A personalized learning path generator based on metadata standards.

Thanks to the technological improvements of recent years, distance education represents a real alternative or support to the traditional formative processes. The Internet allows the design of contents, which are able to raise the quality of the traditional formative process. However, the amount of information students can obtain from the Internet is immense and students can easily be confused. Teachers can also be disconcerted by this quantity of content and they are often unable to suggest the correct content to their students. A solution to these problems can be derived from the ever more detailed description of each content area: in literature this process is defined as creating metadata. This approach, in fact, can support the introduction in an e-learning environment of a new software module: the Intelligent Tutoring System. These modules can easily build personalized learning paths. In fact, the real problem is often that of organizing lessons genuinely based on student profiles and not only a simple sequencing of contents. This article proposes a Java and JSP technology-based tool for metadata creation and management and the automatic selection of contents to form a sequencing of lessons and to complete a learning path. With this tool teachers can describe the contents, student profiles and ontology, according to a standard model that at this moment is "standard de facto." In this tool we have integrated a module that from the standard description of various resources (student profiles, content descriptions, etc.) deduces their digest representative vector. By comparing these vectors this module automatically finds the most suitable set of contents for every student profile.

**********

Some of the tasks that any e-learning platform should carry out and that characterize the whole training process are to allow people involved in training to find, evaluate, and acquire contents to allow easy updating and maintenance of both contents and information about users and to create lessons tailored to each student creating the so-called learning object sequencing (Cirillo, Cozzolino, De Santo, Marsella, Salerno, 2000). These activities are common and easy to carry out in traditional learning processes; however, we cannot say the same when this new technology is used. While designing and organizing a course, a teacher has to choose the most appropriate training contents. This digital content selection presents notable difficulties, also due to the huge amount of information available on the Internet, of which only a minimum part really meets teachers' needs. This problem is still more evident when students must autonomously find the training contents that best suit them. The possibility of accessing to contents that could be useless or not related to the subjects of interest is considerable. A solution to these problems is derived from the ever more detailed description of each training content area, a process known in literature as the process of creating metadata (Gyo Sik Moon, 2001). Metadata is descriptive information; according to the classical definition, it is data about data (Kent & Schuerhoff, 1997). At this present moment, metadata has a wide application: in fact, this concept is used in various applications, for example, in multimedia retrieval, advertising, even if the main aim of metadata remains the same: providing a detailed description of a resource (Bretherton & Singley, 1994). The e-learning industry is concerned with establishing rules to be commonly used in the process of creating metadata and, consequently, in describing contents, users, ontologies, and course structure. The use of standardized metadata allows current e-learning platforms to integrate new and more powerful services. In fact, in addition to reusability and sharing of training resources with other platforms, it is possible to design and implement "intelligent" services able to help students and teachers during the training process (Gyo Sik Moon, 2001; Murphy, 1998; Hernandez-Dominguez & da Silva, 2001; Colace, De Santo, Molinara, & Percannella, 2003). These services can add value to a platform and guarantee improvement in the pedagogical quality of the training process (Jonassen, Peck, Wilson, & Pfeiffer, 1998; Alessi & Trollip, 1996; Colace, De Santo, & Vento, 2003). This article is aimed at designing and implementing a set of web-based services that, by resorting to standards, are able to help e-learning system actors during the various phases of the training process. This software tool (Java based) can be integrated into any given e-learning platform improving it, and can also offer services for creating metadata associated with information, semi-automatically creating course structure, and for content research and evaluation to offer students. For the description of the resources we have chosen the technology that at this moment is considered "standard de facto" in the e-learning field. Using the structure and the features of these standards, we describe a resource as a vector to facilitate the choice of the most suitable contents for each student. The article also reports some results obtained using the proposed tool.

METADATA STANDARDS

The main aim of metadata process is the description of data. Generally, metadata is organised into categories or fields that represent a characteristic of the learning resource and adds information to the data. This process cannot be arbitrarily carried out, otherwise one runs the risk of creating contradictions and incompatibility. The e-learning industry is concerned with establishing rules to be commonly used in the process of creating metadata and, consequently, in describing contents, users, ontologies, and course structure (IMS Global Learning Consortium, September, 2003). The definition of such standards is aimed at improving training resource access, to increase resource reusability allowing semantic interoperability among different knowledge domains, improve the documentation quality in the training field, and make content research more effective through using semantic criteria. At this moment real standards are still not available, but some are becoming standard "de facto." For example, the IMS Global Learning Consortium learning object (IMS Global Learning Consortium), user profile, question, and interoperability standards are used by many e-learning environments at the moment. IMS is a consortium including members from commercial, educational, and government organizations created to develop and promote specifications for facilitating online learning activities. For course description ADL SCORM (Sharable Content Object Reference Model) (ADL SCORM, 2003) is the most common standard, while SHOE (SHOE Project, 2003) is the simplest standard for ontology description. In our proposal we use IMS, SCORM and SHOE standards to describe data.

