Knowledge-level Automated Software Engineering

FIRB - ASTRO DIT-PRJ-04-017

Status NOT active project
DISI role Partner
Project type Research Project
Dimension National
Acquisition date 2002-11-05
Start date 2002-11-05
End date 2005-11-05
SAP code 40100532

Project details

Project astract Although state-of-the-art software engineering techniques provide tools that are suitable for high-quality software development, an obstacle to their application is the difference between the operating system in which software has to operate (in which keywords are ``actors'', ``responsibilities'', ``aims'', and ``goals'') and the system, which is a collection of software modules, data structures, and interfaces.<br/><br/>As a result, the analysis, specification, and development are very complex activities and therefore often delegated to few experts, who, thanks to their experience on the field, are able to map high-level abstract concepts (such as ``actor'', ``role'', and ``goal'') into appropriate software modules; this is done using also advanced methodologies and progressive refinements. This requires high indirect costs for training or for hiring of experienced people, as well as costs due to additional activities such as documentation.<br/><br/>Such costs are often problematic, expecially for small-size and medium-size enterprises, where the shortage of resources does not allow for such investments, and the high level of competition (for example, for applications on the Web, or e-commerce and e-bossiness) forces a decrease of the time to market of the products.<br/><br/>Such problems lead to the lack (or insufficient) of application of the correct methodologies for software development, resulting in a quality loss, and a lack of certification. Even neglecting the intrinsic hardness of software development (which is only partly attenuated by software engineering methodologies anyway), if the developed software has to satisfy the client's needs, what results is a fragile product, which is also hard to modify due to the obscurity of the correspondence between the initial specification and the finalized product.<br/><br/>The aim of this project is to make precise software engineering techniques and methodologies usable for every enterprise, even those of small or medium size. This way, a larger set of developers will be able to realize high-quality certifiable software with a short time to market, and with limited costs.<br/><br/>Such aim will be reached by means of the development of a methodology that includes both a model of the operating system in which the final product has to run, and a model of the environment, with high-level concepts such has "actors'', "roles'', and "goal''. Such a methodology would closely relate the initial specification and the final product. This aspect of the methodology is referred to as "knowledge level software engineering''.
Keywords agent oriented software engineering, TROPOS, knowledge level software engineering, model checking, planning
Fundings 4093000 €
Partners
  • SRA - ITC-IRST
  • DIST - Università di Genova
  • DIS - Università di Roma 1
  • Delta Dator
  • DIST - Universit
  • DIS - Universit

DISI Sub-project details

Project astract The research unit of the University of Trento will be responsible for WP2 "Specification and Analysis based on Goals and Actors" and it will work on WP1 "Methodology and Process based on the goal/actor paradigm" and on WP3 "Automated Reasoning for Formal Verification and Validation". In particular, the research unit will use its competences on Requirements Engineering and Agent-Oriented software Engineering for the definition of the incremental transformational approach for the requirements analysis, for the definition and the implementation of algorithms for the goal diagram analysis, and the study and the definition of the organizational styles and multi-agent patterns. In WP1 the unit will take care of the development of the methodology based on the paradigm goal/actors and in WP3 it will work in the activity for the definition of validation techniques based on prepositional satisfiability. Finally, in WP7 the unit will work on the support tool development and in WP8 on the case study.
Keywords agent oriented software engineering, knowledge level software engineering, model checking, planning
Fundings 816000 €
Manager Fausto Giunchiglia