=Paper= {{Paper |id=None |storemode=property |title=None |pdfUrl=https://ceur-ws.org/Vol-766/paper28.pdf |volume=Vol-766 |dblpUrl=https://dblp.org/rec/conf/istar/ColomerF11 }} ==None== https://ceur-ws.org/Vol-766/paper28.pdf 
        CEUR Proceedings of the 5th International i* Workshop (iStar 2011)




     Implementing Structural Measures over i* Diagrams

                              Daniel Colomer, Xavier Franch

            Software Engineering for Information Systems Research Group (GESSI)
                         Universitat Politècnica de Catalunya (UPC)
                         c/ Jordi Girona 1-3, 08034, Barcelona, Spain
                             {dcolomer, franch}@essi.upc.edu
                        http://www.essi.upc.edu/~gessi




       Abstract. Measuring is a key issue in any software-related activity. In the
       context of the i* framework, we are implementing Measufier, a prototype for
       measuring i* diagrams in terms of properties that may be derived from their
       structure (structural measures). The prototype works over i* diagrams
       represented by the iStarML interchange format, and provides some facilities for
       managing measures' catalogues, customizing the measures to the analyst needs,
       and computing the measure over particular diagrams.

       Keywords: i*, iStar, structural measures, Measufier, iStarML.




1 Introduction

Measuring is a fundamental activity for assessing the quality of conceptual models of
any kind (“you can’t control what you can’t measure”). i* models are not an
exception to this rule. Some theoretical works have been proposed in the i*
community for defining measures over i* diagrams [1]. However, there are not tools
in the i* marketplace offering the capability of defining and applying those measures.
Our proposed Measufier tool is a first step to bridge this gap. This is the first version
of Measufier (1.1) presenting a set of basic functionalities, to be enlarged in future
versions. Its status thus may be considered quite preliminary. The tool may be
downloaded from http://www.essi.upc.edu/~gessi/Measufier/resources.html, where
some basic tutorial may be also found.


2 General Description

Measufier offers three main functionalities:
 – Measure definition. It allows defining structural measures over i* diagrams
      according to the principles presented in [1]. At the current prototype all the
      measures are kept in a single catalogue.
 – Model management. Several i* diagrams represented in the iStarML
      interchange format [2] can be loaded in the context of a user session.




                                            157
        CEUR Proceedings of the 5th International i* Workshop (iStar 2011)




  –    Measure evaluation. Measufier supports the evaluation of measures selected
       from the catalogue over i* diagrams loaded in a session.
   Given the use of iStarML, Measufier may be easily interconnected with tools that
have the ability to export models into this format. Also, the openness of this
interchange format makes it possible to apply Measufier to different variants of i*
supporting thus most i* modeling frameworks.
   Figure 1, left, shows two example i* diagrams. They represent two alternative
social systems for a Pediatrics Hospital. We want to analyse them quantitatively as a
way to support informed decision-making. We decide to use concepts from social
networks. Figure 1, right, shows the name of four measures from this field and select
just one. This measure, InDegreeMeasure, provides an estimation of the strategic
importance of one element to its environment. It is defined as the number of incoming
connections that an element has.




  Figure 1. Two i* diagrams for a Pediatrics Hospital and the outcome of Measufier.




                                           158
         CEUR Proceedings of the 5th International i* Workshop (iStar 2011)




3 Conclusions

Measufier is covering a gap in the current i* tools landscape. Given its capabilities, it
may complement other analysis tools and may complement existing techniques that
require some quantitative analysis of i* diagrams, e.g., analysing the adequacy of the
model as starting point of an MDD process [3].
   Because it is a first version, the tool has still many limitations, being the most
important not having the ability to deal with arbitrary measures.
   Our future work includes:
   – Building a comprehensive catalog of measures. This catalogue will be indexed
       by concept or intended use of the model: e.g., social measures, measures for
       software architectures represented by i* diagrams 4; etc.
   – Implementing the connection with several i* tools like OME, jUCMNav,
       TAOM4E, REDEPEND, etc. We will explore two non-exclusive ways. First,
       implementing an export facility in this tool for generating iStarML (already
       done for OME [5] and jUCMNav [6]). Second, offering Measufier as a service.


Acknowledgements

This work has been partially supported by the Spanish project TIN2010-19130-C02-
01. We would like to thank Carlos Cares and Lidia López for their help.



References

1. X. Franch, G. Grau, C. Quer: “A Framework for the Definition of Metrics for Actor-
   Dependency Models,” RE 2004.
2. C. Cares, X. Franch, A. Perini, A. Susi: “Towards Interoperability of i* Models using
   iStarML,” Computer Standards & Interfaces, 33(1), 2011.
3. F. Alencar, B. Marín, G. Giachetti, O. Pastor, J. Castro, J. Pimentel: “From i* Requirements
   Models to Conceptual Models of a Model Driven Development Process,” PoEM 2009.
4. G. Grau, X. Franch, “On the Adequacy of i* Models for Representing and Analyzing
   Software Architectures,” RIGiM 2007.
5. C. Cares, X. Franch, “A Metamodelling Approach for i* Model Translations,” CAiSE
   2011.
6. D. Colomer, L. López, C. Cares, X. Franch, “Model Interchange and Tool Interoperability
   in the i* Framework: A Proof of Concept,” WER 2011.




                                             159