One of the most important steps in a Web site quality evaluation project is the selection of which aspects to consider. In terms of methodology, this means defining a model for the site. In some cases it is possible to use standardized models, such as “syntactic” models, but this is not possible when the evaluation must also consider aspects that have to do with the domain and the specific aims of the site or more generally when the evaluation aims to consider the “semantics” of the site. The process of identifying and adapting a quality model requires, apart from time and resources, the contribution of experts in the domain of the site. In this paper we propose to use ontologies to improve the efficiency of this “instantitation processs”. To analyze the feasibility of the approach we have looked at two applications in the tourism sector. The results, while preliminary, are encouraging. Moreover, some critical and delicate points were identified as priorities for future research.
Quality in Web sites is determined by several diverse
factors, some of which are general and therefore are
considered for all types of site and in all domains. Such
features include, for example, the correct functioning of
the site, its conformity with standards of language use or
of accessibility as described in normatives such as the
Web Content Accessibility Guidelines of the W3C
(http://www.w3.org/WAI/GL/) or the U.S. Section 508
Guidelines (http://www.section508.gov/), or standards
introduced in Italy through the Stanca law in 2004
(http://www.pubbliaccesso.it/biblioteca/documentazione/s
tudio_lineeguida/), which requires these standards for
government sites. Other factors or characteristics are
more specific and depend on the type as well as the
domain of the site. Therefore in cases where it is not
possible to use “standard”, “syntactic” models -
generalpurpose and domain-independent – it becomes necessary
to develop quality models that take also these features
into consideration (among models having a standard
version we can look at WebQual [2] and WebQEM [
Recent years have seen increased interest in the
development and application of ontologies. This has
meant firstly the definition of languages and
environments to set up ontologies. Examples of languages
are DAML (Description Logic Markup Language), OIL
(Ontology Interchange Language), RDF (Resource
Description Framework) and OWL (Ontology Web
Language), a semantic markup language for publishing
and sharing ontologies on the World Wide Web
(http://www.w3.org/TR/owl-ref/). Among the more
numerous applications of ontologies are projects related
to the Semantic Web, to obtain “an extension of the
current Web in which information is given well-defined
meaning, better enabling computers and people to work in
cooperation.” [3]. An obvious application of ontologies is
in Web services; among the most recent works and most
frequently downloaded (was in the fifth position of the
Top 10 Downloads from ACM’s Digital Library in
December 2004), we can cite [
The proposal described in this paper differs from these insomuch as it is applied at a different level: it is designed to use ontologies to support the development of quality models that are specialized for Web sites. As such it is a conceptual as well as methodological activity, and is applied at a meta-level with respect to the application of the models themselves.
The paper is structured as follows: the first section gives a general description of ontologies for tourism. Following this we look at the concept of tourist destination, where we present the classification used to apply the “Quality Model Factory” methodology, which we are going to extend to include domain ontologies. The third section puts forth two possible applications for the development of specialized and modular evaluation schemes, respectively for hotel Web sites and for tourist destinations. The conclusion summarizes the preliminary results of these applications, underlining the critical points emerging from the application of the methodology and which require further study.
Tourism represents approximately 11% of worldwide GDP, according to the World Travel & Tourism Council (http://www.wttc.org/). Adding to this is the notable growth in the number of tourism-related Internet transactions in recent years (e-commerce). For example, in 2004, 40% of U.S. travelers who use the Internet claimed to make all of their travel purchases online, versus 29% in 2003 (www.tia.org). In this context, the quality of Web sites becomes a vital strategic factor for all actors involved. Because tourism is a transversal sector – or “umbrella industry” – it has contributions from other sectors, thus an analysis of the sector must have input from different fields such as transport, culture, and sport, to name a few. This fact explains the existence of numerous ontologies for tourism. An exhaustive classification can be found in [1].
Existing ontologies are both general for the tourism sector as well as specific, the latter referring to particular domains. In the first category we have the ontology developed for the Harmonise project, whose goal is to develop an ontology-mediated integration of tourist systems following different standards so that organizations can exchange information without changing their data structures (http://www.harmo-ten.info/). In addition there is the Mondeca´s tourism ontology, which includes tourism concepts from the WTO (World Tourism Organization) thesaurus. At this writing this ontology has 1000 concepts that describe accommodations and transportation and a few other secondary elements related to geography, health and immigration (http://www.mondeca.com/).
