This paper describes an approach to the design and implementation of ontology driven dynamic web sites combining ontologies and wiki technologies. The core of the architectural solution proposed is completely based on ontologies rather than on more traditional forms of persistent data storage facilities, such as relational databases. This approach provides a flexible support to the design and implementation of web portals in which navigation schemes are not entirely predetermined but are instead influenced by actual relationships among the contents of the ontology, that are used to generate web pages as well as hyperlinks. A wiki technology is integrated with this approach in order to create pages not directly derived by elements of the ontology, but also to enrich the textual contents with suitable formatting, images and hyperlinks. The application of this approach to the realization of a web portal is also described; the portal is devoted to enable a number of scientific communities to share archaeological knowledge about the Silk Road domain.
The introduction of technologies enabling the development of data driven web sites represented an extremely important step in the process that lead the initial version of the Web to its current state of pervasive diffusion and impressive growth rate. Data driven web sites are able to generate web contents and pages according to the information persistently stored in a suitable module, such as a relational Data Base Management System. This represents the last tier of an architecture encompassing a middle tier that is able to interpret suitable templates for web pages that are instantiated according to actual data stored in the data tier, and that are eventually passed to the first tier responsible of the visualization of the page (i.e. the client tier, a common web browser). Traditionally these templates have been realized by integrating modules developed in common programming languages with a web application (e.g. Common Gateway Interface or Java Servlet technologies), or embedding “programming languages–like” control structures and abstractions in traditional web pages through scripting languages (e.g. PHP, ASP, JSP). Some of the most relevant abstractions related to a data driven web site, and determining its organization and function are thus implemented and expressed in terms of a database logical schema and specific programming language control structures embedded in page templates.
Even if this approach surely represents a huge improvement with respect to static
web sites, that are essentially repositories of HTML documents and embedded
multimedia contents, a set of different research efforts have been carried out in an attempt
to supply higher level abstractions to support the design and implementation of web
based advanced information systems and applications. Some of these approaches, for
instance [
The following section will elaborate the research context in which this work is set,
briefly introducing relevant related works, while Section 3 introduces the architecture
and the various functionalities offered by NaVEditOW [
In his proposal for a global hypertext [
Within a short period, the Internet and World Wide Web have become ubiquitous and today Internet is full of data-driven Web sites: yellow pages, e-Commerce sites, digital libraries are only the most common examples. But also, forum, blog, wiki, video and photo sharing website, hotel booking, online auction website, online community, Web-based email client and Web mapping service are all data-driven Web sites.
A data-driven Web site is much easier to maintain than a static Web site: most content changes require no change to the pages. Instead, changes are made to the data source and this source could be enterprise or organization database. Sharing a common data source, the Web application can be easily integrated within information system.
Client Web Browser
Application Code Web Server So, for example, in a manufacture company when a new product is added to the enterprise database, it could be automatically displayed on the corporate Web site products catalog.
One of the most common architecture used in developing of dynamic data driven web sites, is shown in Figure 1. In particular, the persistent data storage is generally delegated to a relational database management system. The web server generally hosts dynamic pages including server side scripts necessary to query the data tier and collect the information required to compose pages related to the contents of the database. These dynamic pages represent a sort of template for the web pages that must be generated according to the stored data, specifying different aspects ranging from the queries that must be submitted to the database to retrieve them, to the kind of links that must be created to support the navigation inside the data.
An evolution of this approach is the adoption of the the MVC (Model View
Controller) [
Generally, the data is stored in a relational database and the model is represented by a database schema. In our opinion, it is difficult to represent all the relations that could be present in a complex domain (such the archaeology domain) with a database schema. In many cases, is difficult to translate some aspects of a conceptual schema (such the generalization) into the relational model. Some kind of extra-relational constraint are not representable within the database schema and requests database store procedure or external application code. There are also problems to manage the database schema, for example, the is no way to check the schema consistency.
The proposed approach provides the adoption of an ontology [
Radeox Wiki Engine
NavEditOW
tionality. In particular, we employed NaVEditOW [
The main motivations of this architectural modification are related to a more comprehensive exploitation of the explicit relationships among the concepts described in the ontology, and a simpler integration of this kind of architecture with instruments and systems developed in the Semantic Web context.
Other approaches of ontology-driven Web site architectures can be found in the
literature: in [
A large number of approaches for combining Semantic Web and Wiki technologies
are currently under development; some relevant examples are Makna [
In this section we present a system for web based navigation, querying and updating of ontological KBs. The presented software allows exploring the concepts and their relational dependencies as well as the instances by means of hyper-links; moreover, it provides a front-end to query the repository with the SPARQL1 query language.
NavEditOW is an environment for navigating, querying and A-Box2 editing of OWL3 (Web Ontology Language) ontologies through a web-based interface.
