Introduction¹

Our goal for this paper is to introduce the software named Framework System for Meaningful Learning (SIMAS), whose main role is to support online and ontology based knowledge representation. We developed this system following the current programing language tendencies, aiming towards a progressive adaptation to the specific user needs, simplifying the work processes and specializing in solutions that take advantage of the hardware capabilities; using all the available tools. For its design and development, an integration of databases and the programming languages Apache, PHP, and MySQL was conducted; resulting in robust, adaptable, responsive, secure, and effective software capable of accomplishing the administrator and target user’s main objectives.

Along with the pedagogical innovation it offers, this development tries to simplify processes and provide friendly user experiences (UXDesign). This is accomplished by evaluating the system behavior, shape diagramming, and icon development, whose features such as size, color harmony, tone similarities are adjusted to reach the expected user perception.

Ontology and education system development

While logic has to do with the relationship between variables, ontology is concerned with the variable values (^{Sowa, 2000}). Ontology is the study of being, of what something is. In the words of ^{Quine (1969)}, an ontology is the value of a quantified variable. Formal sciences like logic, or graph theory, study structures - relation among variables - while ontology studies meaning, or variable values.

Ontology is interested in entity categories and their relationships (^{Hofweber, 2009}) with the purpose of integrating knowledge about the systems that compose the components of the universe and, beyond that, about the universe itself, when looking into the most general categories. There are abstract entities (sets, concepts, numbers,...), concrete entities (objects, plants, planets,…), common sense entities (ways of analyzing the existence of something), universal entities (property, quality, attribute,…), mental entity (idea, reasoning, memory,…), vacuum entity (absence of matter).

According to ^{Husserl (2005)}, ontology is the science of the essences. His approach leads to the differentiation of a formal level, which includes the most general categories of being, and a separate level including regional ontologies which correspond to specific reality domains. Sciences, as well as digital knowledge representation systems, are dealing with regional ontologies. As a consequence, ontologies are, essentially, meaning based knowledge representations. Both in philosophy and science, ontologies arise from the need of knowledge organization.

For computer science, ontologies are particularly valuable for knowledge representation in specific domains. ^{Feldman y Sanger (2007)} represent the Ontology Domain concept as the tuple O:=(C,≤ c); made of the c concepts of set C, whose elements are named concepts, and the partial order relation ≤ between concepts c of C, named hierarchy of concepts.

For as long as computer science has had to process meaning in the knowledge systems - semantic dimension - it has taken ontology systems as valuable tools. Two examples of this fact are the WorldNet² system and the Semantic Web. The first one is built based on synonymy relationships in the English language with more than 100,000 related items. In the Semantic Web, concepts are words with attached metadata; relating them with each other and implementing intelligent searching processes. In computer science, ontologies have several uses: for knowledge acquisition and its representation, for organizing information, in consultation and use of information, and finally to improve the same learning process that are of particular interest in this work.

Learning ontologies with the support of software tools is a process that moves on progressively (^{Zou, 2007}), and involves identifying concepts with associated properties, naming them, learning relations, and building structures at different abstraction levels. Ontologies, in this case, have to be managed by intelligent agents and should be understandable by human users. The developments are promising, but still challenging.

Text mining is focused on finding relations among sets of words convergent in formal characterization of ontologies in specific knowledge domains (^{Feldman, y Sanger, 2007}). In a review of methods and tools for learning from text, ^{Gómez and Manzano (2005)} find that developed tools are applied to learning relationships, while new concepts are applied to build taxonomies. In a study about the relationship between ontologies, ^{Ibrahim et al. (2012)} realize that this approach amplifies the possibility of relating and integrating specific domains.

Component based Software engineering integrates programs as units, with the possibility of being developed and tested separately. The organization of elements in ontological structures provides a name for each component associated to a concept; assigning specific features in such a way that conditions for intelligent software reasoning are established. This tendency has initiated the development of ontology based languages -OWL-, able to use the Semantic Web advantages (^{Pahl, 2007}). In addition to component based reasoning, OWL supports on line collaborative software development.

Specific science domains are based on ontologies; structures formed by interrelated sets of terms which dynamically evolve as a result of research studies. One consequence of this process is the ontology standardization in every science; which for software system development signifies the requirement of programs that model ontology systems. Protegé³, for example, is structured as a system of frames with kinds (concepts), slots (features), and perspectives (facets), (^{Rubin et al., 2007}). This system is provided with a graphic user interface -GUI- for project development and is accessible online, to make the collaborative work of programmers easier. The systems developed in this type of environment support logical inference available for intelligent software agents. As a result, the information storage as well as the information retrieval is improved (^{Akerman y Tyree, 2006}).

