edit document Doci in library Libi; Link(Doci,Libj,Ck,Dm), UnLink(Doci,Libj,Ck,Dm) to link/unlink document Doci in library Libj to concept Ck of domain Dm; Add(KeyWi,Docj,Libk), Del(KeyWi,Docj,Libk) to add/delete keyword
KeyWi
of
document
Docj
in
library
Libk;
Link(Doci,Libj,To,CSm),
UnLink(Doci,Libj,To,CSm),
Link(Doci,Libj,Ta,To,CSm), UnLink(Doci,Libj,Ta,To,CSm) to link/unlink document Doci in library Libj to topic To or task Ta of course structure CSm. The library authoring tasks trigger a set of operations performed by the library engine over the library educational items collection. These operations ensure data consistency by including checks for conflicts. For instance, for the authoring task Add(Doci,Libi) the library engine performs Chk(Doci,Libi,exist) to check whether document Doci is already in library Libi, then updates the document list in the library with U(Doci,Libi) and performs Add(Doci, Libi, metadata). In the metadata the author includes the keywords, their weights and all requested fields to map the new document to the library structure. The library engine transmits further the results to the library assistant for analysis and presentation in an appropriate format and view(s) to the author in order to customize and support his/her authoring most efficiently. In this case the library assistant allows the author to choose from the options: (1) select a document from the list of documents in the library with the same starting letter; (2) select a document from the list of documents with the same authors; (3) select a document from the list of documents with the same title; (4) add the new document to the library; and (5) delete an existing document from the library. After the initial selection the author is offered to change/edit the metadata of the selected or newly entered document. Vvarious hints and suggestions are presented further to the author in order to keep the consistency of the library metadata, documents list and the related domains and course structures. Notifications are also sent to the other authors with respect to the performed changes.
5. Cooperative Author Support
In this section we map the newly proposed ontology-based cooperative authoring environment over a more general picture of cooperation and collaboration [13], to show in which degree our proposal matches the requirements. In pure collaborative authoring, each author takes over one or more authoring sub-task(s). When each author accomplishes the sub-task(s), the group goal is reached and collaborative mutual interdependent authoring is achieved. In the wider spread cooperative authoring, authors just re-use each other’s materials, learning goal settings, dictionaries, course and tasks linking and sequencing, ontologies, etc. The primitive interaction activities among participants during both cooperative and collaborative authoring, from a macro-granulation perspective, are as follows (listed in their order of priorities): (Task) Planning, (Task) Execution, Creation. Coordination, Control (mainly for collaboration).
Initiative, Supervision (mainly for collaboration) Observation, Suggesting (mainly for collaboration). Data Sharing. Idea Sharing (mainly for collaboration). Dialogue (with Interaction) (only for collaboration). Surprisingly, compared to collaborative learning [13], only ‘Turn Taking’ doesn’t appear as a primitive activity and, supplementary to pure (task) ‘Execution’, ‘Creation’, as a higher-level ‘Execution’ form, is added, to make the distinction from cooperative learning environments, where no new knowledge (here, in the form of concepts, tasks or links) is produced. In the present model, we only create support for the following cooperation activities: ‘(Task) Execution’ and ‘Creation’ – broken down in the atomic operations (‘Add’, ‘Del’, ‘Edit’, ‘U’, ‘L’, ‘V’) and the respective authoring tasks – as exemplified by Table 2 (Section 4); ‘Control’ – implemented in the form of consistency checks of information structures developed by single or different authors; ‘Data Sharing’ – represented by the Information Layer. From a technological point of view, support systems for cooperative authoring should provide all appropriate tools for these activities. Thus, the following resources are required (see figure 3 left side): Individual (personal) work space: this is the space where individual authoring interaction activities, such as ‘Planning’ (partially), ‘Execution’, ‘Creation’ and ‘Initiative’ (partially), take place. Cooperative (shared) work space (shared object space) – possibly extendable to collaborative work space: this is the space where activities, such as ‘Planning’ (partially), ‘Coordination’, ‘Control’, ‘Initiative’ (partially), ‘Supervision’, ‘Idea Sharing’ (partially) and especially ‘Data Sharing’, take place. For collaborative authoring, additionally to the above, the following must be present: Dialogue channel (implicit or explicit) – a space where ‘Initiative’ (partially), ‘Supervision’ (partially), ‘Observation’, ‘Suggesting’, ‘Data Sharing’ (very little), ‘Idea Sharing’ and, obviously, ‘Dialogue’, take place. Technologically mediated remote communication (audio, visual) – optional – with the role of enhancing the ‘Dialogue’ space and extending the same activities. This cooperative and the extended collaborative model can be further elaborated. However, due to lack of space, the other layers and aspects thereof (component objects, their functionality and interactions) are not further detailed in this paper.
