[Previous entry: "Security testing"] [Main Index] [Next entry: "2011 tulip haul"]

Oct 31, 2007 2:40:04 AM Archived Entry: "Software Test Time proposal for StarEast"

Estimating test time: simple models and practical lessons

Summary: As software testers, we are regularly asked how long it will take to test a project, but are rarely given the tools to produce a scientific estimate. This presentation will introduce several ways of producing good estimates, including one specific to software testing that can be implemented in an organization with no prior history of time analysis. It will include customizable formulas for "massaging" estimates and different ways of estimating risk. It will conclude with practical lessons learned from a project that required six months of testing, including how to present your estimates effectively and how to convince people of the value of keeping accurate time records.

Bullet poitns:
What is the value of good time estimation
Different methods of estimating a project
Two easy methods to estimate scientifically in a company that doesn't keep good time records
How to "massage" estimates with formulas that include buffer time
How to sell improving estimation in an organization


Description: The presentation discusses several different ways of calculating estimates ("best guess," experienced person's best guess, and actually using past numbers to calculate the effort size) and the advantages and disadvantages of each. Adaptable formulas that incorporate "buffer time" and risk are presented. Finally, a real-live example of a QA project with solid time measurement is presented, with an explanation of how team size was determined, how the work flow was determined, what the "actual hours" of testing looked like, what unexpected items affected the execution time, and how the project changed the company's "test time formula."

Outline:
1. What is estimation.
2. What is the value of estimation.
3. Why is it difficult to make good estimates.
4. What are the different methods of estimation, and which is best.
5. Which methods can be adapted to suit test planning.
6. Case study: project "Gemini."
7. How to correct course with bad estimates.
8. Test estimation "formulas" and how to use them.
9. The politics of scheduling (if time permits).
10. Q&A.

"Only weak men fear able women" - Marion Boyars