Ontologies to support the definition of the knowledge society

This paper presents Arianna, the approach used in Diviana, a small Italian e-consulting organization, for defining a standard in describing the semantic of e-government services. Such a standard is born sharing (and not imposing!) information with Italian Local Public Administration (LPA) Entities, especially Comuni. In order to describe semantically the LPA services, our approach models ontologies using the Unified Modeling Language (UML). The UML model is automatically converted in a SVG site semantically browsable, as further explained later, and a set of XML Schema Definition (XSD) files, describing data structure used in the services. Moreover, such XSD files represent the communication standard intra and inter LPA entities; in fact the XSD describes the base elements for implementing the application interoperability. The automatic generation of a prototypal XForms user interface completes the approach and allows people without technical skill to validate the semantic of services. All the information contained in the model repository are made available through one e-government catalogue over the internet. The results obtained during the tasks of standardization conclude the paper. Introduction Internet diffusion supported a high standardization level: everyone can browse millions of sites which use heterogeneous technologies without having any problem. In the second half of 90s, the interest for the Application Interoperability grew up rapidly. The growth did not correspond to an efficient technology but with the promises it holds. In fact, nowadays, Application Interoperability is pervasive in many Internet activities and it is transparent for the users. Application Interoperability is the set containing everything needed by two or more applications to interact each others for reaching a specific business goal. For example, let us consider a b2b environment: e-commerce sites are tightly integrated with just-in-time producers, transport/logistics and payment sites. Usually, the Application Interoperability is implemented using ad-hoc interfaces: each site has its specific technology and language. 2 Diviana e-consulting, Via Pisandro, 91, 00124 Rome – Italy (abarone, dipietro)@acm.org, http://Arianna.diviana.net Such sites are islands in the cyberspace: the users can browse them but each site has laws, rules and languages for its own. This situation could be acceptable in the early Internet age but today it is obsolete. In fact the islands want to gather together in archipelagos using common laws, rules and languages. The Italian Public Administration (PA) is an archipelago of Administrative Entities including 8100 Towns (Comuni), 102 Province and 20 Regions, each one with its Administrative Autonomy and the need of interacting and integrating each other. The Central Government and several other government related Agencies complete the picture. In order to accomplish the integration, the critical step is defining a communication standard between Administrative Entities. A standard could be defined using an imposing way (by law) or using a sharing process. The first way has no chance of success because it reduces the decisional autonomy of the Administrative Entities, guaranteed by the Italian Constitutional Law. The latter way aims to share the domain information, following the method successfully used by the Internet Engineering Task Force (IETF) [1], and the W3C[2] for introducing new Internet technologies. Defining and tuning a shared standard is a longer and more complex task than imposing a prefab one. In fact, the process has to overcome difficulties, rivalries and prejudices and the process managers must be super-partes and must have a strong authoritativeness. Moreover, the complexity of defining a PA standard is enormous and it is easy losing the governance of the project. In order to manage such a complexity we need a formal approach using a formal representation. We decided to use UML modelling for representing the standard, as we are going to explain in the following sections. In order to build up a standard it is necessary to face various problems which concern with juridical, managing and monitoring aspects, beside the technology ones. Even the impacts of the change on persons have to be considered. Our goal is to describe this situation in an easy to understand and not ambiguous manner in order to discuss and share the standard. We propose the use of a formal model to accomplish such a goal. The whole process and the tools described in this paper have been implemented by an articulate solution, named Arianna. As in the myth, Arianna gave Theseus a wool thread to find his way out the labyrinth after he killed the Minotaur, our solution give the user the ability to discover his/her path into the complexity of the Public Administration in general and, more specifically, into the Italian one. Arianna can be found at http://Arianna.diviana.net. In the remaining of this paper, Section 2 introduces the concept of ontology, Section 3 summarizes the solution paradigm, Section 4 describes the architectural model, Section 5 describes the Interoperability Pattern, Section 6 makes some consideration about the reuse based approach, Section 7 shows the UML repository structure and section 8 shows the different ways Ontologies to support the definition of the knowledge society 3 used to made available the knowledge base to the users over the Internet. Section 9 makes a short description of the catalogue, while Section 10 describes the Ontology Definition and Usage Process. Section 11 presents some numerical results. Some final remarks and forwards in Section 12 and 13 conclude the paper. The LPA ontology The complexity described in the introduction must be managed, and the first enabling step is a shared knowledge. The key concept driving our approach on sharing knowledge is the LPA ontology definition. The ontology definition is the basis for managing the interaction between a large amount of subjects, each one with its decisional autonomy, to identify the involved concepts, information, elements, subjects and roles and their mutual relationships, giving each one both a semantic definition and an intrinsic structure description. And everything must be shared, with time and patience. To achieve this goal, an analysis of the services offered by the LPA has been made. The result has been the identification of the services, each one with their respective clients and providers. For each service, the following items, have been identified: • Information, each one with the description of its: • ownership definition; • full information structure; • lifecycle. • Normative; • Administrative practices; • Available best practices, if any. Services are also been classified using different taxonomies, giving to specific classes of users/providers an easy way to retrieve and access them. The service provider describes the specific LPA responsible for the service. In the case of complex services (i.e. services involving more than one cooperating LPAs) it describes the single LPA responsible for the entire service, usually representing the one facing with the user. The client describes the specific subject that will use the service. It can be a citizen, a company or a third part acting as an intermediate. The full ontology definition is based on a specific solution paradigm and is described using an UML knowledge base, as described in the remaining of then paper. The solution paradigm The solution paradigm is based on the following three main aspects: 1. The definition of an architectural reference model; 2. The definition of an interoperability pattern; 3. The reuse based approach. 4 Diviana e-consulting, Via Pisandro, 91, 00124 Rome – Italy (abarone, dipietro)@acm.org, http://Arianna.diviana.net Architecture: a Reference Model As the main goal of our work is to define a semantic approach capable of running in different physical environment, we didn’t prescribe a specific architecture, but only describe a generic logical layering, permitting hosting of specific implementations. The result is an n-tier architecture, briefly described in the rest of this chapter and depicted by the following figure: