 |
| Wednesday, June 4, 2003: 10:15 AM Go to 11:15 AM Go to 2:30 PM |
|
| Double-Track Session! |
W4 Measurement Programs
Effectively Using Inspection Data Analysis
Edward F. Weller, Software Technology Transition
|
What you do with the data you collect from inspections is a key differentiating factor between inspections and other forms of peer reviews. That’s because this data is critical to building a closed-loop control system that allows you to improve your inspection process and achieve effectiveness above 70-80%. But inspection data can either be used or abused. This session examines many of the inspection techniques Ed Weller has used in more than 13 years of implementing and leading inspection programs. He then shows you how to convert best-practice techniques into project success.
• Identify the data you should collect
• Learn to avoid traps and pitfalls in data analysis
• Discover how to get the most from your inspection process |
|
|
W5 Predictive Measures
Predicting Software Errors and Defects
Aura Yanavi, JP Morgan Chase
|
Did you know that traditional forecasting techniques can be applied to software to predict the quality of the software and estimate the number of critical defects? This session presents two “new” metrics — the quality factor and the functionality factor — which are based upon the number of test requirements. These two metrics can be used not only to predict the quality of the software and the number of defects going into the system test, but also to estimate the quality of production releases. The model has been validated by actual use where it’s accurately predicted the number of defects delivered into the systems test environment. Aura Yanavi teaches you how to use this practical model to help you predict the number of critical and major defects in your applications. She also shows you the benefits and disadvantages of this type of prediction-based measurement.
• Learn to assess the complexity of major software releases
• Forecast the number of critical defects for the next major release
• Discover two new software metrics — quality and functionality | |
W6 What to Measure
Metrics for Software Leads
Scott Jamison, General Dynamics
|
As more organizations push toward increased cost and schedule management, software metrics have become integral parts of the software or project lead’s job function. This session presents a set of practical measurement tools to help software leads track the schedule and cost status of a software project. Scott Jamison offers an explanation of how software metrics tie into the lifecycle of the system through the concept of “earned value.” You’ll walk away with tools to turn your developers into effective status reporters.
• Find out how to measure the work that must be performed on a development project
• Learn to use tools to divide up the work into measurable units
• Perform risk analysis to measure out-of-scope work and schedule or cost slippage | |
|
| Wednesday, June 4, 2003: 11:15 AM Go to 10:15 AM Go to 2:30 PM |
|
W9 Predictive Measures
On-Time Projects Using the Predictive Confidence Ranking
Tim Stefanini, Velocitos Inc.
|
One of the hardest things to manage in a project is schedule slippage. Even areas that seem predictable can fall unexpectedly behind. What we need is a way to get advance warning of problem areas as soon as possible. By implementing a confidence tracking system, you can quickly pinpoint the areas that need the most attention. The confidence feedback can then be gathered on either a large-scale or a feature-by-feature basis. Tim Stefanini delivers general strategies for delivering a system like this in your organization, and demonstrates how to track confidence ranking versus actual delivery dates.
• Find out how you can set up confidence management ranking in your projects
• Evaluate granularity in confidence measurement
• Learn how to respond to predictive confidence ratings to reach project success | |
W10 What to Measure
Measuring the Right Stuff: Establishing Information Needs
Peter Baxter, Distributive Software
|
Establishing a measurement process has evolved from, “If it moves, count it,” all the way to today’s information-needs-based approach. The problem with the information-needs approach is that when identifying and defining what to measure, this robust and flexible framework focuses on the information needs of managers. This session gives you simple techniques for identifying the diverse information needs within your organization — information needs that become the requirements of your measurement process and lead to more useful, effective measurements. Peter Baxter also examines the relationship between information needs and the success of your measurement process.
• Obtain an overview of the ISO-15939/PSM Measurement Information Model
• Get 11 techniques for determining information needs
• Learn methods for extracting information needs from people, documents, and work products | |
|
| Wednesday, June 4, 2003: 2:30 PM Go to 10:15 AM Go to 11:15 AM |
|
W14 Measurement Programs
CMMI and Agile Methods: Friends or Foes?
Thomas M. Cagley, Jr., The David Consulting Group
|
Agile methods have a reputation for being at odds with structured approaches, but that isn’t necessarily the case. The debate over whether agile methods can be used while implementing the CMMI process improvement construct has been charged with histrionics and negative reactions across the board. In this session, Thomas Cagley shows you why, although the answers are not black and white, agile methods can indeed be incorporated within a CMMI approach. He also offers advice on when to integrate the two methods, and to what degree.
• Examine how the CMMI and agile methods can be incorporated
• Learn the development methodologies needed to support different types of work breakdown structures
• Determine whether or not CMMI and agile method integration are for you | |
W15 Predictive Measures
A Rigorous Model for Estimating Web-Based Applications
Robert Daudt, Pacific Northwest National Laboratory
|
This presentation illustrates the typical problems associated with estimating Web-based systems, and then outlines the business value of using a rigorous estimation versus an ad-hoc process to try to combat them. Robert Daudt presents two primary methodologies — Putnam-Meyers Software Equations and COCOMO II — and gives you criteria so that you can select which model is right for your organization. He walks you through a complete estimation process, demonstrating the use of rigorous methods and the typical pitfalls that should be avoided. The techniques presented can be applied to applications other than Web-based as well.
• See a working example of Web-based application estimating using rigorous techniques
• Learn the fundamentals of rigorous software estimation using the Putnam-Meyers Software Equations and COCOMO II models
• Find out how to select your estimation approach based on existing data | |
W16 What to Measure
Designing a Clean-Slate, Low-Impact Metrics Program
Russell Roundtree, Software Development Technologies
|
For any metrics program to be widely used, the information must be easily digested and understandable within the context that the data represent. When launching a new metrics program, the best way to begin is to set clear boundaries on what you’re doing and what you’re trying to achieve. Russell Roundtree explains how he and his team set out on a metrics-building mission by focusing on just a small number of measures that allowed them to show trends important to the project and management. He’ll also show you how to take small beginnings and turn them into internal process metrics and even organizational metrics down the road.
• Determine what you should worr y about when starting a metrics program
• Learn to set expectations for your program
• Set phases for more sophisticated future metrics | |
