FulFill Pro Smart Client User
Interface

Prepared for PROTOCOL
by XOCOMP

Prepared by Manuel Dennis III
Prepared week of 2/10/2003

System Size (lines of code)
Effort (staff months)

±11%
±197%

25
10.6

Schedule (calendar months)
Completion Date

4.9
4.9 months

±46%
±46%

1/1/2004
$320,202

Cost
Peak Staff (people)

±197%
±104%

3.5
4.6

Average Staff (people)
Overall Estimate Quality

2.3
Fair

This estimate is the 50/50 estimate--the estimate for which there is both a 50 percent chance of overrunning and a 50
percent chance of underrunning the estimate. This is also known as the "nominal" estimate. This estimate is for the "main
build" phase of a project, the time from detailed requirements specification complete to  software acceptance. Earlier
phases of a project are not estimated here.

Management Metric
Effort (staff months)
Schedule (calendar months)

11.3
$246,309

Cost
Peak Staff (people)

2.4
1.7

Average Staff (people)
These planning values meet the project's entire set of constraints and priorities to the maximum extent possible.

Estimates vary in the quality of the assumptions used to create them.
Some of these characteristics can be evaluated programmatically.
This report rates the quality of this estimate on a 5 point verbal scale:
     • Excellent

     • Very Good
     • Good
     • Fair

     • Poor
Overall quality of this estimate: Fair

Estimates calibrated with three or more historical projects are most accurate. Estimates calibrated with one or two
historical projects, cost drivers, or project types are less accurate. This estimate has been calibrated using project types.

Scope estimates created with fine-granularity units such as function points and lines of code include less imprecision than
estimates created with large-granularity units such as classes and subsystems. Scope estimates created on a module-by-

module basis are more accurate than entering only one basic scope estimate.
This estimate's type of scope estimate: Basic Size (Classes/Modules)
Scope estimate quality: Good

Estimates created later in a project are more accurate than estimates created early in the project. Even the best estimates
cannot be very accurate if they are created at a point in the project when comparatively little is known about the software to
be built.

Current project phase: High-Level Design Complete
Estimate quality possible in this phase: Very Good

The table below provides a consistency check by comparing the current project estimate to results from other projects of
similar sizes and types. The estimated project has a type of "Intranet Systems (internal)" and a subtype of "Business
Systems."

Management Metric
Productivity (lines of code per

staff-month)
Schedule (months)

Within Normal Range
Within Normal Range

Effort (staff-months)
Average Staff (people)

19
1.7

Code Generation Rate (lines
of code per month)

• Estimated size is greater than or equal to 5000 lines of code.
• Nominal development is expected to last at least 6 months.
• Nominal effort is expected to be at least 18 staff-months.

When less than two of these criteria are met, the only way to achieve a reliable estimate is to use Historical calibration.
Even when Historical calibration is used, some projects are too small to estimate reliably. This project uses Project Type

(from industry data) calibration, and four of these conditions have been met.
Suitability of Estimate to estimate this project: Excellent

Management Metric
Effort (staff months)
Schedule (calendar

Management Metric
Effort (staff months)
Schedule (calendar

months)
Cost

months)
Cost

Peak Staff (people)
Average Staff

3.3
2.3

Peak Staff (people)
Average Staff

2.4
1.7

Management Metric
Effort (staff months)
Schedule (calendar

Management Metric
Effort (staff months)
Schedule (calendar

months)
Cost

months)
Cost

Peak Staff (people)
Average Staff

Peak Staff (people)
Average Staff

0.9
0.6

The table below contains planning options. These options are computed from the nominal effort, schedule, and cost
produced by the Monte Carlo simulation. Each of these options is considered to be equally achievable in the configurations
shown. The schedule that best meets the constraints and priorities entered by the estimator (if any) is shown with a shaded

1
2

8.5
8.6

61
56

$781,744
$725,707

9.3
8.5

3
4

8.8
9.0

52
48

$673,687
$625,395

7.7
7.0

5
6

9.1
9.3

45
42

$580,566
$538,949

6.4
5.8

7
8

9.5
9.6

39
36

$500,316
$464,452

5.3
4.8

33
31

$431,159
$400,253

10
11

10.0
10.2

4.4
4.0

29
27

$371,562
$344,927

12
13

10.4
10.6

3.7
3.3

25
23

$320,202
$293,389

14
15
16

10.8
11.1
11.3

3.0
2.7
2.4

21
19

$268,820
$246,309

17
18

11.6
11.8

17
16

2.2
1.9

$225,683
$206,785

19
20

12.1
12.3

15
13

1.7
1.6

$189,469
$173,602

21
22

12.6
12.9

12
11

1.4
1.2

$159,065
$145,745

23
24

13.2
13.5

1.1
1.0

$133,540
$122,358

9
9

Estimates can be constrained to hold results within numeric values for cost, effort, schedule, and peak staff. The table
below shows the constraints used to create this estimate.

