´╗┐

Challenge 3: Cost-Effective

The return from using traceability is adequate in relation to the outlay of establishing it.

Introduction

  • Description

    By establishing traceability automatically and early in the vision scenario, the engineer is alerted to product requirements that she overlooked early on in the engineering process, avoiding the need for costly rework later. Such knowledge has been accrued over a myriad of projects thanks to traceability. The engineer is able to focus on her job, and on those analyses that demand her expertise and decision-making skills, and is not distracted by building in this traceability support as she works. Moreover, by having the opportunity of creating or maintaining the traceability on-demand later, the engineer does not have to worry about having a traceability problem in the future; she knows that missing traceability can always be established cost-effectively if and when needed, based upon tried and tested best-of-breed techniques and methods.

  • Problem Addressed

    Complete traceability is often impractical, expensive to establish and not always necessary. Too much time can be invested in establishing traceability that may never be used or useful on a project, such as the provision of rich link semantics that are never actually exploited in traceability-related queries or analyses. It is difficult to know what is

Research Projects

Major Research Project:

Requirements Cost-Benefit Models

Develop cost-benefit models for analyzing stakeholder requirements for traceability and associated solution options.

Supporting Research Projects:

Research Project 2.

Agree metrics for measuring traceability cost-effectiveness.

Research Project 3.

Create decision support tools and impact analysis tools for making traceability return on investment decisions, such as a mechanism to globally and locally optimize the traceability solution based on stakeholder requirements for traceability, the available resources and the return on investment required.

Research Project 4.

Gather and disseminate benchmark empirical studies for researchers to demonstrate the cost-effectiveness (or not) of various traceability processes, techniques, methods and tools, as part of the Traceability Body of Knowledge.

Research Project 5.

Devise a way to associate a cost and a benefit profile with every trace that is brought into existence and maintained.

Research Project 6.

Create a value proposition to demonstrate that traceability addresses the right goals, in a decomposed way, versus goal displacement (where the means used to achieve a goal become more important than the goal itself).

Industry Practice

Industry Practice 1.

Practitioners consult the Traceability Body of Knowledge to understand the cost-effectiveness of existing and new techniques and methods when making traceability strategy decisions.

Industry Practice 2.

Practitioners use decision support tools and impact analysis tools to explore the cost-effectiveness of employing various and mixed traceability strategies on a project.

Industry Practice 3.

Practitioners track the return on investment from traceability on a project and contribute these data to the traceability community.

General Comments

Please login first in order to leave a comment here.
Author: St├ęphane Mercier Date: 2013-06-20 08:37:16
Message: Some suggestions... Traceability should be complete or it's useless. We should first establish a traceability model: which kind of artifacts should be traceable, and how they should be linked. This model shall not be too complex or it wont be usable. We should not first put our focus on automating the traceability but, to determining the rules and artefacts required to create and maintain the traceability. Once this is done, we will be able to create tools that will automating traceability correctly. We need to proceed in the same way than for to create software: 1- Define user requirements (Grand traceability challenge?). 2- Define software specifications (detailed traceability model and its rules that should work if they are applied manually ) 3 - To Create tools that are conform to the specifications to automate traceability functions. Our main problem is that we don't have a commun traceability model: detailed traceability model and its rules. And a traceability model could been defined as: Artifact type ex.: user requirement, class, function, use case, software functionnal specification, etc. - How it is related to another artifact type. ex: The artefact type A is always linked to artefact type B, using link type C. Link type - Rules of propagation (from-to, to-from, bidirectionnal, write, read-only, etc.). Have fun with these suggestions....
© Center of Excellence for Software Traceability