Services for Creating Metadata

These services are essentially teacher-oriented and allow metadata creation. Receiving information on resources from teachers, such services must create XML files starting from DTD files specified by standards. In an e-learning environment, the teacher has to create metadata from the following resources: User profile, Ontologies, Training contents and questions, Learning Objects and Courses. We have therefore designed a Java-based service that implements the use case in Figure 1.

Our modules provide teachers with a friendly environment, which enables them to acquire, through some GUIs, all the information necessary for correctly describing the resource. In the next paragraph we will describe the various modules except for the learning path generator module. We will describe this module in the fourth paragraph after the description of the modelling of the didactic resource.

[FIGURE 1 OMITTED]

User Profile Description Module

Information on students represents one of the most important factors for an e-learning platform to work well. Taking into account didactic objectives, learner's preferences and competences, it is possible to give a teacher or a software application the possibility of choosing among the contents most suited to learners. Besides, the possibility of updating the user profile at the end of each training phase, as well as of allowing the teacher to continuously check user's progress, greatly improves learning process quality. As previously stated a standard provided by the IMS consortium will be used for this kind of description: the Learner Information Package (LIP). In structuring the LIP standard, the consortium has defined 11 fundamental categories, in which information on students are kept (IMS Global Learning Consortium, 2003). Each of these fundamental elements is composed of a variable number of subfields. Some of these fundamental macro-fields play a more important role in our project (i.e., Goal and Activity). Though faithful to the standard, we will define the desired level of detail in our description. In Figure 2 the graphic user interface of the student description tool is depicted.

Ontology Standard Description Module

In general, an ontology is a complex structure made up of a series of elements, each of which is composed of a kind of Relation and a series of related Concepts (Chandrasekaran, Josephson, & Benjamins, 1999). Through an ontology built by the teacher, it will be possible to describe the knowledge domain, the subjects constituting it, the relations among the various subjects, as well as methodologies and means with which they are presented. To describe ontologies in our module, an XML based standard is used: Simple Html Ontology Extension (SHOE). The Parallel Understanding Group of the University of Maryland has proposed this standard, which is essentially an extension of the META tag mechanism of the HTML language. One of the most important aspects is that SHOE was designed with the needs of the Web in mind, taking into account a possible significant interaction between intelligent agents and web documents. It has also limited semantics to make it possible to handle large amounts of data. Therefore, the final objective of this module is to provide a service, which allows the teacher to describe the ontology and which permits its digital record according to the SHOE standard. On the other hand the ontology description can improve the performance of an automatic learning generator path, because in this way it can build a correct learning path for the students. Figure 3 shows the graphic user interface of the ontology description tool.

[FIGURE 2 OMITTED]

Learning Object Description Module

This service allows the teacher to describe the contents. Metadata associated with contents facilitates the achievement of the following objectives: to sum up data meaning and analyze, evaluate, and filter a series of documents. As for the standard, we suggest using the one provided by IMS. At present, it appears to be the most complete standard, which is also widely accepted by organizations. The structure proposed for describing training resources is hierarchical. At the top of the hierarchy, there is the "root" element containing various subelements that can have further subelements inside them. Starting from the root, there are nine subelements: general, lifecycle, metametadata, technical, educational, rights, relation, annotation, classification. Each of these contains a complex series of subelements. Obviously, some of them have a greater importance in defining the best pedagogical context where resources are to be included, as well as in better characterizing these resources for the student. Figure 2 shows the graphic user interface of learning object metadata description tool.

