Annotating Regulations Using Cerno: An Application to Italian Documents - Extended Abstract

Abstract

The increasing complexity of software systems and growing demand for regulations compliance require effective methods and tools to support requirements analysts activities. Internationalization of information systems due to both economics and Web based architectures call for the application of regulations written in different languages. Thus far existing approaches for extracting rights and obligations have concentrated on English documents. In this paper, we describe the results of the application of Cerno, a lightweight framework for semantic annotation, to legal documents written in Italian. In addition, we investigate critical issues for semantic annotation tools in a different cultural and environmental context. Results obtained, while preliminary, allow us to quantify the effort needed to port tools based on Cerno and give some insight on directions of future development of a multilingual system to support semantic annotation of regulations not only in different domains, but also written in different languages.

How to integrate legal requirements into a requirements engineering methodology for the development of security and privacy patterns

“…1) require legal experts to describe patterns in natural language. Such a description is parsed by a natural language processor on the basis of a semantic template (Breaux and Antón 2008;Zeni et al 2008). The annotated description is then used to automatically generate graphical models of patterns, which are revised by security engineers using a CASE tool.…”

Requirements model generation to support requirements elicitation: the Secure Tropos experience

“…This approach allows for the development of a wide range of semantic annotation tools on top of Cerno in order to support different annotation models and the analysis of different types of input documents. For instance, Cerno has been applied in the tourism sector (Kiyavitskaya et al 2006(Kiyavitskaya et al , 2007b and in the domain of regulatory documents to extract obligations and rights (Kiyavitskaya et al 2007a;Zeni et al 2008). The approach, however, needs a preliminary adaptation of the architecture of Cerno to each particular application domain.…”