GRALD: an Approach for Goal and Risk Analysis in the Development of Information Systems for the Web of Data

Marcio Louzada De Freitas,Renata Silva Souza Guizzardi,Vítor Estêvão Silva Souza

doi:10.5753/jserd.2019.18

Abstract

The publication of Linked Data on the Web regarding several application domains leads to new problems related to Requirements Engineering, which needs to take into account aspects related to new ways of developing systems and delivering information integrated with the Web of Data. Tasks such as (functional and non-functional) requirements elicitation and ontology-based conceptual modeling can be applied to the development of systems that publish Linked Data, in order to obtain a better shared conceptualization (i.e., a domain ontology) of the published data. The use of vocabularies is an intrinsic activity when publishing or consuming Linked Data and their choice can be supported by the elicited requirements and domain ontology. However, it is important to assess the risk when choosing external vocabularies, as their use can lead to problems, such as misinterpretation of meanings due to poor documentation, connection timeouts due to infrastructure problems, etc. Thus, risk identification, modeling and analysis techniques can be employed, in order to identify risks and their impacts on stakeholder goals. In this work, we propose GRALD: Goals and Risks Analysis for Linked Data, an approach for modeling goals and risks for information systems for the Web of Data.

Highlights

The Semantic Web was presented by Berners-Lee et al (2001) as the Web version that seeks to make content understandable by both humans and machines, improve search engines by giving meaning to the published content and take into account contextual information of time, space and states of things
The evaluation of this proposal was conducted by the first author of this paper and an undergraduate Computer Science student (Silva, 2017), using Web-based Information Systems developed by students of the Web Development & Semantic Web course of our Postgraduate Program in Computer Science, all of which aim to publish Linked Data
We evaluated our proposal by creating goal and risk models for these systems and searching for vocabularies based on these models

Summary

Introduction

The Semantic Web was presented by Berners-Lee et al (2001) as the Web version that seeks to make content understandable by both humans and machines, improve search engines by giving meaning to the published content and take into account contextual information of time, space and states of things. According to Bizer et al (2009), Linked Data is a set of data interconnected by URIs (Uniform Resource Identifiers) whose contents can be processed by machines, forming a Web of Data. The published content is based on the RDF (Resource Description Framework) standard and data can be extracted using SPARQL4 queries. Published data and their interconnections are described through vocabularies, i.e., schemas that describe the existing entities and the relationships between them. Such data can refer to several domains, such as Geographic, Media, Social Media, Governmental, Libraries and Education, Life Sciences and so on (Heath and Bizer, 2011)

Objectives

Results

Conclusion