Outline Outline Review of PSP Levels Overview Resource planning - PDF document

Outline Outline Review of PSP Levels Overview Resource planning Planning IV: Planning IV: Estimating development and task time Combining multiple estimates Resource & Schedule Estimating Resource & Schedule Estimating Multiple regression Schedule estimating Earned value tracking Estimating accuracy Homework #5 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide Humphrey Ch. 6 - slide 1 1 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 2 Humphrey Ch. 6 - slide 2 Review of PSP Levels (Humphrey, 1995, p. 11) Overview (cf. Humphrey, 1995, p. 145) Review of PSP Levels (Humphrey, 1995, p. 11) Overview (cf. Humphrey, 1995, p. 145) This chapter covers: PSP3 Cyclic Cyclic development • How to make plans for small programs • How to combine these into larger PSP2 PSP2.1 Quality Mgt consolidated plans Code reviews Design templates Design reviews Schedule planning includes: PSP1.1 PSP1 • Resource loading Planning Task planning Size estimating Schedule planning Test report • Resource utilization PSP0.1 PSP0 Coding standard • Earned value tracking Size measurement Current process Baseline Process improvement Time recording proposal (PIP) Defect recording • … Defect type standard AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 3 Humphrey Ch. 6 - slide 3 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide Humphrey Ch. 6 - slide 4 4 Review of the PSP Project Review of the PSP Project Planning Framework Planning Framework Resource Planning (cf. Humphrey, 1995, p. 145-147) Resource Planning (cf. Humphrey, 1995, p. 145-147) (cf. Humphrey, 1995, p. 146) (cf. Humphrey, 1995, p. 146) In the PSP, the resource is your time. Customer NOTE: Real life is Define NOTE: Real life is Items Need Requirements not as linear as this not as linear as this Productivity framework suggests. framework suggests. Tasks Produce Conceptual • Hours required / unit of work Design • Each job has many unique conditions Historical Estimate Size Product Size and factors which affect productivity Customer Database Historical Estimate – See “cement” example, p. 148. Productivity Resources Database • Estimate productivity by calculating the Produce Resources Management Schedule Available average and range from prior jobs Delivered Develop Analyze the Tracking (homework assignments) Size, Resource, Product Product Process Reports Schedule Data AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 5 Humphrey Ch. 6 - slide 5 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 6 Humphrey Ch. 6 - slide 6 1

Development Time Development Time Estimating Task Time Estimating Task Time Planning Process ( Humphrey, 1995, p. 149) Planning Process ( Humphrey, 1995, p. 149) (cf. Humphrey, 1995, p. 145) (cf. Humphrey, 1995, p. 145) The SW development task is a special instance of Walk through Walk through general tasks for which time estimates must be diagram and diagram and LOC Size Estimate steps, p. 148-153. steps, p. 148-153. made. Obtain historical data • See Fig. 6.3, p. 156, and general task-estimation steps. Are there sufficient Estimating data for a regression For SW development we prefer to base our Choice C calculation? No Yes estimates on historical data. Calculate historical Are there sufficient estimate productivity in actual new Estimating Estimating data for a regression & changed LOC per hour calculation? Choice A We have three types of historical data which may Choice B No Yes Do the regression Do the regression be used: Calculate the calculation on actual calculation on estimated hours required object LOC object LOC and actual and actual hours hours • A: Estimated object LOC & total actual development hours Calculate the shortest and Calculate the time Calculate the time longest likely times required required • B: Actual object LOC & total actual development hours • C: Actual total new/changed LOC & total actual Calculate the prediction Calculate the prediction Time Estimate intervals interval development hours Time Estimate Time Estimate AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide Humphrey Ch. 6 - slide 7 7 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide Humphrey Ch. 6 - slide 8 8 Development Time Planning Development Time Planning Combining Multiple Estimates Combining Multiple Estimates Example ( Humphrey, 1995, p. 149) Example ( Humphrey, 1995, p. 149) (cf. Humphrey, 1995, p. 158-163) (cf. Humphrey, 1995, p. 158-163) Walk through example on p. 153-155 Assume 4 estimates: a, b, c, d. The estimated hours and standard deviations are: See how regression parameters are H a , H b , H c , H d and σ a , σ b , σ c , σ d • calculated and used. When estimates are independent (e.g. come from separate databases) and unbiased (not all from same project, under same manager, etc.): H t = total hourly time = Σ H i • σ t = total standard deviation = sqrt( Σ σ i ) • H upper = H t + σ t • H lower = H t - σ t • AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 9 Humphrey Ch. 6 - slide 9 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 10 Humphrey Ch. 6 - slide 10 Combining Multiple Estimates Combining Multiple Estimates Multiple Regression Multiple Regression Under Dependence (cf. Humphrey, 1995, p. 158-163) Under Dependence (cf. Humphrey, 1995, p. 158-163) (cf. Humphrey, 1995, p. 162-166) (cf. Humphrey, 1995, p. 162-166) Must use more involved calculation The problem: for the prediction interval when • We don’t have detailed enough data. • e.g. We have total hours, new LOC, reused estimates to be combined are not LOC, & modified LOC, but not hours by each of statistically independent these LOC categories. Use formulas on p. 160-162 The solution: • Multiple regression estimates the relative contributions. Example regression equation: • Hours t = β o + β 1 New k + β 2 Reuse k + β 3 Modified k AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 11 Humphrey Ch. 6 - slide 11 AU INSY 560, Winter 1997, Dan Turk AU INSY 560, Winter 1997, Dan Turk Humphrey Ch. 6 - slide 12 Humphrey Ch. 6 - slide 12 2