With respect to ontology navigation, since individuals play a fundamental source of knowledge for people accessing an ontology, A-Box navigation should be supported. From this perspective, it is important to support not only navigation of concept hierarchies defined by isA relations, but also other forms of ordering on the individuals domain. As a first example, locations can be linked through a partOf relation and it should be possible to group locations under the location of which they are all subparts (e.g. browsing all countries of which Europe is composed of starting from Europe); as a second example consider a number of historical periods ordered according to a relations such as followedBy: it should be possible to exploit this relation to sort such individuals from the first to the last one.
With respect to editing, although T-box4 maintenance require a certain knowledge about ontological formalisms, A-Box editing should be supported taking into account: (i) cardinality and range restrictions defined in the T-Box need to be respected; (ii) ranges of properties and individuals stored in the ontologies can be also exploited to drive and suggest instance update. Moreover, contextual editing, that is, the editing of the A-box while browsing the ontology, should be supported. 1 SPARQL (SPARQL Protocol and RDF Query Language) is an RDF query language standardized by RDF Data Access Working Group of the World Wide Web Consortium. For more information, see http://www.w3.org/TR/rdf-sparql-query/ 2 The A-Box is the “assertion component” of a knowledge base. 3 http://www.w3.org/TR/owl-features/ 4 The T-Box is the “terminological component” of a knowledge base.
Although end-users may not be familiar with query languages, the possibility to perform expressive queries should be supported. From the one hand a language as much as similar to well-known query languages for relational databases language should be preferred. On the other hand, interfaces enabling non expert users querying the ontology should be developed (e.g. query forms).
In the following paragraphs, we presents more details about each of these tree basic functionalities and the application architecture. 3.1
With the ontology navigation interface the users can view ontology individuals and
their properties and browse properties via hyperlinks. Browsing the ontology is
essential for the user in order to explore the available information and it also helps non-expert
users to refine their search requirements, when they start with no specific requirement in
mind [
The root of the navigation tree is the OWL class Thing, and the rest of the tree is organized as follows: under the root node, there are the top-level classes (i.e. direct subclasses of Thing); each class can be expanded to show its subclass hierarchy and its individual members; individual-to-individual tree connections are defined according to a number of selected tree-like properties (e.g. partOf ); finally if total order relations are selected, they are exploited to order individuals within a given level of the tree. In order to distinguish between classes and individuals, the former are represented in petrol blue while the latter are shown in shocking pink. An example of a navigation tree is presented in the Figure 3. In this example Geographical Region is a class and Cental Asia, Uzbekistan, etc... are its instances. This individuals, in turn, are connected each other by the partOf directly property. 3.2
The application allows the users to create, edit and remove individuals of the ontology, their properties and, in particular, their labels. In fact, To ensure multi-languages support, it’s possible to define several labels in different languages for every individual.
The properties of each classes are defined in the T-Box. Two types of properties are distinguished: object property is a binary relation between two individuals and datatype property is a binary relation between an individual and a literal (a primitive type, like string or number). Cardinality and range restrictions for properties are used to support users while editing. For example, in an archeological ontology, the class TypologyOfArchaeologicalObject has the property builtOf. This property has no cardinality restriction (so it can have zero, one ore more values) but Material is specified as range (co-domain). For instance, Sword is an instance of TypologyOfArchaeologicalObject and has the property builtOf Metal, where Metal is an instance of Material.
There are a datatype and an object editor in the framework: the datatype editor allows editing a literal values, displayed as a text input box, object allows defining the property values presenting the user a tree for selecting the values; the individuals displayed in the tree are only those that are valid for the property range.
Literal values can be not only plains strings but also Wiki text. This allows the users to insert long formatted text with link in the generated pages. For example, we can have the ’Uzbekistan’ individual in our ontology. ’Uzbekistan’, as an instance of Country, has some properties such population, capital city, part-of, currency, history, foreign relations, etc. Same (e.g. population) are plain literals (numbers, dates, strings), others are references to other ontology instances (e.g. the filer of the property ’capital city’ is Tashkent which is an instance of the class City). History of foreign relations are Wiki Text, so they can be formatted, contains images an hyperlinks. Wiki Text can also contains ’macro’ that allows to include in the text information from the ontology or results of SPARQL queries.