Test and Training Contents Description Module

Particular attention was paid to metadata associated with test and training contents. In any e-learning platform, there is a need for evaluating a student's learning grade. A solution to the presentation of assessment tests and response processes is that of describing the questions by using the standard Question & Test Interoperability (QTI) as basic tool for implementation. QTI specifications defined within the IMS standard are engaged in both offering a solution for the presentation of assessment tests to users and facilitating the automatic response process. Question description also allows associating incorrect responses with retrieval programs from which a student can directly benefit. Figure 5 depicts the graphic user interface of question metadata description tool.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

MODELLING OF THE DIDACTIC RESOURCE

The opportunity of better defining a resource by using its didactic and pedagogical characteristics through the description standard fields induces us to represent it with a model. The idea is to generate, a "digest" of learning objects. Our aim is to better qualify the resource, making it clear to the software module, which interacts with the contents, the knowledge domain to which it belongs and its more peculiar characteristics. At the same time, an opportune modelling allows quantifying the resource, making it possible to establish a relationship among metadata by using appropriate metrics. The objective quantification makes it possible for an intelligent software tool to propose the contents that are suitable to the student needs. We have therefore implemented a software module able to model the single described training resource through a string vector whose components summarize specific resource features. This representative vector has been so defined:

   Didactic Resource ={Typology, Ontology, Pedagogical educational
   properties, Technical requisites, Rights}

Each component of this vector is still a string vector, and represents a particular aspect of the resource and gathers the most important information obtained combining standard description fields. We have chosen to use a vector since this structure better organizes the information associated with the resource allowing its easier retrieval. It is clear that the possibility of presenting this vector representative of learning object semantic content to an intelligent software module, which is able to semi-automatically infer decisions concerning the training contents utilization, as an Intelligent Tutoring System is able to do, improves and optimizes its performances in the retrieval process. We are going to explain in detail the descriptive vector components:

{Typology}: The main aim of this category is to give a global and general vision of the resource. This vector contains all information useful for classifying the learning object. The vector {Typology} is so structured:

   {Typology}:= {Typology, Identifier, Title of the resource, Author of
   the resource, Date of creation of the resource, Language,
   Description, Keywords}

In Table 1 there is an explanation, according the standard specifications, of every component.

{Ontology}: this vector has to explain in which didactic context the resource can be inserted. This component has to provide information in order to contextualize the training resource in a well defined knowledge domain. Because of this vector, it is possible to associate each resource with other training resources that belong to established knowledge domains allowing the organization of training paths. This type of representation appears to be particularly suitable for locating and recovering the resource itself within the domain. The vector {Ontology} is so structured:

  {Ontology}: {Purpose, Taxonpath, Taxon, Description, Keyword,
  Relation, Kind, Resource}

In Table 2 there is an explanation, according the standard specifications, of every component.

{Pedagogical Educational Properties}: this vector describes the pedagogical and educational characteristics defining the resource. It is possible through the analysis of this component to know the interactivity level of the resource with the user, its semantic density, and in general to pedagogically define it. The vector {Pedagogical educational properties} is so structured:

  {Pedagogical Educational Properties}: {Pedagogical Educational
  Properties, Interactivity, Resource Type, Interactivity Level,
  Semantic Density, Resource Users, Teaching Context, Age Range,
  Difficulty, Learning Time, Description}

In Table 3 there is an explanation, according the standard specifications, of every component.

{Technical requisites}: this vector has to describe the technical requisites necessary to the correct utilization of the resource. In particular, it is engaged in defining what its technological format is, what operating system makes it work, and what software is necessary for its correct utilization. In addition, it makes it possible to find the actual location of the resource. The vector {Technical requisites} is so structured:

  {Technical requisites}:{Technical requisites, Format, Size, Location,
  Required software resources, Duration}

In Table 4 there is an explanation, according the standard specifications, of every component.

{Rights}: This vector describes the billing modes and the costs associated with the resource. The vector {Rights} is so structured:

  {Rights}:={Rights, cost, copyright, rating}

In Table 5 there is an explanation, according the standard specifications, of every component.

Every component matches with the respective descriptive IMS standard field. In this way we can work with a well defined set of standard information, which is also the most meaningful, and we can use all descriptive fields when more detailed information is needed.