Figure 3. Mapping of the Cooperative/Collaborative Authoring on 2D-layers. Figure 3 presents the global model of cooperative and collaborative authoring (left side of figure), as well as the mapping of the presented 2D model over it (right side of figure). In this paper we focused on the Performance Layer and Information Reference Layer, with their respective ontological representations. The activities presented in Section 4 refer to activities at the level of the individual workspace, but their results are visible from the cooperative workspace. More specifically, the Information Layer (right side) corresponds to the resources for authoring in the Information Reference Layer (left side). Respectively, the Operation Layer is especially aimed at the individual work space, while the Author Assisting Layer is bridging the individual and cooperative work spaces. The extension towards the collaborative work space is not done yet, although the presented (Section 3) Operation and Author Assisting Layers set the basis for such a development. For the inclusion of the other primitive interaction activities, a new layer or module, such as the Extended Author Assisting Layer sketched in Fig. 3, could be added to the present model.
6. Conclusion
In this paper we have introduced a layered approach for Web-based courseware authoring, which was transformed into a cooperative ontology engineering process. We have started by integrating our previous research into the main research steps for courseware authoring, and by stressing the need for cooperation and even collaboration among authors, especially in Web-authoring, where the information is plentiful. We based our analysis and discussions of cooperative course-authoring tools on the AIMS system. Cooperation implies dividing the course contents into re-usable, sharable parts. In AIMS, these parts are concepts, tasks and topics. We have shown the authoring tools present in AIMS and discussed that although the
structure allows cooperation and is adequate for this purpose, the actual authoring support needs improving. The paper proposes an additional semantic layer for intelligent authoring assistance and details the various types of support that such a layer should provide for various types of authoring actions possible in an ontology-based courseware-authoring environment (such as AIMS, but not restricted to it). We have shown how the concept-based ontological authoring support we envision maps over a cooperative (and extended collaborative) authoring model and also pointed to future possible extensions. We focused on the issues of re-usage and cooperative information sharing, as well as pointed to possible steps towards collaborative authoring, in the sense of simultaneous performance of authoring activities. The processing presented in this paper is self-contained. However, many more aspects can be analysed and further research direction pursued. A direction already pointed to in Section 5 is towards pure collaborative authoring environments. This would mean a merge between re-usage based cooperative environments, such as the one presented in this paper, and collaborative means of working extracted from previous research on collaborative learning environments [13].
References
1.
P. De Bra, P. Brusilovsky, T. Murray, & M. Specht, Adaptive Web-based Textbooks. (Panel introduction) Proc. WebNet Conf., 2001, 269-271.
2.
J. Kay, & Kummerfeld, Adaptive Hypertext for Individualised Instruction, Workshop on Adaptive Hypertext and Hypermedia, UM'94, 1994, Cape Cod.
3.
R. Hubscher, & S. Puntambekar, Navigation Support for Learners in Hypertext Systems: Is More Indeed Better? AIED, IOS Press, 68, 2001, 13-20.
4.
L. Aroyo, & D. Dicheva, AIMS: Learning and Teaching Support for WWW-based Education, IJCEELL, 11(1/2), 2001, 152-164.
5.
T. Berners-Lee, Semantic Web road http://www.w3.org/DesignIssues/Semantic.html.
6.
T. Gruber, Toward principles for the design of ontologies used for knowledge sharing. In Formal Ontology in Conceptual Analysis and Knowledge Representation, N. Guarino and R. Poli (Eds.), Kluwer Academic, 1993.
7.
L. Aroyo, D. Dicheva, & I. Velev, A Concept-Based Approach to Support Learning in a Web-based Course Environment. J. Moore, C.L. Redfield, W.L. Johnson (Eds.) AIED, IOS Press, 68, 2001, 1-10.
8.
L. Aroyo, D. Dicheva, & A. Cristea, Ontological Support for Web Courseware Authoring, Proc. Int. Conf. On Intelligent Tutoring Systems (ITS’02), 2002, Biarritz, France, 270-280.
9.
R. Mizoguchi, J. Bourdeau, Using Ontological Engineering to Overcome Common AI-ED Problems, Int. Journal of . AI in Education, 11, 2000, pp. 1-12.
map.
Internal
note,
World
Wide
Web
Consortium,
1998,
10. P. Brusilovsky, Adaptive Educational Hypermedia, Proceedings of PEG conference, Finland, 2001, 8-12. 11. Cristea, & T. Okamoto, OO Collaborative Course Authoring Environment supported by Concept Mapping in MyEnglishTeacher, ET, 4 (2), 2001. 12. P. De Bra, & L. Calvi, AHA! An open Adaptive Hypermedia Architecture, The New Review of Hypermedia and Multimedia, 4, 1998, 115-139.
13. T. Okamoto, M. Kayama, & A. Cristea, Consideration of building a Common Platform of Collaborative Learning Environment, ICCE’01, Korea, AACE, Ed.: C.-H. Lee, 2, 2001, 800-807. 14. Lambiotte, J.G., Dansereau, D.F., Cross, D.R. and Reynolds, S.B., Multirelational semantic maps. Educational Psychology Review 1(4), 1989, 331-367 15. Sowa, J., Building, Sharing, and Merging Ontologies, http://www.jfsowa.com/ontology/ontoshar.htm
Darina Dicheva is Associate Professor of Computer Science at the Winston-Salem State University. She joined WSSU after working for a long time at the Faculty of Mathematics and Informatics, University of Sofia. Her research interests include applying AI methods and tools in education, intelligent tutoring systems and shells, adaptive information retrieval and filtering, Web-based flexible and distance learning. She has authored and coauthored over 70 research papers and textbooks. She is Associate Editor of the International Journal of Continuing Engineering Education and Life-long Learning (IJCEELL), Associate Editor of the IEEE Educational Technologies and Society Journal (East-European section), and Member of the Review Board of the Interactive Multimedia Electronic Journal of Computer-Enhanced Learning (IMEJ). She has guest edited a number of special issues of BJET, JCAL, and IJCEELL and has been a member of the program committees of more than 15 international conferences in the recent years.
Dr. Lora Aroyo is a full-time assistant professor at the Information Systems Group, Fac. of Math. and Comp. Science, Eindhoven University of Technology, The Netherlands, and a part-time at the Appl. Protocol Systems Group, Comp. Science Dept., University of Twente, The Netherlands. Her research interests are in the field of Adaptive Information Retrieval, User Modelling, Concept Mapping, Ontologies and the application of AI methods and tools in education for ITS and Web-based Educational Environments. She authored and co-authored over 20 research papers. She has guest edited special issues of IJCEELL and JILR. She is a member of the Int. Society of AI in Education and of ICAI’02 program committee. She is a reviewer for ET&S and an IJCEELL Journal Associate.
Alexandra I. Cristea received her IS Dr. title and worked at the University of Electro-Communications, Tokyo, Japan. She is presently Assistant Professor at the IS Group, Faculty of Mathematics & Comp. Science, Eindhoven, University of Technology, The Netherlands. Her research interests include AI, Neural Networks, Adaptive Systems, UM, Knowledge Computing, Concept Mapping, ITS, Web-based Educational Environments. She authored and co-authored over 80 research papers and course booklets. She is a member of IEEE, was program committee member of ICCE’01, ICAI’02, IKE’02 and was reviewer or session chair for many conferences, such as ICCE’99, ITS’00, CEC’00, AH’02. She was reviewer for the Know BIS book and IEEE TNN, IEEE TKDE, KI journals. She is executive peer reviewer of the ET&S Journal.