NavEditOW adopts an editing policy. Knowledge engineers interacting with domain experts, are supposed to create the ontology and edit the Tbox with standard design-time editing tools such as Prote´ge´ [17]. They are supposed to test its quality with reasoners such as Racer [18] or Fact++ [19] in order to check if the Tbox (i) is consistent and (ii) it captures the intended meaning (by concept hierarchy inference). End-users can edit the Abox; they can insert new instances, asserting that they are member of a concept of the ontology, and assert relation statements; this means, that they can edit the textual descriptions contained in the ontology. The assumption behind this editing policy is that end users are not familiar with formal semantics. First, learning to use concept constructor and ontology axioms is difficult for end users, if one goes beyond simple Tbox updates, such as the insertion of a new concept as subconcept of a concept in the ontology. Second, every axiom has logical implications that are difficult to control, especially for people non expert in formal semantics. 3.3
The first implemented query interface is the SPARQL query form in which users can write query in the SPARQL language, display results in paginated tabular form and navigate through results via hyperlinks. This interfaces is very flexible because the users can write arbitrary queries but is not suitable for end users. Another kind of query interfaces is based on a predefined set of queries. Every predefined queries is composed of a description in natural language, a SPARQL query with eventually free parameters and a list of parameters. Every parameters have a label, a type and eventually a restriction on the valid values (e.g. a parameter can be filled only with instances of a specific class). For this interface, users can select a query by its description, fill the query parameters and execute it. The results are presented as the results of the other query form. The queries can be inserted in the Wiki Text through a macro, for example the following macro shows in the resulting page all the provinces of the Uzbekistan: {sparql | SELECT ?x WHERE { ?x silkrode:partOf silkrode:Uzbekistan . ?x rdf:type silkrode:Province }} 3.4
From an architectural point of view, the functionalities (ontology editing, navigation and querying) of the user interfaces are based on the Application API, as shown in Figure 5. The main purpose of this API is to support the manipulation and querying of the ontology through the standard SPARQL query language and through a set of specific adapters, shielding the user from the underlying semantic framework. A plug-in interface, in fact, makes the application independent from the adopted specific semantic framework (as long as SPARQL is supported). A different adapter for every semantic framework is needed because SPARQL is only a query language and does not offers any data manipulation statements (e.g. INSERT, UPDATE, DELETE).
The adapter API supports the manipulation and query of the RDF graphs in two different ways: frame-centric and statement-centric. The former view is similar to the object-oriented paradigm: every resource is viewed as an object and properties as attribute. This view is used for ontology navigation and resource manipulation. Statementcentric is a lower level view in which the graph is represented as a set of triples. Each triple contains three components: subject, predicate and object.
Currently two semantic framework adapters have been implemented: the first one is a wrapper for the Jena Semantic Web Toolkit, the other for the Sesame Framework. The former is an open-source Semantic Web Toolkit5 aimed at supporting the development of applications that use the Semantic Web information models an languages [20]. We have initially adopted this framework since it matched our requirements, it is widely used within the Semantic Web research community and well documented. This first adapter implementation works well with small ontologies, but fails with larger ones since Jena is not suitable to manage an huge amount of data (a performance evaluation of several frameworks suitable for large OWL ontologies is presented in [21]). For this reason, we choose to develop a new adapter for the Sesame Framework6 [22]. Sesame provides a number of functionalities for handling (querying and manipulating) RDF graphs. It also supports various types of storage facilities and inference mechanisms. 5 http://jena.sourceforge.net/ 6 http://www.openrdf.org
Remote SPARQL Client External Reasoning Service
A-Box Editor
Semantic Navigation
Interface
Semantic Query
Interface Radeox Wiki Engine
Application API
Web Interface SPARQL
Semanti Framwework Adapter Sesame
Jena Persistent Storage itc rk anm eowm lryea eS fra The default implementation supports inferencing and querying on RDF Schema but lacks a specific support for OWL. 4
SilkRoDE (Silk Roads in the Digital Era) is a project that aims to collect, structure and diffuse all knowledge concerning the Cultural Heritage of Central Asia, including not only archaeological material, sites and historical monuments but also data from fields such as geography, sociology and ethnography. It is an open and evolutive project, which functions as an intelligent network linking all interested institutions, research groups and scholars, that worked and working in Central Asia.
As a first step, SilkRoDE aims to create a Wiki, used on a daily basis by all specialists of Central Asia, interested members of the general public and those involved in Cultural Management. The creation of this Wiki will be possible thanks to a collaboration between specialists from the Humanities and from Computer Sciences in a very close multidisciplinary context. One of the keys to the success of the SilkRoDE project is the decision of all participating scholars, institutions and research groups to work together as Equal Partners. This does not mean that all resources are simply pooled together but that each resource is clearly associated to the authors and funding agencies that enabled its creation. SilkRoDE thus aims to be a Network rather than a new institution. In this perspective, the NavEditOW system represents the general platform adopted by the Project to organize and manage the various different contents of SilkRoDE Wiki. The main aim is to share information and knowledge, and all partners should thus be enabled to employ the system to publish, connect personal data and information, and to identify other groups or Institutions working or interested in same themes. This need for a collaborative and collective participation in data and information collection phases lead us to choose a Wiki–based approach to develop the SilkRoDE portal.
