status report #5 · status report #5 (program data through january 1996) executive summary...

Quantitative Software Management, Inc.

PROJECT X Status Report #5

(Program Data Through January 1996)

Executive Summary (Assessment Yellow- Caution)

q The overall status of this project is “Yellow Light”.q So far this is a low to moderate risk situation.q There is a predicted slippage of 1.7 months. This has a potential impact on cost,

schedule, and reliability.q q The predicted Productivity Index is 5.9 vs. the 6.4 of the initial plan.

Major Observations: Upon review of the data covering performance through January1996 we offer the following observations (graphically portrayed in the “at a Glance”assessment). Detailed observations and comments are found in the body of this report.

1. The estimated code size at completion has grown from 40,270 SLOC initially up to44,760 in December and has now fallen back to 42,267 at the end of January (increaseof 1997 SLOC -- approximately 5% -- from the beginning).

2. All product construction metrics are below the expected values on the plannedproduction curves. New code increased during January; unit testing slowed down andintegration of existing code seemed to have stopped.

3. Application of people to this program fluctuates up and down due to shifting ofresources; on average the applied headcount is 20-25% lower than was originallyplanned. Assigned headcount for January was back to 40 people, but resource shiftingcontinues.

PROJECT X Status Report #5 January 1996

Quantitative Software Management, Inc. 2

Project Assessment at a GlanceDesign Units (Cum)

0

40

80

S 1 2 3 4 57S 1 2

DU

's

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Integrated Code (Cum)

0

20

40

60S 1 2 3 4 57S 1 2 IC

(thousands)

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Total Defect Rate

0

200

400

S 1 2 3 4 57S 1 2

Defects

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Unit Coded (Cum)

0

20

40

60S 1 2 3 4 57S 1 2 U

C (thousands)

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Size Est Variance (Rate)

0

20

40

60S 1 2 3 4 57S 1 2 S

EV

(thousands)

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Gantt Chart

C&T

RQ_D

S 1 2 3 4 57S 1 2

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Unit Tested (Cum)

0

20

40

60S 1 2 3 4 57S 1 2 U

T (thousands)

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Aggregate Staffing Rate

0

20

40

60S 1 2 3 4 57S 1 2

People

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Total Cum Effort

0

400

800

1200

S 1 2 3 4 57S 1 2

PM

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Current Plan Actual Interpolated Green Control Bound Yellow Control Bound Life Cycle includes RQ_D, C&TS = Start, 1 = PDR, 2 = Bld_1, 3 = CDR, 4 = Bld_2, 5 = TRR, 7 = Bld_3

Figure #1. Program metrics compared to plan and statistical control bounds.



Product Size Estimate:

The size estimate decreased from 44,760 last month to 42,267 at end of January. This isapproximately 5% larger than the original estimate. It is normal to see some fluctuation asthe design and code process progresses; after continued growth for the last three reportingperiods, the latest decrease may indicate that the size projection has stabilized. Size needsto be monitored closely.

Current Size Estimate vs. Original Size Estimate

Size Est Variance (Rate)

0

10

20

30

40

50

60S 1 2 3 4 5 7S 1 2

SE

V (thousands)

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Actual

Interpolated

Plan

Green CB

Yellow CB

S = Start

1 = PDR

2 = Bld_1

3 = CDR

4 = Bld_2

5 = TRR

7 = Bld_3

Size Est Variance(K)PI 6.4MBI 2.8

Date 2/14/96 (11.50 mos)

Plan Actual Diff40

Monthly Re-estimate of Size at Completion

Original Size Estimate

Figure #2. Current code size estimate at completion vs. original code size estimate.Note that estimate at completion has increased for 3 months and has decreased inthe last month. The current size estimate is approximately 5% larger than theoriginal size estimate.



Product Construction Performance:

All of the product construction metrics are below the planned completion.

Several factors could be causing this performance. 1) COMPANY A has beenunderstaffed (see report section on staffing & effort). 2) The COMPANY A plan is basedon a PI of 6.4. This is aggressive since the best that they have ever accomplished on pastefforts was a PI of 5.4. They are currently working at a PI of 5.9.

Product Construction Metrics(Plan vs. Actuals)

Some newcode thismonth; stillwell behind Unit testing

sloweddown; wellbelowexpected

Design Units (Cum)

0

20

40

60

80

100S 1 2 3 4 5 7S 1 2

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul


0

10

20

30

40

50

60S 1 2 3 4 5 7S 1 2

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Unit Coded (Cum)

0

10

20

30

40

50

60S 1 2 3 4 5 7S 1 2

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Unit Tested (Cum)

0

10

20

30

40

50

60S 1 2 3 4 5 7S 1 2

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Current Plan Actual Interpolated Green Control Bound Yellow Control BoundS = Start, 1 = PDR, 2 = Bld_1, 3 = CDR, 4 = Bld_2, 5 = TRR, 7 = Bld_3

Design unitspicked up in lastmonth

IntegratedCode is startingto fall behind

Unit tested sloweddown; well belowexpected

Figure #3. Product construction status vs. plan.



Staffing & Effort Performance:

The contractor had planned to be at a full time equivalent staffing of 39 people from thestart of the project. Because of resource sharing between programs at COMPANY A, thecontractor has only been able to staff at about 75% on average of what had been planned.

Staffing Metrics(Plan vs. Actuals)


0

10

20

30

40

50

60S 1 2 3 4 5 7S 1 2

People

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Actual

Interpolated

Plan

Green CB

Yellow CB

S = Start

1 = PDR

2 = Bld_1

3 = CDR

4 = Bld_2

5 = TRR

7 = Bld_3

Agg. StaffPI 6.4MBI 2.8

Date 2/14/96 (11.50 mos)

Plan Actual Diff39.00

Consistently understaffed frombeginning until thelast two months.Reflects headcount rather thanapplied personhours.