Constraint
Maximum Schedule Allowed

Value
12.0 months

Maximum Effort Allowed
Maximum Cost Allowed
Maximum Peak Staff Allowed

20 staff-months
n/a
5.0 staff

Minimum Mean Time to Defect
(MTTD) Allowed

Priorities can be used to determine the optimal project plan. The table below shows the priorities used to create this
estimate.

Priority
Schedule Priority
Effort Priority

Value
High Priority
Low Priority

Cost Priority
Peak Staff Priority

Medium Priority
Low Priority

Mean Time to Defect (MTTD)
Priority

Defect-related constraints and priorities are currently used only for documentation purposes and do not have any
quantitative effect on the estimates produced.

The table below contains project milestones for the optimal project plan. This estimate is for the "main build" phase of a
project, that is, the time between detailed requirements specification complete to  software acceptance. Dates are not
provided for milestones prior to the project start date.

Start of Feasibility Study
Feasibility Study / Product Concept

Complete
General Requirements Complete
Detailed Requirements / UI Design

Complete
High-Level Design Complete
Detailed-Design Complete
Feature Complete / Code Complete
Start of User-Oriented System Test

69
148
196

4/23/2003
7/11/2003
8/28/2003

20%
43%
57%

5%
24%
40%

(beta test)
Development and Test Complete
Software Accepted

320
344

93%
100%

90%
100%

~35 forms and dialogs 

~155 buttons which should directly map to 155 method calls

~40 drop down lists, ~25 appeared to be unique table lookups 

~15 grids with an average of 

~8 columns defined per grid 

~17 label areas

~225 textboxes which will require validation code, unique textboxes were code could be reused was not determined.

~2 tree views with multiple node levels 

~35 option buttons which may increase the number of methods due to specific case instances 

~45 check boxes which likely equate to conditional logic

The 35 forms can be broken down into 6 complex, ~21 simplex, and ~9 dialogs.

From these forms we derived the figure of 30 base classes with a possible additional class for each of the 25 unique drop down list to give us 55 class modules.

~10 major tabs were identified across the complex forms to indicate a higher order of complexity for the ~6 base classes.

No estimate can be presented on the component interface since we have user interface samples and database structures but no Business Logic Layer to review.

The table below contains a staffing profile for the optimal project plan. These staffing figures include development staff,
quality assurance staff, and first-level management.

Team Size (at
End of Month)

1
2

2/13/2003
3/15/2003

0.4
0.8

0
1

3
4

4/15/2003
5/15/2003

1.2
1.5

2
3

5
6

6/15/2003
7/15/2003

1.8
2.0

5
7

7
8

2.2
2.3

9/14/2003
10/14/2003
11/14/2003

11
13
16

2.4
2.4

10
11

The table below contains the cash flow for the optimal project plan. These costs include fully burdened labor costs for
development staff, quality assurance staff, and first-level management.  This is totally subjective because 

XOCOMP has no way of knowing the actual cost per PROTOCOL employee/developer.

1
2

2/13/2003
3/15/2003

$2,625.54
$7,795.72

$10,421.26
$23,148.60

3
4

4/15/2003
5/15/2003

$12,727.33
$17,274.83

$40,423.43
$61,735.36

5
6

6/15/2003
7/15/2003

$21,311.93
$24,737.37

7
8

$27,479.01
$29,495.90

$113,951.73
$143,447.64

9/14/2003
10/14/2003
11/14/2003

$30,778.58
$31,347.16

$174,226.22
$205,573.38

10
11

The table below contains scope estimates by probability. These scope estimates are expressed in lines of code. If the scope
estimates were not originally expressed by the estimator in lines of code they have been converted to lines of code. The
scope estimates are based on parameters that have been entered by the estimator, including the following:

Scoping Method: Basic Size (Classes/Modules)
Project Phase: High-Level Design Complete
Number of Simulations: 500

Than
41,483
46,550
49,716

1.0
5.0

-27%
-18%
-12%

10.0
20.0

51,616
54,150

-9%
-4%

30.0
40.0

55,416
56,683

50.0
60.0

0%
4%

59,216
61,116

70.0
80.0

62,383
64,283

10%
13%

90.0
95.0

66,816
70,616

18%
25%

The table below contains effort estimates by probability. These effort estimates are expressed in staff-months. They are
based on parameters that have been entered by the estimator, including the following:

Calibration Method: Project Type (from industry data)
Project Phase: High-Level Design Complete
Number of Simulations: 500

1.0
5.0

6
8

-75%
-66%

10.0
20.0

10
12

-61%
-50%

30.0
40.0

14
18

-42%
-27%

50.0
60.0

25
32

28%
76%

70.0
80.0

44
70

184%
9,590%

90.0
95.0

2,401
3,498

14,016%
19,726%

The table below contains cost estimates by probability. They are based on parameters that have been entered by the
estimator, including the following:  Once again, XOCOMP does not know the actual PROTOCOL employee/developer cost.