There are over ten elements on the list of domainspecific ontologies that can be useful for the tourist sector, including geographic ontologies, means of transportation ontologies, gastronomy ontologies, etc. [1].
General – or sometimes called upper – ontologies also exist and aim to gather definitions and concepts that together make up what is known as unspecialized common knowledge. One of the best known of these is WordNet – more appropriately referred to as a lexical reference system (http://www wordnet.princeton.edu/) which was extended from solely English into other languages through the EuroWordNet (http://www.globalwordnet.org/gwa/wordnet_table.htm).
Looking at the ontologies for hotels and tourist destinations (the organizations examined for the feasibility analysis of the approach proposed here) the following facts emerged.
Since the concept of hotel is part of common knowledge, the notion is present in WordNet. More specifically, for each concept – in this case hotel – WordNet gives information on the generalization, specialization and also on “part of” relationships. A description of the linked concepts is available at Answer.com (www.answer.com). In short, in WordNet, focusing to concepts that are directly linked to hotel, thus exploring only relationships represented by arcs going out from the node of hotel, we obtain (see also figure 1): hotel is a kind of: • building, edifice — a structure that has a roof and walls and stands more or less permanently in one place kinds of hotel: • hostel, hostelry, inn, lodge — a hotel providing overnight lodging for travelers • motel, motor hotel, motor inn, motor lodge, tourist court, court — a hotel for motorists; provides direct access from rooms to parking area • resort hotel, spa — a fashionable hotel • Ritz — (informal) an ostentatiously elegant hotel • ski lodge — a hotel at a ski resort Parts of hotel: • hotel room — a bedroom (usually with bath) in a hotel
Furthering the analysis by using the concept of hotel room, WordNet gives other information: kinds of hotel room: • adjoining room — a hotel room that shares a wall with an adjoining room but is not connected by a door • connecting room — a hotel room that shres a wall with an adjoining room and is connected by a private door hotel room is a kind of: • bedroom, sleeping room, chamber, bedchamber — a room used primarily for sleeping. hotel room is a part of: • hotel — a building where travelers can pay for lodging and meals and other services projects. Answers.com also provides hypertext definitions (where they exist) found in other sources such as Wikipedia (http://www.wikipedia.org/), which can be useful to integrate with those contained in WordNet, and also their translation into diverse languages.
The more challenging problem is to identify an ontology for tourist destinations, although this difficulty is justified by the complex definition of destination, a concept which, albeit only recently, is assuming an increasingly important role in the tourist sector in general. Basically, the features that identify a tourist destination and distinguish it from what is simply a local offering of a product or service that can be of interest to tourists are: a well-defined geographic area with identifiable borders and a territorial identity; the presence of numerous operators with different prospectives and objectives that makes it necessary to devise a shared strategy in presenting the offering consisting of attractions and services specifically catering to tourists in the location; an understanding of the nature of the potential demand for the tourist products offered; awareness of the need to balance tourism’s exploitation of resources with ecological, environmental and community stewardship.
A classification of the destinations serving the leisure tourist segment identifies eight distinct types of destination based on the goals for the vacation and the principal attractions present at the destination (table 1) [8]. The table shows key information about the defining features of a destination. Once established, these aspects can then serve as input when determining the requirements and the quality factors for the Web site of the destination.
The definition of tourist destination and the classification of diverse types of destination show how the necessary concepts belong to diverse domains. WordNet gives no treatment to the concept of tourist destination. Numerous ontologies were examined as part of this research, but none proved able to cover all the elements that characterize a destination: it is thus necessary to use different ontologies to describe the geographic area, the different attractions, sports, transport, etc. On the other hand, general ontologies (upper ontologies) contain a lot of information that is not useful because it is related to concepts that differ greatly from those used in the tourism context. In addition, general ontologies for tourism cover only some of the necessary areas (for example, Mondeca gives good treatment to accommodations and transportation) but not for some indispensable aspects of a destination (examples being cultural, natural and artistic attractions or events).
