NSF Grant CCR-0306613
Transforming Requirement Specifications into Architectural Prescriptions

Project Description:

Requirements to architecture is the first and hardest step in the process of engineering software systems. Our goal is to understand how to transform requirements into architectures, and based on that, to provide practical and effective techniques, methods, processes and tools to assist in that venture. In doing this research, we take an empirically-based rather than a technology-based approach in our research -- that is, we use empirically based techniques to understand the problem deeply and to evaluate the results of our research rigorously.

Please note that any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation

Project Members (over the life of the project):


The first two papers represent our first fruits of our interviews with our architects. The third paper is a consolidation of the comparison of the van Lamsweerde and the Brandozzi/Perry methods of transforming requirements to architecture. The fourth paper explores ideas about architectural styles relevant such non-functional requirements as adaptability, dependability, and self-healing.

Currently in progress are the technical reports incorporating the results of our analyses for all the questions and the individual case studies, and the conference and journal papers from Divya's and Damien's (with Axel van Lamsweerde, U. c. de Louvain, Brussels) theses.

Summer 2004

During the summer of this first year we began addressing a third project goal in the context of a newly formed collaboration with Prof. Steve Easterbrook, University of Toronto.

Wendy Liu, a student of Prof. Easterbrooks, visited The University of Texas at Austin this summer (the first of several projected visits, as well as reciprocal visits that I will make to the University of Toronto as a member of her thesis committee) to begin studying what practicing architects do to create architectural structures appropriate for their non-functional requirements. Ms Liu has an IBM Fellowship and an entree into IBM in Toronto to complement my contact with IBM here in Austin. We expect this to be a fruitful exploration and collaboration.

The results of her visit were 1) several iterations over the questionaire we used for our architects interviews and 2) her participation in several of our initial interviews here in Austin.

The questionaire is aimed at providing as comprehensive an interview as possible regarding requirements and architecture and transforming the former into the latter. Our intent is to use the questionaire as a starting point for a conversational style, open ended, semi-structured interview.


In the first year of the project we addressed and made a good start on two of the our project goals: I was particularly fortunate to have Damien Vanderveken, a student of Prof. Axel van Lamsweerde, Universite catholique de Louvain, visiting for the fall semester. This enabled us to do an interesting comparative evaluation of Prof. van Lamsweerde's and my methods for transforming goal-directed requirements specifications into architecture specifications. We used as the basis for this comparison a Trio specification of a power plant [1, 2] that we specified in KAOS and enhanced with non-functional requirements. We then used this latter enhanced specification to create architecture specifications using both methods. The results are reported in the following report that we expect to publish as a workshop paper.

Divya Jani and Damien Vanderveken, Experience report: Deriving Architecture Specifications from KAOS Specifications. December 2003. Project Supervisor: Prof. Dewayne E. Perry.

Damien then returned to Belgium and the two students then continued using this comparison as the basis for their respective masters theses, both taking the same general approach of modifying the respective methods and extending them to incorporate in various ways the use of architectural styles and patterns to transform a functional architectural structure into an appropriate non-functional one. The two theses are listed below.

Prof. Van Lamsweerde and I expect to merge the results of these two theses into suitable conference and journal papers. I expect this fruitful collaboration to continue throughout the project.

Preliminary work

Manuel Brandozzi and Dewayne E Perry. "From Goal-Oriented Requirements to Architectural Prescriptions: The Preskiptor Process." International Workshop From Software Requirements to Architectures (STRAW2), International Conference on Software Engineering 2003, Portland OR, May 2003, pp 107-113. [pdf]

Manuel Brandozzi and Dewayne E. Perry. "Verifying the Preskiptor Process with Students: an Experimental Design", April 2002. [pdf]

Manuel Brandozzi and Dewayne E Perry, ``Architectural Prescriptions for Dependable Systems'', International Workshop on Architecting Dependable Systems, International Conference on Software Engineering 2002, Orlando FL, May 2002 [pdf] [slides for the workshop talk] [an expanded version, unpublished].

Manuel Brandozzi. "From Goal Oriented Requirements Specifications to Architectural Masters Thesis, Electrical and Computer Engineering, The University of Texas at Austin, December 2001. Supervisor: Prof. Dewayne E. Perry. [pdf]

Manuel Brandozzi and Dewayne E Perry. "Transforming Goal Oriented Requirements Specifications into Architectural Prescriptions." Workshop From Software Requirements to Architectures (STRAW1), International Conference on Software Engineering 2001, Toronto, May 2001, 54-61. [pdf]

Dewayne E. Perry - This information last updated February 2005
Copyright © 2005. All rights reserved.