Calibration Method: Project Type (from industry data)
Project Phase: High-Level Design Complete
Number of Simulations: 500

Than
$80,346

1.0
5.0

-75%
-66%

$107,334
$125,465

10.0
20.0

-61%
-50%

$161,183
$185,901

30.0
40.0

-42%
-27%

$234,258
$320,202

50.0
60.0

$411,145
$564,969
$910,270

28%
76%

70.0
80.0

184%
9,590%

90.0
95.0

$31,027,029
$45,199,973

14,016%
19,726%

The table below contains schedule estimates by probability. These schedule estimates are expressed in calendar months.
They are based on parameters that have been entered by the estimator, including the following:

Calibration Method: Project Type (from industry data)
Project Phase: High-Level Design Complete
Number of Simulations: 500

1.0
5.0

6.9
7.5

-35%
-29%

10.0
20.0

8.0
8.6

-24%
-19%

30.0
40.0

9.2
9.8

50.0
60.0

10.6
11.6

10%
23%
44%

70.0
80.0

13.0
15.3

90.0
95.0

49.8
55.6

370%
425%

This estimate has been calibrated by choosing a description of the project type from a list of project types. The list of types
includes business systems, shrinkwrap products, avionics, and many other types. Estimate creates a project estimate by

using its database of productivity data for each project type. Because this data accounts for experience with software
development across entire industries, estimates created in this  way are subject to more variation than estimates created
using other calibration methods.

Project Type: Intranet Systems (internal)
Project Subtype: Business Systems

Expected Value
(50% Probability)

Standard
Deviation

Productivity Parameter (PP)
Productivity Index (PI)

n/a
2.2

Nominal Manpower Buildup Parameter (MBP)
Nominal Manpower Buildup Index (MBI)

1.0
-1.6

n/a
n/a

Minimum MBI Used in Creating Planning Options
Maximum MBI Used in Creating Planning Options

MBI Used to Create Optimum Plan
Lines of Code Used to Create Simulation

0.4
56,683

n/a
6,333

outcomes of
productivity, current project phase, and other parameters

entered by
of
different

estimator. It then estimates the likelihood
various project

assigns risk
complex

levels to
situations

Software
estimates created

planning
a

options.
uncertainty,

with the
more

aid
accurate

of automated
than

estimation
created

involve
Estimate
otherwise be impossible to model.

lot
to

of
create

this
meaningful

methodology allows
estimates

software
manual estimation

methods. Automated
software

estimation tools
be delivered

ultimately
lower cost than manual methods do.

Construx has
software

of
estimation

of the
model

industry)
virtually useless
environment in which it

that
unless it

generic
is calibrated

Estimate
tool which employs

project estimation
sophisticated modeling

for the specific
Estimate employs

a variety of
techniques to create accurate software project estimates.

will be used.
methods,

three
degrees of estimation accuracy.

• Project Types
The

Estimate
approaches.

accurate)
the type

method
being estimated

involves describing
by
list

of project
common project

• SLIM
SLIM

a list
includes

types. The
shrinkwrap

business
many other

was
early 1970s and first offered as a commercial product in
1978. The

products, avionics, and
creates
productivity

types. Estimate then
using

estimate
for

by
project

its
type.

database
Because

of
this

since its initial
Putnam
Excellence" (Yourdon Press, 1992).

offering and is fully described
co-authored

in a book
"Measures for

data accounts for experience with software development
across
calibration
estimates created using other calibration methods.

industries,
method

estimates
subject to

created
more variation

The
efficiently-run

is
projects

on
follow

software
can be

well-defined
of exponential

• Cost Drivers
A

patterns that
equations. These
approach to creating cost, effort, schedule, peak staffing,

modeled with
equations form

a set
the core

specific projects is to describe both the project type and
additional

and defect estimates.
• Cocomo 2.0

project
parameters include
the

project attributes
team will

(such as
be co-located

at the
geographically

Cocomo
Barry W. Boehm in the 1970s and described in his 1981

facility
facilities),
project is well-known or unprecedented), and personnel

spread
product

among
attributes

dispersed
whether

book,
Hall).

attributes
below
calibration method is used,
its database of industrywide productivity data.

(such
average,

whether
average,

the
or

team
When

is
this

exceptional).
Estimate again makes use of

Since 1981, additional work
Cocomo model and

has been done to refine the
apply it to projects other than the

U.S. Department
originally
extended into Cocomo 2.0, which allows estimates to be
created for

of Defense
developed.

projects for
present,

which it
model

was
has been

• Historical Projects
The most

accurate method
data

of calibrating
completed

enter
organization

projects
perform

within
project

set of cost drivers. Estimate
as a supplement
calibrated using cost

uses the Cocomo 2.0 model
to the SLIM

that
estimated.
estimating assumptions

will
approach

the
through

being
raft

when estimates are
drivers. A productivity baseline is

cuts
affect

that
personnel

productivity such
quality,

as
the

established
productivity factor
Cocomo 2.0 productivity.

the
is then

type
adjusted using

settings;
the computed

product
organization's
effectiveness,
other than historical data

existing
and

code
on.

organizational
calibration

When
are used, each

• Monte Carlo Simulation
Estimate

of these factors
estimator,

must
introduces
each

approximated
additional

by
imprecision

into
project

estimate at
data

complex interactions
assumptions.

in the face
Estimate

of uncertain estimating
simulates

organization to calibrate
projects, this approach provides accurate estimates with