Accession Number : ADA623581


Title :   Consistent Evolution of Software Artifacts and Non-Functional Models


Descriptive Note : Final rept. 1 Oct 2011-30 Sep 2014


Corporate Author : AQUILA UNIV (ITALY)


Personal Author(s) : Cortellessa, Vittorio ; Di Marco, A ; Di Ruscio, D ; Pierantonio, A ; Arcelli, D ; Eramo, R ; Trubiani, C ; Tucci, M


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/a623581.pdf


Report Date : 14 Nov 2014


Pagination or Media Count : 6


Abstract : This project has focused on the problem of filling the gap between non-functional analysis of software models and software development artifacts. Its main motivation stems from the fact that, still today, non-functional analysis and software development activities are quite loosely interconnected, thus the additional value that they can lead to each other is often lost. Instead it is commonly recognized that a tighter interconnection of practices, instruments and notations would give rise to more accurate analyses and more effective feedback from analysis results back to software artifacts. The confidence that advanced model-driven techniques can be suitable within this context came from previously successful collaborations among members of our research group that includes researchers from the Model-Driven Engineering and from the Non-functional Analysis areas. The seed of this project's activities was our first experiments on bidirectional transformations between software artifacts and performance models. Our intent was to understand whether this type of transformations could close the loop in both directions that are: transforming software artifacts into analyzable performance models (forward path), interpreting performance analysis results in order to remove problems from performance models and propagate the changes back to software artifacts (backward path). Starting from that point, the project has addressed three main research questions that, along the project development, have emerged as key points in the backward path context, which is the most problematic one. They can be summarized as follows: (1) Is it more advantageous working on the performance side or on the software side to solve problems identified with performance analysis? Can synergy be found among techniques adopted on these sides? (2) How can advanced model-driven techniques help in the model refactoring that is required to remove performance problems detected from the analysis?


Descriptors :   *SOFTWARE ENGINEERING , ITALY , NANOPARTICLES , PERFORMANCE(ENGINEERING)


Subject Categories : Computer Programming and Software


Distribution Statement : APPROVED FOR PUBLIC RELEASE