LEARNING PATH GENERATOR MODULE

One of the aims pursued by our work is to create a software service to be integrated into a given e-learning platform to allow the automatic definition of a course structure and its standard description on the basis of the information acquired through metadata associated with user profiles, ontologies, and learning object resources. Course structure includes organization of various lessons and contents related to them, management of users and their progress, and evaluation management. It is possible to use a metadata standard for defining course structure. SCORM is one of the best standards destined to become a "de facto" standard. This software service which starts from the user, ontology, training resource, and test description must automatically design the best course structure for each single student providing its description in the SCORM format. Course structure design includes lesson scheduling and the choice of the most suitable contents to be associated with lessons. In addition, lesson scheduling includes students' evaluation phases. As already stated, this software module, that has an intelligent tutoring role in an e-learning platform, has to choose, organize, and propose the best training path to the student on the basis of the student profile. The choice of the best training path obviously involves the choice of the learning objects more suitable to the student preferences. The resource, using ontology standard description, can be chosen taking into account the pedagogical context in which the user attends the lesson. From the point of view of size, light resources (in byte) should be preferred in case of non high-quality Internet connections. Another aspect to be considered is related to the time that the user can dedicate to the lesson. The system must therefore offer resources whose learning time estimated by the teacher should not overcome the time that the user would like to spend attending the lesson. Our module has to acquire the following information from the standard descriptive fields of the user profile: Interactivity level preferred by the student in the resource, student learning level, time dedicated by the student to the lesson, connection type usually used by the user, and preferred user language. In this way we can create a vector similar to the learning object digest vector that has been introduced in the previous paragraph. The information contained in the fields of this vector will be interpreted, manipulated, and kept in a special structure, in this case a numerical vector, that represents, from a numerical point of view, the resource and the user profile. The structures are so defined:

   User = {Difficulty[.sub.u]; Interactivity[.sub.u]; Size[.sub.u];
   Time[.sub.u]}
   Training resource = {Difficulty[.sub.r]; Interactivity[.sub.r];
   Size[.sub.r]; Time[.sub.r]}

The Difficulty field in the vector User is closely related to the results obtained by the user during the courses and contains a numerical value representing the arithmetical mean of the results obtained until the present time. Our software module therefore uses this numerical information, from 0 to 10, to create a range of value. The Difficulty field in the vectors' Training resource is obtained from the field of Pedagogical educational properties of vectors that describe the learning object. In this case our module manipulates and arranges this information in order to obtain a numerical value in the range of 0-10. As previously stated, the vector Pedagigical Educational Properties contains in its fields, numbers and strings obtained from learning object description. The first step is to transform the string values in a numerical value in the range 0-10. Obviously we manipulate only the information of interest and that can be transformed (for example not description). At the end of this phase we obtain the difficulty field number calculating a weighted average of all values. In particular we give a greater weight to features as difficulty and semantic density. The Interactivity Level field contains the interactivity level preferred by the student within a training resource. Our module divides the interactivity level in 10 sublevels (from very low level to very high level) and assigns a numerical value (from 1 to 10) to each level. Also in this case the system retrieves information from the description of learning object through the most appropriate fields of {Pedagogical Educational Properties} (for example interactivity, interactivity level and so on) and {technical requisites} (for example format: in this case we give greater values for format such as video and flash animation and lower values for formats such as .doc, .pdf, .ppt, etc.). The Size field describes the connection capacity generally used by the student and so the size in bytes that the resource should have to be more easily available to the student. Also in this case for the description of the student our tool creates a 10 level range and assigns a numerical value (from 1 (0-100 Kbytes to 10 (upper 2 Megabytes)) to each level. For the learning object the information are obtained manipulating the field Size of the vector {Technical requisites}. The software module executes the same operation for the Typical Learning Time features that describes the time usually spent by the user in attending the lesson. In order to obtain the best correspondence between User Resource and Training Resource, we will calculate a correspondence index (Ind) by using the following formula, where [alpha], [beta], [chi], [delta] are integer values:

Ind = [alpha]|Difficulty[.sub.u]-Difficulty[.sub.r]|+[beta]|Interactivity[.sub.u]-Interactivity[.sub.r]|+ [chi]|Size[.sub.u]-Size[.sub.r]|+ [delta]|Time[.sub.u]-Time[.sub.r]|

As can be deduced from the formula, all the single contributions are taken into account, opportunely evaluated by the weight [alpha], [beta], [chi], [delta], provided by the single components. The a value is higher than all the others in order to emphasize the difficulty of each contribution since it is not appropriate to offer training contents with difficulty levels that are superior to those presented by the student profile. The formula gives back Ind value 0 when the resource observed presents exactly the same values as the user resource obtained from the user profile. The more the index values are distant from the value 0, the more the resource observed is distant from that student needs. Our tool builds a lesson using a fundamental content (the most suitable content for the user), some additional contents and a final test. This content can also be used during the normal lessons in the presence of the teacher. After the evaluation of the final test the user profile and the course structure is updated. In fact once the course structure and the contents most suited to the student are defined, he/she will start to attend lessons through the scheme shown in Figure 6.

The structure of the course described through the SCORM standard alternates lessons with evaluation phases (in accordance with the pedagogical model). After attending a lesson, the student will answer multiple choice questions connected to the subjects that have been studied. The platform will automatically make the test on the basis of both the user profile described in LIP and the objectives connected to the lesson and specified in the Objectives tag of the SCORM standard. It will choose among the questions previously included by the teacher and described through the QTI standard. In fact, in such a tag, it is possible to specify experiences and knowledge that should have been acquired by the student at the end of a lesson. At the end of the evaluation phase, if the student overcomes a given threshold, he will access a "deepening" section where it is possible to autonomously look for further contents. Obviously the outputs of our tool are standard files (in XML language) that the current e-learning platform and also a simple web browser can interpret (Figure 7).

EXPERIMENTAL RESULTS

In our experimentation we have considered the course of "Introduction to Java programming" at the Electrical Engineering Faculty of the University of Salerno. At the start of the course the teacher has described the ontology, the contents related to every topic, the tests, and the user profile of about 50 students using the previously described tool. The result of this process is depicted in Figure 8.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

The entire number of contents of this course is about 150 and there are about 15 different contents for every subject. The contents are power point presentations, movies, word documents, html pages, flash animation, and combinations of these. The teacher of this course prepared some of these contents while others were retrieved from other similar course. The same teacher described every content. For every subject we have scheduled a maximum of two lessons and for every lesson the system has to propose a main content and a wealth content. With this information our automatic contents selector can build a customized course for every student according to the rules previously described. To measure the exactness of our approach we have introduced a "similarity index" that measures the deviation of selected contents from a given student profile. We have defined "similarity index" as:

Similarity Index = [[.sub.i]average[.sub.i]]/4

with i=Distance, Interactivity, Size and Time and where average[.sub.i] is equal to:

average[.sub.i] = [N.[k=1]][[I[C.sub.ki] - [U.sub.i]]/N]

where N is the number of selected contents, [C.sub.ki] and [U.sub.i] represent the numerical value of interest coefficient. As previously stated, the main aim of this index is to estimate the distance between student needs and selected contents. Obviously if this index is close to 0 the learning path and its linked contents are really close to user requirements. In Table 6 we show the contents that better match a student with the numerical representative vector {1,5,5,2}. This vector is obtained by analysing the user description according the rules previously described for contents. Besides the similarity index another important parameter is the length of representative content and user vector (i.e. the module values of the vectors). Obviously a similar value of length for the user profile digest vector and the content description digest vector means a suitable match. In Table 6 we can see that the maximum similarity index value is lower than 1, in other words, the selector chooses contents that on average do not exceed by one level every vector component.

In Figures 9 and 10 we have shown the distance length between the user profile (gray color) and the selected contents (black color). In Figure 9 we can say how the selected main contents show a trend similar to user description vector length. In fact the difference between the two vectors is very low (maximum 0.33). In Figure 10 we can see how the situation for the wealth contents is slightly different. In fact these contents do not have the best match with the user profile and they must only support the student's further in depth study. The content selector can also propose to the student, the 10 best contents.

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

CONCLUSION

In this article we have illustrated a web-based tool that can help teachers and students in the search of contents that can support the traditional formative processes. The module that creates the contents sequence forming the learning path is based on standards description of data and makes use of a combined acquisition of information coming from both the user profile and content description to realize the best match between students' demands and contents features. Using this tool a teacher can reach an important objective: the possibility to organize, in a simple way, teaching material tailored to specific learning goals. This tool also generates metadata according to a standard that every platform can use. In particular the course description that represents the sequencing of contents can also be viewed in a normal web browser. In the future we are aiming to improve the scoring function: in general we are trying to use data mining techniques (Bayesian Networks for example) to search the best match between contents and students.