Figure #4. Staffing and effort plan vs. actual headcount. Not all people shown areworking full time on the program.



Effort Metrics(Plan vs. Actuals)

Total Cum Effort

0

200

400

600

800

1000

1200

1400S 1 2 3 4 5 7S 1 2

PM

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Actual

Interpolated

Plan

Green CB

Yellow CB

S = Start

1 = PDR

2 = Bld_1

3 = CDR

4 = Bld_2

5 = TRR

7 = Bld_3

Total Cum Effort (PM)PI 6.4MBI 2.8

Date 1/30/96 (10.98 mos)

Plan Actual Diff426.97 337.78 -89.19

Effort consistently underapplied. About 21% at thispoint in time.

Figure #5. Planned vs. actual effort.

Schedule Status:

Milestone 2 as shown above was completed on 2/3/96. This milestone slipped about 1month.

Significant Events:

Each month we ask the developer to provide a short narrative on any significant events(positive and negative) that will help us better understand the behavior of the metrics.These are the significant events reported by COMPANY A for January 1996.

1. The entire GN&T and Software Groups have been moved into a new closed area.This will allow desktop execution of classified all-software sIBUlations. Previouslythis had been done in different parts of the building with manual movement of parts ofthe program. Full transition to classified processing will be complete end of February.

2. There have been a number of problems associated with executing on target processorwith an IBU. All problems have been configuration issues with Company B’shardware and software. Since the configuration is still changing, Company B’ssoftware delivered earlier was found to be incompatible with hardware deliveries.Company B was at COMPANY A’s in early February to sort out these configurationissues and support COMPANY A’s integration goals.

3. There has been some size growth due to MIL STD 1760 requirements. This haspushed final unit testing into early February. Some people have been pulled off the



coding effort to work on getting OTF running on the target processor. A significantamount of time has been spent generating updates to all the software CDRLs fordelivery ahead of CDR.



Forecast Projections:

As of the end of December 1995 sufficient data were available to perform an adaptiveupdate forecast based on what has occurred so far. This allows us to forecast when theproject will be done, the associated cost in personmonths, and later, the related reliability(so far the actual reliability data is not sufficiently robust to determine much).

The following charts show the results of that forecast.

Adaptive Forecast vs. PlanSummary View

Design Units (Cum)

0

20

4060

80S 1 2 3 4 57S 1 2 3 4 57

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Unit Tested (Cum)

0

20

40

S 1 2 3 4 57S 1 2 3 4 57

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul


0

20

40

S 1 2 3 4 57S 1 2 3 4 57

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Unit Coded (Cum)

0

20

40

S 1 2 3 4 57S 1 2 3 4 57

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul


0

20

40

S 1 2 3 4 57S 1 2 3 4 57

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Total Cum Effort

0

400

800

1200S 1 2 3 4 57S 1 2 3 4 57

5 11 17 23 29 *Jan'95

Jul Jan'96

Jul Jan'97

Jul

Gantt Chart

C&T

RQ_D

S 1 2 3 4 5 7S 1 2 3 4 5 7

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Current Plan Actual Interpolated Current Forecast Life Cycle includes RQ_D, C&TS = Start, 1 = PDR, 2 = Bld_1, 3 = CDR, 4 = Bld_2, 5 = TRR, 7 = Bld_3

Figure #6. Summary of plan vs. actuals for the key metrics.



Planned vs. Actual Effort with Projection to Completion:

Planned vs. Projected EffortProjected Overrun isabout 65 PMTotal Cum Effort

0

200

400

600

800

1000

1200S 1 2 3 4 5 7S 1 2 3 4 5 7

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Actual

Interpolated

Forecast

Plan

S = Start

1 = PDR

2 = Bld_1

3 = CDR

4 = Bld_2

5 = TRR

7 = Bld_3

Total Cum Effort (PM)PI 6.4 5.9MBI 2.8 2.7

Date 6/11/97 (27.38 mos)

PlanActual/

ForecastEst. to

Complete1063.48 1010.33 64.96

Figure #7. Effort projection to completion.



Planned vs. Actual Code Production with Projection to Completion:

Integrated Code Projection

About 1.3 monthsslippage indicatedfor code to beintegrated(1/27/97)


0

10

20

30

40

50S 1 2 3 4 5 7S 1 2 3 4 5 7

IC (thousands)

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

Actual

InterpolatedForecast

Plan

S = Start

1 = PDR

2 = Bld_1 3 = CDR

4 = Bld_2

5 = TRR

7 = Bld_3

Integrated Code(K)PI 6.4 5.9MBI 2.8 2.7

Date 1/27/97 (22.86 mos)

PlanActual/

ForecastEst. to

Complete42.16 42.27 0.00

Figure #8. Integrated code production projection to completion.



Planned vs. Actual Schedule with Projection to Completion:

Schedule Projection

Projected 1.7month slip

Gantt Chart

C&T

RQ_D

S 1 2 3 4 5 7S 1 2 3 4 5 7

2 5 8 11 14 17 20 23 26 29 *Jan'95

Apr Jul Oct Jan'96

Apr Jul Oct Jan'97

Apr Jul

PlanActualForecast

S = Start 1 = PDR 2 = Bld_1 3 = CDR 4 = Bld_2 5 = TRR 7 = Bld_3

Elapsed MonthsPI 6.4 5.9MBI 2.8 2.7

Date 1/30/96 (10.98 mos)

PlanActual/

ForecastEst. to

Complete10.90 10.90 18.03

Figure #9. Schedule projection.



SLIM-Control Output:

The SLIM-Control output for the January 1996 analysis is attached.

status report #5 · status report #5 (program data through january 1996) executive summary...

Documents