For this reason we have used WordNet in our feasibility analysis, and the decision as to which concepts to consider was informed by the definition of destination itself and by the table. For example, for alpine destinations we used the concepts of sport, landscape, nature, etc. We were thus able to simulate the nucleus of an ontology for tourist destinations. 3. The use of ontologies to define specialized models
In [9] we described a modular and scalable approach – the Quality Model Factory - to define specialized quality models identifying the specific features of tourist destination Web sites. Its goal was to introduce a systematic way to define a “personalized” evaluation framework. The use of modules derives from the application of reuse of artefacts [13] as a viable practice for definition of evaluation models. Scalability is obtained thanks to the adoption of a general conceptual framework, for Web site quality, the 7Loci meta-model (www.economia.unitn.it/etourism/pubblicazioni.asp).
This model introduces seven dimensions used to classify the numerous features of a Web site that can then be evaluated. The dimensions are Identity, Content, Services, Location, Maintenance, Usability and Feasibility.
The foundational procedure that serves as the starting point in developing a modular model for a given class of Web sites is outlined in the steps in table 2.
The procedure has two parts; the first is the instantiation of the 7Loci model, reusing where possible a model already defined for one of the types of site in a class. This step is necessary in order to define the model using requirements as the starting point. In the second part the specific elements for the type of site are identified. When talking about the sites of tourist destinations, this means translating these elements into points (characteristics) within the evaluation model; for example, the unique features of the site for a seaside destination will be converted into points in the model used to evaluate that type of site, while features for another type of destination (religious, for example) may not contain those same points. In our previous work we looked at the aspects in table 1, which in general can be Instantiating Web Sites Quality Models: an Ontologies driven Approach 55 found in classification schemes of the sites of a specific category. When describing the procedure, for the sake of simplicity we refer to “questions” to insert in the evaluation modules. In reality this is only one way of formulating the points or factors of the evaluation model; besides interrogatives (e.g., in boolean questions) they can also be described in declarative form.
In this paper we want to broaden the approach to be able to develop models for diverse types of sites, and such generalization is conceptually based on the use of ontologies. The steps of the procedure where ontologies can be used are shown in table 2.
The logical architecture of the Quality Model Factory is depicted in figure 2 (a Quality Factory to support “information quality” assessment is described in [4]), where the database “Web site classification” is substituted with a database containing the ontologies.
Requirement elicitation
Web sites classification Quality models repository Requirements repository
Web site evaluation project 7Loci instantiation Quality requirements Quality evaluation Modules
Quality model factory Quality evaluation Validation
From among the different actors in the tourist sector present in the territory we focused on hotels and tourist destinations when studying our approach. For both categories tourism is the core business and the Web site is a strategic tool. For the tourist destination the Web site serves to promote and commercialize the products as well as to support the different actors.
In a previous project we built a model for the comparative evaluation of the Web sites of about 200 hotels. The model consists of 18 questions: four for the dimension Identity, nine for Content, two for Services and one for Localization, Management and Usability. The model is a useful reference to check the results that can be obtained with the Quality Model Factory which is strengthened through the use of ontologies. In this case we have used the concepts related to hotel in WordNet. In addition, for the requirements of the owner and user, we have put together information emerging from the research conducted on the hotels [6]. In this case it meant developing a model more similar to that constructed with the contribution of experts, and also applicable to all hotels. Thus we did not intend to construct different models for different types of hotels.
Applying the Quality Model Factory approach to tourist destination Web sites without using ontologies produced a series of models, each containing about 100 factors that are a reference for evaluating and which can be produced with ontologies [9]. Moreover, we thus obtained important results for the adaption process, where ontologies are necessary to specialize the general-purpose models. Also emerging, in particular, was how the first two dimensions of the 7Loci meta-model depend more on the type of site (with an average of about 80% of specialized factors); about one third are specialized for Services, ten percent for Usability, Maintenance and Localization (related principally to the presence of different target tourist groups). Development of a model for the evaluation of Web site quality for tourist destinations is underway as part of the joint project IFITT/WTO (www.ifitt.org/)sito. The model foresees two levels of assessment: the first looks at general aspects and therefore at requirements that all tourist destination Web sites must satisfy; the second is called Strategy Based Full Web Site Evaluation. A limitation of the IFITT/WTO model is that there is a measurable gap between the two levels. Essentially the first level uses a general model for all destinations while the second must be defined ad hoc. We propose the use of ontologies to specialize the quality models in a systematic way.
The project is still underway, the results obtained thus far have revealed the following aspects:
The dimensions of the 7Loci meta-model where the use of ontologies is more straightforward are those which are more “semantic”, thus Content, and Services; for these dimensions it is possible to use the hierarchies contained in the ontologies as check-lists to identify aspects for which the site must give information or support for services.
For the dimension Identity it is necessary to integrate the ontologies with the aims of the site and be able to connect them to the concepts in the ontology that mainly contribute to the creation of the image of the organization; these concepts are usually specializations of “father” concepts; in the case of hotels, their specialization is for specific target markets, for example motor hotels.
For all the dimensions of the 7Loci meta-model, the combined use of ontologies, of the standardized version accompanied by a list of the main aims of the site (no more than five elements: e.g., target, business functions, general links), by people with limited experience working in Web site quality (undergraduates students), made it possible to develop “draft” specialized models in a short time and which can be rapidly verified and completed by an expert. On the whole, the methodology proposed makes it possible to notably improve the efficiency of the process of defining specialized models.
Substantial initial effort is required to identify ontologies for the different domains necessary to cover the tourist sector. Moreover, existing ontologies are heterogeneous with regard to the coverage of the domain they refer to. Nonetheless, they can be reused for numerous categories of operators and entities.
Statistical analysis of the terms used in sites under analysis provides useful information for the choice of ontology, but principally to choose the concepts to use within the ontology to instantiate the quality models.
Most existing ontologies are in English language; but it could be necessary to have ontologies in other languages.
As for their implementation, most ontologies that can be used in the tourism context are written in DAML and some in OWL. This means that if we want to create environments to support the “Quality Model Factory” it is necessary to extract a version of at least two different types of file that even non-experts can understand. 10. References [1]. Bachlechner D., D10 v0.2 Ontology Collection in view of an E-Tourism Portal, E-Tourism Working Draft 5, Oct. 2004, http://www.deri.at/research/projects/etourism/2004/d10/v0.2/20041005/ [2]. Barnes S. J., Vidgen R. T., WebQual: An Exploration of Web Site Quality, H. Mahrer (ed), Proc. of the 8th European Conf. on Information Systems, Vienna, July 3-5, 2000, pp. 298-305 [3]. Berners-Lee T., Hendler J., Lassila O., The Semantic Web, Scientific American, May 2001, 284(5): 34-43 [4]. Capiello C., Francalanci C., Pernici B., A rule based methodology to support information quality assessment and improvemnet, Studies in Communication Sciences, 4(2): 137-154, 2004 [5]. Falbo R.A., Guizzardi G., Duarte K.C., Natali A.C.C., Developing Software for and with Reuse: An Ontological Approach, Proc. CSITeA’2002, USA ACIS, 2002, pp. 311-316 [6]. Franch M., Martini U., Novi Inverardi P.L., Buffa F., Awareness and exploitation of the potential of the Web by SMTEs: The case of alpine hotels in Italy and France, Frew A.J. (ed.), Springer Verlag LNCS 2722: 318-327, 2005 [7]. Hüsemann B., Vossen G., Ontology-Driven Multimedia Object Management for Private Users – Overview and Research Issues, AIS SIGSEMIS Bulletin, 1 (1), April 2004 [8]. Martini U., Da luoghi a destinazioni turistiche. In Franch, M. (ed.), Destination management: Governare in turismo tra locale e globale, Turin, Giappichelli, pp. 67111, 2002 (in Italian) [9]. Mich L., Franch M., Martini U., A Modular Approach to Quality Evaluation of Tourist Destination Web Sites: The Quality Model Factory, Frew A.J. (ed.), Springer Verlag, LNCS 2722: 555-565, 2005 [10]. Mich L., Franch M., Novi Inverardi P., Marzani P., Choosing the "rightweight" model for Web site quality