Table 1 Typology's Fields Description

Field                             Explanation

Typology                          A briefly description of the learning
                                  object
Identifier                        Global unique label for learning
                                  object
Title of the resource             Learning object's name
Author of the resource            Learning object's author name
Date of creation of the resource  Resource's data of creation
Language                          Learning object's language. For
                                  example: "en"--"it"
Description                       Describes learning object's content
Keywords                          Contains keyword description of the
                                  resource

Table 2 Ontology's Fields Description

Field        Explanation

Purpose      Charateristics of the resource described by this
             classification entry. For example: Discipline,
             Prerequisities, Idea, Educational Objectives
Taxonpath    A taxonomic path in a specific ontology
Taxon        An entry in a classification. An ordered list of
             Taxons creates a taxonomic path: this is a path from a more
             general to more specific entry in a classification. For
             example: Physics/Acoustics/Instruments/Stethoscope
Description  Description of the other resource identified in Resource
             field
Keyword      Contains keyword description of learning object relative to
             its stated purpose
Relation     Features of the resource in relationship to other learning
             object
Kind         Nature of the relationship between the resource being
             described and the one identified by Resource field
Resource     Resource the relantionship holds for

Table 3 Pedagogical Educational Properties Fields Description

Field                    Explanation

Pedagogical Educational  A briefly description of the Pedagogical
Properties               Educational Properties of learning object
Interactivity            The type of interactivity supported by the
                         learning object
Resource Type            Specific kind of resource, most dominant kind
                         first.
                         For example: Exercise, Simulation, Slide,
                         Narrative,...
Interactivity Level      Level of interactivity between an end user and
                         the learning object.
                         For example: very low, low, medium, high, very
                         high
Semantic Density         Subjective measure learning object's usefulness
                         as compared to its size or duration
Teaching Context         The typical learning environment where use of
                         learning object is intended to take place. For
                         example: Primary Education, University First
                         Cycle ...
Age Range                Age of the typical intended user. For example:
                         suitable for children over 7.
Difficulty               How hard it is to work through the learning
                         object for the typical target audience. For
                         example: very easy, easy, medium,...
Learning Time            Approximate or typical time it takes to work
                         with the resource.
                         For example: 1h and 30 minutes

Table 4 Pedagogical Educational Properties Fields Description

Field                 Explanation

Technical requisites  A briefly description of the Technical requisites
                      of learning object
Format                Technical data type of the resource. For example:
                      text/html, video/mpeg
Size                  The size of the digital resourxe in Bytes
Location              A location or a method that resolves to a location
                      of the resource
Required Software     Needs in order to access the resource
Resources
Duration              Time a continous learnig object takes when played
                      at intended speed, in seconds

Table 5 Rights Fields Description

Field      Explanation

Rigths     Condition of use of the resource
Cost       Whether use of the resource requires payment
Copyright  Whether copyright or other restrictions apply
Rating     Description of the other resource identified in Resource
           field

Table 6 Obtained Results for a {1, 5, 5, 2} Student Profile

Student       166/324

              Difficulty  Interaction  Size  Time  User Vector Length
User Profile      1            5         5     2           7.42

Course Schedule for Student 166/324

Lesson                         Learning Content    Feature

Introduction: Main Content     Introduction.ppt    1,5,5,3
Introduction: Wealth Content   Introduction.pdf    1,5,6,3
OOP: Main Content              OOP.ppt             1,6,4,2
OOP: Wealth Content            OOPIntr.exe         1,4,4,2
Exception 1: Main Content      ExceptionIntro.avi  1,5,5,2
Exception 1: Wealth Content    Exception.ppt       1,4,5,3
Exception 2: Main Content      Exception.html      1,5,4,3
Exception 2: Wealth Content    ExceptionJava.doc   1,5,6,4
Thread: Main Content           Java_Thread.ppt     1,5,5,3
Thread: Wealth Content         Thread.doc          1,6,5,2
Swing: Main Content            Swing.ppt           1,5,4,3
Swing: Wealth Content          Java_Gui.ppt        1,5,6,2
Stream 1: Main Content         JavaStream.avi      1,5,5,2
Stream 1: Wealth Content       Stream.avi          1,5,7,3
Stream 2: Main Content         JavaIO.ppt          1,5,4,3
Stream 2: Wealth Content       Stream.pdf          1,5,7,2
Socket: Main Content           Socket.html         1,5,5,2
Socket: Wealth Content         Socket.rtf          1,5,5,3