The use of this kind of systems is present in the sciences and in disciplines for which ontology system modeling is a challenge (^{Zhang et al., 2013}; ^{Nick et al. 2002}). In the case of education, two pruposes are sought: on one hand, producing learning content in sciences and disciplines, and on the other hand, implementing learning systems to develop skills for handling the ontology systems of those sciences and disciplines.

The commitment of organizing content for learning specific ontologies is taken by some researchers in a similar way that word meanings are treated in the semantic web. The content units, as digital objects, are named learning objects, which are organized as elements of formal ontology structures, based on associated metadata. This approach makes logical inference feasible, easing the search and the use of components in different contexts (^{Sicilia y García, 2005}).

The concept of ontology in specific domains has been taken as the basis for building ontology networks. In a review of studies, ^{Pernas et al. (2012)} describes an application of ontology in learning systems as a tool to improve the adaptability of the learning system to the student context and background. In ^{Wang et al. (2012)}, an ontology of a specific knowledge domain is integrated with an ontology of learning objects, and one of learning strategies, in order to develop an ecological learning system. Said system is consistent with the point of view of the subject matter and both, the point of view of the learning resources and the achievement of learning skills.

In a different example presented by (^{Pernas et al. 2012}), a learning context taxonomy, a learner profile ontology, a knowledge domain ontology, and a resource and tools for learning ontology are integrated, with the purpose of improving the learning system adaptability to the user conditions. In a similar approach, ^{Knight et al. (2006)}, integrate a learning content ontology with a design process of learning object ontology, and a learning context ontology (LOCO) producing a higher level ontology using the Protegé development environment. These results provide the basis for implementing computational intelligence in the field of on line education (^{Vasilakos et al., 2004}) in which the knowledge domain, the student model, the context, the learning objects and the educational model are represented using ontologies; allowing the computer to operate based on networks of meaning.

CmapTools is created as a tool for the students to elaborate graphical networks, named conceptual maps, in which concept names are the nodes, and the links show relations between pairs of concepts (^{Novak, 1977}). In one condition, there is input information (for example, a text or a video) and the student elaborates a concept map to represent the structure of the received message; in a different condition, the student is asked to express his/her own knowledge by means of a conceptual map.

Concept maps may be used to represent hierarchical or other type of ontologies, but most of the times are applied to informal schemas (^{Novak and Cañas, 2006}). In this direction several tools have been developed (Table 1): some like Webprotégé and SWO are oriented to build ontology based applications, but most of them are applied to represent knowledge structures informally. The Table 1 shows some of the features of these tools.

Most of the software is oriented to build concept structures. However, usually they do not support a classification of relationships. The hierarchical classification is the privileged relation, but does not include other kind of systematic organization, like time sequence, systemic integration, or causal relations. Distinguishing these kind of relations helps the student to integrate specific knowledge perspetives based on the knowledge domain and specific problems.

Conceptual framework

The knowledge representation practice, as a method for building meaningful learning, is consistent with constructistic approaches in the fields of psychology and pedagogy.

Research reflected in ontological structures standardized by the academic communities is projected in education as learning content. The most widespread approach in educational practice has been the development of instructional systems, which currently tend to be based on these ontologies. Environments are also created to help students learn by means of reconstructing ontologies of the knowledge domain of study.

In a previous study, college education majors enrolled in science education courses that developed ontology reconstruction excercises for specific geography units, and futher prepared and conducted teaching practices with students. Said study showed greater inclusion of these structures in the training of their own students (^{Maldonado et al., 2001}).

Learning abstract structures is usually challenging for most students. In our approach, student participation in the contruction of knowledge representations using multimedia information, and the possibility of comparing their own ontological structure with the structure prepared by the teacher, strongly anchors the student learning process. Accordingly, a better structuring of the long term memory is expected.

The SIMAS environment - presented here - evolves from previous projects (^{Maldonado et al., 2010}), to configure a general structure - shell - that allows the user to perform the following functions: design an easy editing ontological structure (as a graph), automatically generate a structure navigable hypertext with a page for each node of the ontological structure and bind options to enable multi-media information for each page. It can be used by the teachers to develop their own hypertexts as learning resources, or by the students to develop their own hypertext projects both individually and collaboratively.

The software induces collaborative work online, the use of multiple digital contents prepared by different authors, and the creation of resources. Looking forward, this stage will facilitate the integration of software agents for different pedagogical functions such as generating questions, metamemory judgments activation, goal management, monitoring of learning strategies, and management of online collaborative activities.

The software follows the contemporary developments on ontologies and systems (^{Sowa, 2000}) and focuses particularly on hierarchical, temporal, spatial, systemic and causal relationships.