The introduced approach and the NavEditOW system were adopted to represent and organize basic information about institution, research groups and scholars having active researches in Central Asia, as well as information about relevant bibliography and important cultural and archaeological sites. The platform will be integrated with webGIS: the possible connection between specific entities of the SilkRoDE ontology and georeferred entities stored in a GIS will support the visualization of spatial position and distribution of various entities in dynamically generated maps.
The ontological approach provided the required expressiveness and flexibility even in these starting phases of the project, in particular in order to support rich forms of navigation among stored contents. In particular, this approach provided the possibility of representing and managing relationships like “is-a” and “part-of” without flattening the related entities in a single table or splitting them in different tables, as would have been necessary adopting a traditional relational database. For instance institutions, research groups and individual scholars are all actors of the SilkRoDE ontology, in other words, they are individuals belonging to classes that are related to the Actor class by an “is-a” relation. It is now possible to define a generic relationship binding an archaeological site to an instance of the Actor class or one of its subclasses, without the need to define different relationships. This flexibility in defining and establishing relationships among individuals will support further types of analysis aimed, for instance, at identifying possible connections among actors that are working on similar or related research issues, or geographic areas, or adopting similar methodologies. 5
The paper has described an ontology driven approach to the modeling, design and implementation of dynamic web sites. In particular, we aimed at simplifying the realization of web based systems that exploit and give access to a shared ontology, but these systems should also look like traditional web sites and support simple forms of navigation. To this aim, we endowed the system with a wiki engine and we included documents, and in particular wiki pages, in the ontological tier, to support a simple form of editing of pages that give the site an ordinary structure and appearance, that can be enhanced by means of the exploitation of the underlying domain ontology.
The motivations of this effort, as well as related work and the research context were introduced, and the NavEditOW framework was described, in terms of provided functionalities and architecture. A case study providing the application of the introduced approach and framework was also presented.
Future works are mainly aimed at extending the range of the represented and managed concepts, with particular reference to the various topics that can be used to characterize relevant scientific publications, in an effort similar to the one described in [23]. Moreover, in the medium term, the project will consider the possibility to realize specific wrappers able to to export contents complying to the CIDOC Conceptual Reference Model [24], so as to achieve a high level of interoperability with this relevant standard for cultural heritage information organization.
With respect to the editing policy, it will be tested the possibility to enable endusers to insert new concepts and the respective subclass-relations. Moreover the more extended use of reasoning capabilities will be investigated. 17. Gennari, J.H., Musen, M.A., Fergerson, R.W., Grosso, W.E., Crube´zy, M., Eriksson, H., Noy, N.F., Tu, S.W.: The Evolution of Prote´ge´: an Environment for Knowledge-Based Systems Development. Int. J. Hum.-Comput. Stud. 58(1) (2003) 89–123 18. Haarslev, V., Mo¨ ller, R.: Racer: a Core Inference Engine for the Semantic Web. In Sure, Y.,
Corcho, O´ ., eds.: EON. Volume 87 of CEUR Workshop Proceedings., CEUR-WS.org (2003) 19. Tsarkov, D., Horrocks, I.: Fact++ Description Logic Reasoner: System Description. In Furbach, U., Shankar, N., eds.: IJCAR. Volume 4130 of Lecture Notes in Computer Science., Springer (2006) 292–297 20. McBride, B.: Jena: a Semantic Web Toolkit. IEEE Internet Computing 6(6) (2002) 55–59 21. Guo, Y., Pan, Z., Heflin, J.: An Evaluation of Knowledge Base Systems for Large OWL Datasets. In McIlraith, S.A., Plexousakis, D., van Harmelen, F., eds.: International Semantic Web Conference. Volume 3298 of Lecture Notes in Computer Science., Springer (2004) 274–288 22. Broekstra, J., Kampman, A., van Harmelen, F.: Sesame: an Architecture for Storing and Querying RDF Data and Schema Information. In Fensel, D., Hendler, J.A., Lieberman, H., Wahlster, W., eds.: Spinning the Semantic Web, MIT Press (2003) 197–222 23. Bonomi, A., Mantegari, G., Vizzari, G.: A Framework for Ontological Description of Archaeological Scientific Publications. In Tumarello, G., Bouquet, P., Signore, O., eds.: Semantic Web Applications and Perspectives 2006, Proceedings of SWAP 2006, the 3rd Italian Semantic Web Workshop, Pisa, Italy. Volume 201 of CEUR Workshop Proceedings. (2006) 24. Doerr, M.: The CIDOC CRM, an Ontological Approach to Schema Heterogeneity. In Kalfoglou, Y., Schorlemmer, W.M., Sheth, A.P., Staab, S., Uschold, M., eds.: Semantic Interoperability and Integration. Volume 04391 of Dagstuhl Seminar Proceedings., IBFI, Schloss Dagstuhl, Germany (2005)