                                                         Content Vector
Lesson                         Typology of contests      Length

Introduction: Main Content     Power Point Presentation      7.75
Introduction: Wealth Content   Pdf File                      8.43
OOP: Main Content              Power Point Presentation      7.55
OOP: Wealth Content            Flash File                    6.08
Exception 1: Main Content      Avi Movie                     7.42
Exception 1: Wealth Content    Power Point Presentation      7.14
Exception 2: Main Content      Web Pages                     7.14
Exception 2: Wealth Content    Word Document                 8.83
Thread: Main Content           Power Point Presentation      7.75
Thread: Wealth Content         Word Document                 8.12
Swing: Main Content            Power Point Presentation      7.14
Swing: Wealth Content          Power Point Presentation      8.12
Stream 1: Main Content         Avi Movie                     7.42
Stream 1: Wealth Content       Avi Movie                     9.17
Stream 2: Main Content         Power Point Presentation      7.14
Stream 2: Wealth Content       Pdf File                      8.89
Socket: Main Content           Web Pages                     7.42
Socket: Wealth Content         Word Document                 7.75

Lesson                         Similarity Index

Introduction: Main Content            0.25
Introduction: Wealth Content          0.5
OOP: Main Content                     0.5
OOP: Wealth Content                   0.5
Exception 1: Main Content             0
Exception 1: Wealth Content           0.5
Exception 2: Main Content             0.5
Exception 2: Wealth Content           0.75
Thread: Main Content                  0.25
Thread: Wealth Content                0.25
Swing: Main Content                   0.5
Swing: Wealth Content                 0.25
Stream 1: Main Content                0
Stream 1: Wealth Content              0.75
Stream 2: Main Content                0.5
Stream 2: Wealth Content              0.5
Socket: Main Content                  0
Socket: Wealth Content                0.25

References

ADL SCORM, (2003, September). Retrieved April 17, 2005 from http://www.adlnet.org/

Alessi, S, & Trollip, S. (1996). Computer based instruction methods and development. Upper Saddle River, NJ: Prentice Hall.

Bretherton, F.P., & Singley, P.T. (1994). Metadata: A user's view. Proceedings of the Seventh International Conference on Scientific and Statistical Database Management, 2, 166-174.

Chandrasekaran, B., Josephson, J.R., & Benjamins, R. (1999). What are ontologies, and why do we need them? IEEE Intelligent Systems, 14(2).

Cirillo, F., Cozzolino, A., De Santo, M., Marsella, M., & Salerno, S. (2000). A metadata based distance learning platform. Proceedings of IEEE International Conference on Systems, Man, and Cybernetics, 1, 76-82.

Colace, F., De Santo, M., Molinara, M., & Percannella, G. (2003). An automatic learning contents selector based on metadata standards. Proceedings of IEEE International Conference on Information Technology: Research and Education, 2, 243-249.

Colace, F., De Santo, M., & Vento, M. (2003). Evaluating online learning platforms: A case study. Proceedings of IEEE 36th Hawaii International Conference on Systems Sciences, 3, 437-446.

Hernandez-Dominguez, A., & da Silva, A.D.S. (2001). An adapted virtual class based on intelligent tutoring system and agent approaches. Proceedings of IEEE International Conference on Advanced Learning Technologies, 1, 243-244.

IMS Global Learning Consortium (2003, September). Retrieved April 17, 2005 from http://www.imsproject.org/

Jonassen D.H., Peck K.L., Wilson B.G., & Pfeiffer W.S. (1998). Learning with technology: A constructivist perspective. Upper Saddle River, NJ: Prentice Hall.

Kent, J.J., & Schuerhoff, M. (1997). Some thoughts about a metadata management system. Proceedings of the Ninth International Conference on Scientific and Statistical Database Management, 1.

Moon, G.S. (2001). Design and implementation of metadata management system for WWW coursewares. Proceedings of the 31st Annual Frontiers in Education Conference, 2, 6-10.

Murphy, L.D., (1998). Digital document metadata in organizations: Roles, analytical approaches, and future research directions. Proceedings of the 31st Hawaii International Conference on System Sciences, 2, 267-276.

SHOE Project (2003, September). Retrieved April 17, 2005 http://www.cs.umd.edu/projects/plus/SHOE/

F. COLACE, M. DE SANTO, M. VENTO

Universita degli Studi di Salerno, Fisciano (SA), Italy

[e-mail]

[e-mail]

[e-mail]