Figure 1 A representation developed by a student using the hierarchical tool 

Figure 2 A representation developed by a student sing the time line tool 

Figure 3 A representation developed by a student using the spatial tool 

Figure 4 Representation developed by a student using the systemic tool 

Figure 5 Represetation developed by the kinder teacher usin the Causal Tool 

The software provides conditions for the teachers to organize online learning resources to teach specific content domains and skills. For students, the software is an environment for studying learning units provided by the teacher; or even better, for developing knowledge representation projects, guided by the teacher, in individual or collaborative conditions. As such, teacher and peer reviews are facilitated.

The SIMAS device is oriented to help the student achieve skills for knowledge representation, collaborative ontology building, long term memory structuring. Few software developments are found with these characteristics; reason why SIMAS is considered not only a contribution for people’s learning but for the structuring of communities of learning. One of the special features is the possibility of using multimedia information available on line, implementing the content learning management.

Software architecture

The architecture to develop this Web application, was implemented in three layers - representation, logical, and data layers - and two levels - application and data levels -. The logical and the representation layers belong to the application level, and the data layer, to the data level.

Layer programming is client-server architecture; which main objective is to split the business or content logic from the design logic. The “level” expression corresponds to the way the logic layers are phisically distributed.

Software development

Open Source Technology used in the development of the software is Apache / PHP / MySQL. The construction of application and data layers was performed with the following versions:

Here the databases appear: in the case of the web pages, one of the most used managers is MYSQL, a system that allows, through a series of sentences, to have information stored in a database and retrieve it efficiently and quickly, when needed (Figure 6).

Figure 6 The main interface to enter the software (http://i3campus.co/SIMAS/) 

Interface

Viewed from its shape, SIMAS is an easy access portal and a powerful device for digital content management; it has a modern design that follows current trends, consistent with the category of intended educational environment. It is implemented to create a pleasant user experience and to support the necessary conditions to keep working on the website for developing knowledge representation projects (Figure 7). Their main features are:

Figure 7 The wellcome page, once the user has entered the software 

Figure 8 Interface to access an already created ontology, or to create a new representation 

A set of user roles were implemented to access to the platform:

Figure 9 Interface to edit a new ontology 

Figure 10 Interface to draw a new ontology 

Testing process

Once the software implementation was finished, a pilot test was conducted. Bugs were detected and fixed until a stable version was accomplished.

For an effective use of the tool, a tutorial and a manual were prepared and provided on line for the users (http://i3campus.co/SIMAS/como-funciona.php). These documents provide a detailed support for each component of the system.

With the collaboration of teachers, we observed SIMAS being used by one group of kindergarten, one group of elementary, one group of high school and one group of technical school students (Table 2). Ten observation sessions were conducted for each of the four groups and registered in field notes. At the end on this process, the teacher opinion was given explicitly trough an interview.

In a first step, the teachers built their own ontologies using SIMAS and used them to support learning strategies with the students. Different categories of ontologies were developed: hierarchical, systemic, causal, spatial, and on time line.

In a second step, the teachers asked the student to build collaboratively an ontological representation using SIMAS on a specific topic. In the case of preschool, the teacher used her own ontology to support dialogs and different learning experiences.

Table 3 summarizes the main results according to the observer field notes and the interviews with the teachers.

Observers reported verbal interaction relevant to the assigments when using the SIMAS platform. A proactive attitude on the part of all groups were also found and an emergent sense of satisfaction with the teaching-learning process mediated by ICT.

All the teachers expressed that the software was easy to use. The performance of the systems was also good enough to support the different kind of practices they developed. The interest of this particular study was not to compare the student results, but the software performance in the different activities that the teachers could implement. For example, the kindergarden teacher developed an ontological representation of the colors to support observation activities; the elementary teacher help the students to organize historical events on the time line tool; high school and technical school students search and organize information using specific ontological approaches.

The different tasks have different degree of difficulty; however the teacher reports show that in each kind of task the software was usefull, and easy so use. The student satisfaction was a common feature of the reports.

Result analysis and discussion

The observation of users working with SIMAS in real learning activities shows the platform’s great capabilities of supporting meaningful learning. The users may produce digital objects as inputs for building the meaning of an ontological representation of knowledge, but they also can integrate information available on the web. Consequently, users have a tool for the effective management of digital learning content. Positive user motivation and the improvement of long term memory skills are some of the outstanding effects reported by teachers.

Although SIMAS was designed as an application with several modules, many of them with a graphic interface for working, significant processing of information is required. Consequently, stable Internet connection is a must, in order to prevent data loss and quality decay of graphics when the user is creating ontologies.

Therefore, for subsequent versions, SIMAS would be implemented as an enriched internet application so as to reach a better performance in the following aspects:

SIMAS is a revolutionary application in its domain of use, and has a good performance for an early stage. These considerations are made thinking of later versions that expand its use to areas with low connectivity and difficult conditions.

Conclussion

The current version of the software for ontology knowledge representation on line - SIMAS - exhibits the following features and capabilities: