Chapter 4 software project planning

Software Engineering (3rd ed.), By K.K Aggarwal & Yogesh Singh, Copyright © New Age International Publishers, 2007

1

Software Project Planning
After the finalization of SRS, we would like to
estimate size, cost and development time of the
project. Also, in many cases, customer may like to
know the cost and development time even prior to
finalization of the SRS.


2

In order to conduct a successful software project, we
must understand:
Scope of work to be done
The risk to be incurred
The resources required
The task to be accomplished
The cost to be expended
The schedule to be followed

3

Software planning begins before technical work starts, continues as
the software evolves from concept to reality, and culminates only
when the software is retired.
Size estimation

Cost estimation

Development time
Resources
requirements

Fig. 1: Activities during Software
Project Planning

Project
scheduling


4

Size Estimation
Lines of Code (LOC)
If LOC is simply a count of
the number of lines then
figure shown below contains
18 LOC .
When comments and blank
lines are ignored, the
program in figure 2 shown
below contains 17 LOC.

Fig. 2: Function for sorting an array
1.
2.

int. sort (int x[ ], int n)
{

3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.

int i, j, save, im1;
/*This function sorts array x in ascending order */
If (n<2) return 1;
for (i=2; i<=n; i++)
{
im1=i-1;
for (j=1; j<=im; j++)
if (x[i] < x[j])
{
Save = x[i];
x[i] = x[j];
x[j] = save;
}
}
return 0;
}


5

Growth of Lines of Code (LOC)
2,500,000

Total LOC ("wc -l") -- development releases
2,000,000

Total LOC ("wc -l") -- stable releases
Total LOC uncommented -- development releases

Total LOC

Total LOC uncommented -- stable releases
1,500,000

1,000,000

500,000

0
Jan 1993

Jun 1994

Oct 1995

Mar 1997

Jul 1998


Dec 1999

Apr 2001

6

Furthermore, if the main interest is the size of the program
for specific functionality, it may be reasonable to include
executable statements. The only executable statements in
figure shown above are in lines 5-17 leading to a count of
13. The differences in the counts are 18 to 17 to 13. One
can easily see the potential for major discrepancies for
large programs with many comments or programs written
in language that allow a large number of descriptive but
non-executable statement. Conte has defined lines of code
as:

7

“A line of code is any line of program text that is not a
comment or blank line, regardless of the number of
statements or fragments of statements on the line. This
specifically includes all lines containing program header,
declaration, and executable and non-executable
statements”.
This is the predominant definition for lines of code used
by researchers. By this definition, figure shown above
has 17 LOC.

8

Function Count

Alan Albrecht while working for IBM, recognized the
problem in size measurement in the 1970s, and
developed a technique (which he called Function Point
Analysis), which appeared to be a solution to the size
measurement problem.


9

The principle of Albrecht’s function point analysis (FPA)
is that a system is decomposed into functional units.
Inputs

:

information entering the system

Outputs

:

information leaving the system

Enquiries

:

requests for instant access to
information

Internal logical files

:

information held within the
system

External interface files

:

information held by other system
that is used by the system being
analyzed.


10

The FPA functional units are shown in figure given below:
User

Inquiries

Inputs

ILF

User
Outputs
System

Other
applications
EIF

ILF: Internal logical files
EIF: External interfaces

Fig. 3: FPAs functional units System

11

The five functional units are divided in two categories:
(i) Data function types
Internal Logical Files (ILF): A user identifiable group of
logical related data or control information maintained
within the system.
External Interface files (EIF): A user identifiable group of
logically related data or control information referenced by
the system, but maintained within another system. This
means that EIF counted for one system, may be an ILF in
another system.

12

(ii) Transactional function types
External Input (EI): An EI processes data or control information
that comes from outside the system. The EI is an elementary
process, which is the smallest unit of activity that is meaningful
to the end user in the business.
External Output (EO): An EO is an elementary process that
generate data or control information to be sent outside the
system.
External Inquiry (EQ): An EQ is an elementary process that is
made up to an input-output combination that results in data
retrieval.

13

Special features
Function point approach is independent of the language,
tools, or methodologies used for implementation; i.e. they
do not take into consideration programming languages,
data base management systems, processing hardware or
any other data base technology.
Function points can be estimated from requirement
specification or design specification, thus making it
possible to estimate development efforts in early phases of
development.

14

Function points are directly linked to the statement of
requirements; any change of requirements can easily
be followed by a re-estimate.
Function points are based on the system user’s
external view of the system, non-technical users of
the software system have a better understanding of
what function points are measuring.


15

Counting function points
Functional Units
External Inputs (EI)
External Output (EO)
External Inquiries (EQ)
External logical files (ILF)
External Interface files (EIF)

Weighting factors
Low
Average
High
3
4
6
4
5
7
3
4
6
7
10
15
5
7
10

Table 1 : Functional units with weighting factors

16

Table 2: UFP calculation table
Functional
Units
External
Inputs
(EIs)
External
Outputs
(EOs)
External
Inquiries
(EQs)
External
logical
Files (ILFs)
External
Interface
Files (EIFs)

Complexity
Totals

Count
Complexity
Low x 3
Average x 4
High x 6

=
=
=

Low x 4
Average x 5
High x 7

=
=
=

Low x 3
Average x 4
High x 6

=
=
=

Low x 7
Average x 10
High x 15

=
=
=

Low x 5
Average x 7
High x 10

Functional
Unit Totals

=
=
=

Total Unadjusted Function Point Count

17

The weighting factors are identified for all
functional units and multiplied with the functional
units accordingly. The procedure for the
calculation of Unadjusted Function Point (UFP) is
given in table shown above.


18

The procedure for the calculation of UFP in mathematical
form is given below:

5

3

UFP = ∑∑ Z ij wij
i =1 J =1
Where i indicate the row and j indicates the column of Table 1
Wij : It is the entry of the ith row and jth column of the table 1
Zij : It is the count of the number of functional units of Type i that
have been classified as having the complexity corresponding to
column j.

19

Organizations that use function point methods develop a criterion for
determining whether a particular entry is Low, Average or High.
Nonetheless, the determination of complexity is somewhat
subjective.
FP = UFP * CAF
Where CAF is complexity adjustment factor and is equal to [0.65 +
0.01 x ΣFi]. The Fi (i=1 to 14) are the degree of influence and are
based on responses to questions noted in table 3.


20

Table 3 : Computing function points.
Rate each factor on a scale of 0 to 5.
0
1
No
Incidental
Influence
Number of factors considered ( Fi )

2

3

4

5

Moderate

Average

Significant

Essential

1. Does the system require reliable backup and recovery ?
2. Is data communication required ?
3. Are there distributed processing functions ?
4. Is performance critical ?
5. Will the system run in an existing heavily utilized operational environment ?
6. Does the system require on line data entry ?
7. Does the on line data entry require the input transaction to be built over multiple screens or operations ?
8. Are the master files updated on line ?
9. Is the inputs, outputs, files, or inquiries complex ?
10. Is the internal processing complex ?
11. Is the code designed to be reusable ?
12. Are conversion and installation included in the design ?
13. Is the system designed for multiple installations in different organizations ?
14. Is the application designed to facilitate change and ease of use by the user ?

21

Functions points may compute the following important metrics:
Productivity

=

FP / persons-months

Quality

=

Defects / FP

Cost

=

Rupees / FP

Documentation

=

Pages of documentation per FP

These metrics are controversial and are not universally acceptable.
There are standards issued by the International Functions Point User
Group (IFPUG, covering the Albrecht method) and the United
Kingdom Function Point User Group (UFPGU, covering the MK11
method). An ISO standard for function point method is also being
developed.

22

Example: 4.1
Consider a project with the following functional units:
Number of user inputs

= 50

Number of user outputs

= 40

Number of user enquiries

= 35

Number of user files

= 06

Number of external interfaces

= 04

Assume all complexity adjustment factors and weighting factors are
average. Compute the function points for the project.

23

Solution
We know

5

3

i =1 J =1

UFP = 50 x 4 + 40 x 5 + 35 x 4 + 6 x 10 + 4 x 7
= 200 + 200 + 140 + 60 + 28 = 628
CAF = (0.65 + 0.01 ΣFi)
= (0.65 + 0.01 (14 x 3)) = 0.65 + 0.42 = 1.07
FP = UFP x CAF
= 628 x 1.07 = 672

24

Example:4.2
An application has the following:
10 low external inputs, 12 high external outputs, 20 low
internal logical files, 15 high external interface files, 12
average external inquiries, and a value of complexity
adjustment factor of 1.10.
What are the unadjusted and adjusted function point counts ?


25

Solution
Unadjusted function point counts may be calculated using
as:
5

3

i =1 J =1

FP

= 10 x 3 + 12 x 7 + 20 x 7 + 15 + 10 + 12 x 4
= 30 + 84 +140 + 150 + 48
= 452
= UFP x CAF
= 452 x 1.10 = 497.2.

26

Example: 4.3
Consider a project with the following parameters.
(i) External Inputs:
(a) 10 with low complexity
(b)15 with average complexity
(c) 17 with high complexity
(ii) External Outputs:
(b)13 with high complexity
(iii) External Inquiries:
(b) 4 with average complexity
(c) 2 high complexity

27

(iv) Internal logical files:
(a) 2 with average complexity
(b)1 with high complexity
(v) External Interface files:
In addition to above, system requires
i. Significant data communication
ii. Performance is very critical
iii. Designed code may be moderately reusable
iv. System is not designed for multiple installation in different
organizations.
Other complexity adjustment factors are treated as average. Compute
the function points for the project.

28

Solution: Unadjusted function points may be counted using table 2
Functional
Units
External
Inputs
(EIs)
External
Outputs
(EOs)
External
Inquiries
(EQs)
External
logical
Files (ILFs)
External
Interface
Files (EIFs)

Complexity
Totals

Count

Complexity

10

Low x 3
Average x 4
High x 6

=
=
=

Low x 4
Average x 5
High x 7

=
=
=

24

Low x 3
Average x 4
High x 6

=
=
=

9

Low x 7
Average x 10
High x 15

=
=
=

Low x 5
Average x 7
High x 10

=
=
=

Functional
Unit Totals

15
17
6
0
13
3
4
2
0
2
1
9
0
0

30
60
102

192

0
91

115

16
12

37

0
20
15

35

45
0
0

Total Unadjusted Function Point Count

45
424

29

14

∑ F = 3+4+3+5+3+3+3+3+3+3+2+3+0+3=41
i

i =1

CAF = (0.65 + 0.01 x ΣFi)
= (0.65 + 0.01 x 41)
= 1.06
FP = UFP x CAF
= 424 x 1.06
= 449.44

Hence

FP = 449

30

Relative Cost of Software Phases


31


Relative Cost to detect and correct fault

Cost to Detect and Fix Faults
200
180
160
140
120
100
80
60
40
20
0

Cost

Req

Des

I nt


32

Cost Estimation
A number of estimation techniques have been developed and are
having following attributes in common :
Project scope must be established in advance
Software metrics are used as a basis from which estimates are made
The project is broken into small pieces which are estimated individually

To achieve reliable cost and schedule estimates, a number of options
arise:
Delay estimation until late in project
Use simple decomposition techniques to generate project cost and
schedule estimates
Develop empirical models for estimation
Acquire one or more automated estimation tools

33

MODELS

Static, Single
Variable
Models

Static,
Multivariable
Models


34

Static, Single Variable Models
Methods using this model use an equation to estimate the desired
values such as cost, time, effort, etc. They all depend on the same
variable used as predictor (say, size). An example of the most
common equations is :

C = a Lb

(i)

C is the cost, L is the size and a,b are constants

E
= 1.4 L0.93
DOC = 30.4 L0.90
D
= 4.6 L0.26
Effort (E in Person-months), documentation (DOC, in number of
pages) and duration (D, in months) are calculated from the number
of lines of code (L, in thousands of lines) used as a predictor.

35

Static, Multivariable Models
These models are often based on equation (i), they actually depend
on several variables representing various aspects of the software
development environment, for example method used, user
participation, customer oriented changes, memory constraints, etc.

E

= 5.2 L0.91

D

= 4.1 L0.36

The productivity index uses 29 variables which are found to be
highly correlated to productivity as follows:
29

Ι = ∑ Wi X i
i =1

36

Example: 4.4
Compare the Walston-Felix model with the SEL model on a
software development expected to involve 8 person-years of effort.
(a)Calculate the number of lines of source code that can be
produced.
(b)Calculate the duration of the development.
(c)Calculate the productivity in LOC/PY
(d)Calculate the average manning


37

Solution
The amount of manpower involved = 8 PY = 96 person-months
(a) Number of lines of source code can be obtained by reversing
equation to give:
L = (E/a)1/b
Then
L(SEL) = (96/1.4)1/0.93 = 94264 LOC
L(SEL) = (96/5.2)1/0.91 = 24632 LOC.


38

(b) Duration in months can be calculated by means of equation
D(SEL) = 4.6 (L)0.26
= 4.6 (94.264)0.26 = 15 months
D(W-F) = 4.1 L0.36
= 4.1(24.632)0.36 = 13 months
(c) Productivity is the lines of code produced per person/month (year)

94264
P( SEL) =
= 11783 LOC / Person − Years
8
24632
P(W − F ) =
= 3079 LOC / Person − Years
8

39

(d) Average manning is the average number of persons required per
month in the project.

96 P − M
M ( SEL ) =
= 6.4 Persons
15 M
96 P − M
M (W − F ) =
= 7.4 Persons
13 M


40

The Constructive Cost Model (COCOMO)
Constructive Cost model
(COCOMO)

Basic

Intermediate

Detailed

Model proposed by
B. W. Boehm’s
through his book
Software Engineering Economics in 1981

41

COCOMO applied to

Organic
mode

Semidetached
mode

Embedded
mode


42

Mode

Organic

Project size

Typically
2-50 KLOC

Semi
detached

Typically
50-300 KLOC

Embedded Typically over
300 KLOC

Nature of Project

Innovation

Deadline of
the project

Small size project, experienced
developers in the familiar
environment. For example, pay
roll, inventory projects etc.

Little

Not tight

Medium size project, Medium
size team, Average previous
experience on similar project.
For example: Utility systems
like
compilers,
database
systems, editors etc.

Medium

Medium

Large project, Real time
systems, Complex interfaces,
Very little previous experience.
For example: ATMs, Air Traffic
Control etc.

Significant

Tight

Development
Environment
Familiar & In
house

Medium

Complex
Hardware/
customer
Interfaces
required

Table 4: The comparison of three COCOMO modes

43

Basic Model
Basic COCOMO model takes the form

E = ab ( KLOC )

D = cb ( E )

bb

db

where E is effort applied in Person-Months, and D is the
development time in months. The coefficients ab, bb, cb and db are
given in table 4 (a).

44

ab

bb

cb

db

Organic

2.4

1.05

2.5

0.38

Semidetached

3.0

1.12

2.5

0.35

Embedded

3.6

1.20

2.5

0.32

Software
Project

Table 4(a): Basic COCOMO coefficients


45

When effort and development time are known, the average staff size
to complete the project may be calculated as:

E
Average staff size ( SS ) =
Persons
D
When project size is known, the productivity level may be
calculated as:

KLOC
Productivity ( P ) =
KLOC / PM
E


46

Example: 4.5

Suppose that a project was estimated to be 400 KLOC.
Calculate the effort and development time for each of the three
modes i.e., organic, semidetached and embedded.


47

Solution
The basic COCOMO equation take the form:

E = ab ( KLOC ) bb
D = cb ( KLOC )

db

Estimated size of the project = 400 KLOC
(i) Organic mode
E = 2.4(400)1.05 = 1295.31 PM
D = 2.5(1295.31)0.38 = 38.07 PM


48

(ii) Semidetached mode
E = 3.0(400)1.12 = 2462.79 PM
D = 2.5(2462.79)0.35 = 38.45 PM
(iii) Embedded mode
E = 3.6(400)1.20 = 4772.81 PM
D = 2.5(4772.8)0.32 = 38 PM


49

Example: 4.6
A project size of 200 KLOC is to be developed. Software
development team has average experience on similar type of
projects. The project schedule is not very tight. Calculate the effort,
development time, average staff size and productivity of the project.


50

Solution
The semi-detached mode is the most appropriate mode; keeping in
view the size, schedule and experience of the development team.
Hence

E = 3.0(200)1.12 = 1133.12 PM
D = 2.5(1133.12)0.35 = 29.3 PM

E
Average staff size ( SS ) =
Persons
D
1133.12
=
= 38.67 Persons
29.3

51

KLOC
200
Productivity =
=
= 0.1765 KLOC / PM
E
1133.12

P = 176 LOC / PM


52

Intermediate Model
Cost drivers
(i) Product Attributes
Required s/w reliability
Size of application database
Complexity of the product
(ii) Hardware Attributes
Run time performance constraints
Memory constraints
Virtual machine volatility
Turnaround time

53

(iii) Personal Attributes
Analyst capability
Programmer capability
Application experience
Virtual m/c experience
Programming language experience
(iv) Project Attributes
Modern programming practices
Use of software tools
Required development Schedule

54

Multipliers of different cost drivers
Cost Drivers

RATINGS
Very low

Low

Nominal

High

Very
high

Extra
high

0.75

0.88

1.00

1.15

1.40

--

DATA

--

0.94

1.00

1.08

1.16

--

CPLX

0.70

0.85

1.00

1.15

1.30

1.65

TIME

--

--

1.00

1.11

1.30

1.66

STOR

--

--

1.00

1.06

1.21

1.56

VIRT

--

0.87

1.00

1.15

1.30

--

TURN

--

0.87

1.00

1.07

1.15

--

Product Attributes
RELY

Computer Attributes


55

Cost Drivers

RATINGS
Very low

Low

Nominal

High

Very
high

Extra
high

1.46

1.19

1.00

0.86

0.71

--

1.29

1.13

1.00

0.91

0.82

--

1.42

1.17

1.00

0.86

0.70

--

1.21

1.10

1.00

0.90

--

--

1.14

1.07

1.00

0.95

--

--

1.24

1.10

1.00

0.91

0.82

--

1.24

1.10

1.00

0.91

0.83

--

1.23

1.08

1.00

1.04

1.10

--

Personnel Attributes
ACAP
AEXP
PCAP
VEXP
LEXP
Project Attributes
MODP
TOOL
SCED

Table 5: Multiplier values for effort calculations

56

Intermediate COCOMO equations

E = ai ( KLOC ) bi * EAF
D = ci ( E ) d i
Project

ai

bi

ci

di

Organic

3.2

1.05

2.5

0.38

Semidetached

3.0

1.12

2.5

0.35

Embedded

2.8

1.20

2.5

0.32

Table 6: Coefficients for intermediate COCOMO

57

Detailed COCOMO Model
Detailed COCOMO

Phase-Sensitive
effort multipliers
Cost
drivers

Three level product
hierarchy
Modules subsystem

design
& test

System level

Manpower allocation for
each phase

58

Development Phase
Plan / Requirements
EFFORT

:

DEVELOPMENT TIME :

6% to 8%
10% to 40%

% depend on mode & size


59

Design
Effort
Time

:
:

16% to 18%
19% to 38%

Programming
Effort
Time

:
:

48% to 68%
24% to 64%

Integration & Test
Effort
Time

:
:

16% to 34%
18% to 34%


60

Principle of the effort estimate
Size equivalent
As the software might be partly developed from software already
existing (that is, re-usable code), a full development is not always
required. In such cases, the parts of design document (DD%), code
(C%) and integration (I%) to be modified are estimated. Then, an
adjustment factor, A, is calculated by means of the following
equation.
A = 0.4 DD + 0.3 C + 0.3 I
The size equivalent is obtained by
S (equivalent) = (S x A) / 100

Ep = µ pE
Dp = τ p D

61

Lifecycle Phase Values of
Mode & Code
Size

µp

Plan &
Requirements

System
Design

Detailed
Design

Module
Code & Test

Integration
& Test

Organic Small
S≈2

0.06

0.16

0.26

0.42

0.16

Organic
medium S≈32

0.06

0.16

0.24

0.38

0.22

Semidetached
medium S≈32

0.07

0.17

0.25

0.33

0.25

Semidetached
large S≈128

0.07

0.17

0.24

0.31

0.28

Embedded
large S≈128

0.08

0.18

0.25

0.26

0.31

Embedded
extra large
S≈320

0.08

0.18

0.24

0.24

0.34

Table 7 : Effort and schedule fractions occurring in each phase of the lifecycle

62

Lifecycle Phase Values of
Mode & Code
Size

τp

Plan &
Requirements

System
Design

Detailed
Design

Module Code
& Test

Integration
& Test

Organic Small
S≈2

0.10

0.19

0.24

0.39

0.18

Organic
medium S≈32

0.12

0.19

0.21

0.34

0.26

Semidetached
medium S≈32

0.20

0.26

0.21

0.27

0.26

Semidetached
large S≈128

0.22

0.27

0.19

0.25

0.29

Embedded
large S≈128

0.36

0.36

0.18

0.18

0.28

Embedded
extra large
S≈320

0.40

0.38

0.16

0.16

0.30

Table 7 : Effort and schedule fractions occurring in each phase of the lifecycle

63

Distribution of software life cycle:
1.

Requirement and product design
(a) Plans and requirements
(b)System design

2.

Detailed Design
(a) Detailed design

3.

Code & Unit test
(a) Module code & test

4.

Integrate and Test
(a) Integrate & Test

64

Example: 4.7
A new project with estimated 400 KLOC embedded system has to be
developed. Project manager has a choice of hiring from two pools of
developers: Very highly capable with very little experience in the
programming language being used
Or
Developers of low quality but a lot of experience with the programming
language. What is the impact of hiring all developers from one or the
other pool ?


65

Solution
This is the case of embedded mode and model is intermediate
COCOMO.

E = ai ( KLOC )

Hence

di

= 2.8 (400)1.20 = 3712 PM
Case I: Developers are very highly capable with very little experience
in the programming being used.
EAF

= 0.82 x 1.14 = 0.9348

E

= 3712 x .9348 = 3470 PM

D

= 2.5 (3470)0.32 = 33.9 M

66

Case II: Developers are of low quality but lot of experience with the
programming language being used.
EAF

= 1.29 x 0.95 = 1.22

E

= 3712 x 1.22 = 4528 PM

D

= 2.5 (4528)0.32 = 36.9 M

Case II requires more effort and time. Hence, low quality developers
with lot of programming language experience could not match with
the performance of very highly capable developers with very litter
experience.


67

Example: 4.8
Consider a project to develop a full screen editor. The major components
identified are:
I. Screen edit
II. Command Language Interpreter
III. File Input & Output
IV. Cursor Movement
V. Screen Movement
The size of these are estimated to be 4k, 2k, 1k, 2k and 3k delivered source
code lines. Use COCOMO to determine
1. Overall cost and schedule estimates (assume values for different
cost drivers, with at least three of them being different from 1.0)
2. Cost & Schedule estimates for different phases.

68

Solution
Size of five modules are:
Screen edit

= 4 KLOC

Command language interpreter

= 2 KLOC

File input and output

= 1 KLOC

Cursor movement

= 2 KLOC

Screen movement

= 3 KLOC

Total

= 12 KLOC


69

Let us assume that significant cost drivers are
i.

Required software reliability is high, i.e.,1.15

ii. Product complexity is high, i.e.,1.15
iii. Analyst capability is high, i.e.,0.86
iv. Programming language experience is low,i.e.,1.07
v. All other drivers are nominal
EAF = 1.15x1.15x0.86x1.07 = 1.2169


70

(a) The initial effort estimate for the project is obtained from the
following equation
E = ai (KLOC)bi x EAF
= 3.2(12)1.05 x 1.2169 = 52.91 PM
Development time

D = Ci(E)di
= 2.5(52.91)0.38 = 11.29 M

(b) Using the following equations and referring Table 7, phase wise
cost and schedule estimates can be calculated.

Ep = µ pE
Dp = τ p D

71

Since size is only 12 KLOC, it is an organic small model. Phase wise
effort distribution is given below:
System Design

= 0.16 x 52.91 = 8.465 PM

Detailed Design

= 0.26 x 52.91 = 13.756 PM

Module Code & Test

= 0.42 x 52.91 = 22.222 PM

Integration & Test

= 0.16 x 52.91 = 8.465 Pm

Now Phase wise development time duration is
System Design

= 0.19 x 11.29 = 2.145 M

Detailed Design

= 0.24 x 11.29 = 2.709 M

Module Code & Test

= 0.39 x 11.29 = 4.403 M

Integration & Test

= 0.18 x 11.29 = 2.032 M


72

COCOMO-II
The following categories of applications / projects are identified by
COCOMO-II and are shown in fig. 4 shown below:
Application
generators &
composition aids
End user
programming

Application
composition

Infrastructure

System
integration

Fig. 4 : Categories of applications / projects

73

Stage
No

Model Name

Application for the
types of projects

Applications

Stage I

Application composition
estimation model

Application composition

In
addition
to
application
composition type of projects, this
model is also used for prototyping
(if any) stage of application
generators, infrastructure & system
integration.

Stage II

Early design estimation Application generators,
model
infrastructure & system
integration

Used in early design stage of a
project, when less is known about
the project.

Stage III Post
architecture
estimation model

Application generators,
infrastructure & system
integration

Used after the completion of the
detailed architecture of the project.

Table 8: Stages of COCOMO-II

74

Application Composition Estimation Model

Fig.5: Steps for the estimation of effort in person months

75

i. Assess object counts: Estimate the number of screens, reports and
3 GL components that will comprise this application.

ii. Classification of complexity levels: We have to classify each
object instance into simple, medium and difficult complexity levels
depending on values of its characteristics.

Table 9 (a): For screens

76


Table 9 (b): For reports


77

iii. Assign complexity weight to each object : The weights are used
for three object types i.e., screen, report and 3GL components using
the Table 10.

Table 10: Complexity weights for each level


78

iv. Determine object points: Add all the weighted object instances to
get one number and this known as object-point count.

v. Compute new object points: We have to estimate the percentage
of reuse to be achieved in a project. Depending on the percentage
reuse, the new object points (NOP) are computed.
(object points) * (100-%reuse)
NOP = ------------------------------------------100
NOP are the object points that will need to be developed and differ from
the object point count because there may be reuse.


79

vi. Calculation of productivity rate: The productivity rate can be
calculated as:
Productivity rate (PROD) = NOP/Person month

Table 11: Productivity values

80

vii. Compute the effort in Persons-Months: When PROD is known,
we may estimate effort in Person-Months as:
NOP
Effort in PM = -----------PROD


81

Example: 4.9
Consider a database application project with the following characteristics:
I. The application has 4 screens with 4 views each and 7 data tables
for 3 servers and 4 clients.
II. The application may generate two report of 6 sections each from 07
data tables for two server and 3 clients. There is 10% reuse of
object points.
The developer’s experience and capability in the similar environment is
low. The maturity of organization in terms of capability is also low.
Calculate the object point count, New object points and effort to develop
such a project.


82

Solution
This project comes under the category of application composition
estimation model.
Number of screens = 4 with 4 views each
Number of reports = 2 with 6 sections each
From Table 9 we know that each screen will be of medium
complexity and each report will be difficult complexity.
Using Table 10 of complexity weights, we may calculate object point
count.
= 4 x 2 + 2 x 8 = 24
24 * (100 -10)
NOP = -------------------- = 21.6
100

83

Table 11 gives the low value of productivity (PROD) i.e. 7.
NOP
Efforts in PM = ----------PROD
21.6
Efforts = ----------- = 3.086 PM
7


84

The Early Design Model
The COCOMO-II models use the base equation of the form

PMnominal = A * (size)B
where
PMnominal = Effort of the project in person months
A = Constant representing the nominal productivity, provisionally set to 2.5
B = Scale factor
Size = Software size


85

Scale factor

Explanation

Remarks

Precedentness

Reflects
the
previous
experience
on
similar
projects. This is applicable to
individuals & organization
both in terms of expertise &
experience

Very
low
means
no
previous
experiences, Extra high means that
organization is completely familiar with
this application domain.

Development flexibility

Reflect the degree of flexibility
in the development process.

Very low means a well defined process
is used. Extra high means that the client
gives only general goals.

Reflect the degree of risk
analysis carried out.

Very low means very little analysis and
Extra high means complete and through
risk analysis.

Architecture/
resolution

Risk

Cont…
Table 12: Scaling factors required for the calculation of the value of B

86

Scale factor

Reflects
the
management skills.

Remarks
team

Very
low
means
no
previous
experiences, Extra high means that
organization is completely familiar with
this application domain.

Reflects the process maturity
of the organization. Thus it is
dependent on SEI-CMM level
of the organization.

Team cohesion

Process maturity

Explanation

Very low means organization has no
level at all and extra high means
organization is related as highest level
of SEI-CMM.

Table 12: Scaling factors required for the calculation of the value of B


87

Scaling factors

Very
low

Low

Nominal

High

Very
high

Extra
high

Precedent ness

6.20

4.96

3.72

2.48

1.24

0.00

Development
flexibility

5.07

4.05

3.04

2.03

1.01

0.00

Architecture/ Risk
resolution

7.07

5.65

4.24

2.83

1.41

0.00

Team cohesion

5.48

4.38

3.29

2.19

1.10

0.00

Process maturity

7.80

6.24

4.68

3.12

1.56

0.00

Table 13: Data for the Computation of B

The value of B can be calculated as:
B=0.91 + 0.01 * (Sum of rating on scaling factors for the project)

88

Early design cost drivers
There are seven early design cost drivers and are given below:
i.

Product Reliability and Complexity (RCPX)

ii. Required Reuse (RUSE)
iii. Platform Difficulty (PDIF)
iv. Personnel Capability (PERS)
v. Personnel Experience (PREX)
vi. Facilities (FCIL)
vii. Schedule (SCED)

89

Post architecture cost drivers
There are 17 cost drivers in the Post Architecture model. These are rated
on a scale of 1 to 6 as given below :

The list of seventeen cost drivers is given below :
i.

Reliability Required (RELY)

ii. Database Size (DATA)
iii. Product Complexity (CPLX)
iv. Required Reusability (RUSE)

90

v. Documentation (DOCU)
vi. Execution Time Constraint (TIME)
vii. Main Storage Constraint (STOR)
viii.Platform Volatility (PVOL)
ix. Analyst Capability (ACAP)
x. Programmers Capability (PCAP)
xi. Personnel Continuity (PCON)
xii. Analyst Experience (AEXP)

91

xiii. Programmer Experience (PEXP)
xiv. Language & Tool Experience (LTEX)
xv. Use of Software Tools (TOOL)
xvi. Site Locations & Communication Technology between Sites (SITE)
xvii. Schedule (SCED)


92

Mapping of early design cost drivers and post architecture cost
drivers
The 17 Post Architecture Cost Drivers are mapped to 7 Early Design Cost
Drivers and are given in Table 14
Early Design Cost Drivers

Counter part Combined Post
Architecture Cost drivers

RCPX

RELY, DATA, CPLX, DOCU

RUSE

RUSE

PDIF

TIME, STOR, PVOL

PERS

ACAP, PCAP, PCON

PREX

AEXP, PEXP, LTEX

FCIL

TOOL, SITE

SCED

SCED
Table 14: Mapping table


93

Product of cost drivers for early design model
i. Product Reliability and Complexity (RCPX): The cost driver combines
four Post Architecture cost drivers which are RELY, DATA, CPLX and
DOCU.


94

ii. Required Reuse (RUSE) : This early design model cost driver is same as
its Post architecture Counterpart. The RUSE rating levels are (as per
Table 16):


95

iii. Platform Difficulty (PDIF) : This cost driver combines TIME, STOR
and PVOL of Post Architecture Cost Drivers.


96

iv. Personnel Capability (PERS) : This cost driver combines three Post
Architecture Cost Drivers. These drivers are ACAP, PCAP and PCON.


97

v. Personnel Experience (PREX) : This early design driver combines three
Post Architecture Cost Drivers, which are AEXP, PEXP and LTEX.


98

vi. Facilities (FCIL): This depends on two Post Architecture Cost Drivers,
which are TOOL and SITE.


99

vii.Schedule (SCED) : This early design cost driver is the same as Post
Architecture Counterpart and rating level are given below using table
16.


100

The seven early design cost drivers have been converted into numeric
values with a Nominal value 1.0. These values are used for the calculation
of a factor called “Effort multiplier” which is the product of all seven early
design cost drivers. The numeric values are given in Table 15.

Table 15: Early design parameters

101

The early design model adjusts the nominal effort using 7 effort multipliers
(EMs). Each effort multiplier (also called drivers) has 7 possible weights as
given in Table 15. These factors are used for the calculation of adjusted
effort as given below:

PM adjusted

 7

= PM nominal × ∏ EM i 
 i =7


PMadjusted effort may very even up to 400% from PMnominal
Hence PMadjusted is the fine tuned value of effort in the early design phase


102

Example: 4.10
A software project of application generator category with estimated 50
KLOC has to be developed. The scale factor (B) has low
precedentness, high development flexibility and low team cohesion.
Other factors are nominal. The early design cost drivers like platform
difficult (PDIF) and Personnel Capability (PERS) are high and others
are nominal. Calculate the effort in person months for the
development of the project.


103

Solution
Here B = 0.91 + 0.01 * (Sum of rating on scaling factors for the project)
= 0.91 + 0.01 * (4.96 + 2.03 + 4.24 + 4.38 + 4.68)
= 0.91 + 0.01(20.29)=1.1129
PMnominal = A*(size)B
= 2.5 * (50)1.1129 = 194.41 Person months
The 7 cost drivers are
PDIF = high (1.29)
PERS = high (0.83)
RCPX = nominal (1.0)
RUSE = nominal (1.0)
PREX = nominal (1.0)
FCIL = nominal (1.0)
SCEO = nominal (1.0)

104


= 194.41 * [1.29 x 0.83)
= 194.41 x 1.07
= 208.155 Person months


105

Post Architecture Model
The post architecture model is the most detailed estimation model and is
intended to be used when a software life cycle architecture has been
completed. This model is used in the development and maintenance of
software products in the application generators, system integration or
infrastructure sectors.

PM adjusted

 17

 i =7


EM : Effort multiplier which is the product of 17 cost drivers.
The 17 cost drivers of the Post Architecture model are described in the
table 16.

106


Table 16: Post Architecture Cost Driver rating level summary

Cont…

107



Cont…

108



Cont…

109



110

Product complexity is based on control operations, computational
operations, device dependent operations, data management operations and
user interface management operations. Module complexity rating are given
in table 17.
The numeric values of these 17 cost drivers are given in table 18 for the
calculation of the product of efforts i.e., effort multiplier (EM). Hence PM
adjusted is calculated which will be a better and fine tuned value of effort
in person months.


111

Control
Operations

Computational
Operations

Devicedependent
Operations

Data management
Operations

User Interface
Management
Operations

Very
Low

Straight-line code
with a few nonnested structured
programming
operators: Dos.
Simple module
composition via
procedure calls or
simple scripts.

Evaluation of
simple
expressions: e.g.,
A=B+C*(D-E)

Simple read,
write statements
with simple
formats.

Simple arrays in
main memory.
Simple COTSDB
queries, updates.

Simple input
forms, report
generators.

Low

Straight forward
nesting of
structured
programming
operators. Mostly
simple predicates

Evaluation of
moderate-level
expressions: e.g.,
D=SQRT(B**24*A*C)

No cognizance
needed of
particular
processor or I/O
device
characteristics.
I/O done at
GET/PUT level.

Single file sub
setting with no data
structure changes,
no edits, no
intermediate files,
Moderately
complex COTS-DB
queries, updates.

User of simple
graphics user
interface (GUI)
builders.

Table 17: Module complexity ratings

Cont…

112

Control Operations

Computational
Operations

Devicedependent
Operations

Data
management
Operations

User Interface
Management
Operations

Nominal

Mostly simple nesting.
Some inter module
control Decision tables.
Simple callbacks or
message passing,
including middleware
supported distributed
processing.

Use of standard
maths and
statistical
routines. Basic
matrix/ vector
operations.

I/O processing
includes
device
selection,
status
checking and
error
processing.

Multi-file input
and single file
output. Simple
structural
changes, simple
edits. Complex
COTS-DB
queries,
updates.

Simple use of
widget set.

High

Highly nested
structured
programming operators
with many compound
predicates. Queue and
stack control.
Homogeneous,
distributed processing.
Single processor soft
real time control.

Basic numerical
analysis:
multivariate
interpolation,
ordinary
differential
equations. Basic
truncation, round
off concerns.

Operations at
physical I/O
level (physical
storage
address
translations;
seeks, read
etc.)
Optimized I/O
overlap.

Simple triggers
activated by data
stream contents.
Complex data
restructuring.

Widget set
development
and
extension.
Simple voice
I/O
multimedia.


Cont…

113

Control Operations

Computational
Operations

Device-dependent
Operations

Data
management
Operations

User
Interface
Management
Operations

Very
High

Reentrant and
recursive coding.
Fixed-priority
interrupt handling.
Task
synchronization,
complex callbacks,
heterogeneous
distributed
processing. Single
processor hard real
time control.

Difficult but
structured
numerical analysis:
near singular
matrix equations,
partial differential
equations. Simple
parallelization.

Routines for
interrupt
diagnosis,
servicing,
masking.
Communication
line handling.
Performance
intensive
embedded
systems.

Distributed
database
coordination.
Complex
triggers. Search
optimization.

Moderately
complex
2D/3D,
dynamic
graphics,
multimedia.

Extra
High

Multiple resource
scheduling with
dynamically
changing priorities.
Microcode-level
control. Distributed
hard real time
control.

Difficult and
unstructured
numerical analysis:
highly accurate
analysis of noisy,
stochastic data.
Complex
parallelization.

Device timing
dependent coding,
micro
programmed
operations.
Performance
critical embedded
systems.

Highly coupled,
dynamic
relational and
object
structures.
Natural
language data
management.

Complex
multimedia,
virtual reality.


114

Cost
Driver

Rating
Very Low

RELY

0.75

Low

Nominal

High

Very
High

Extra High

0.88

1.00

1.15

1.39

0.93

1.00

1.09

1.19

0.88

1.00

1.15

1.30

1.66

0.91

1.00

1.14

1.29

1.49

0.95

1.00

1.06

1.13

TIME

1.00

1.11

1.31

1.67

STOR

1.00

1.06

1.21

1.57

0.87

1.00

1.15

1.30

DATA
CPLX

0.75

RUSE
DOCU

0.89

PVOL
ACAP

1.50

1.22

1.00

0.83

0.67

PCAP

1.37

1.16

1.00

0.87

0.74

Table 18: 17 Cost Drivers

Cont…
115

Cost
Driver

Rating
Very Low

Low

Nominal

High

Very
High

PCON

1.24

1.10

1.00

0.92

0.84

AEXP

1.22

1.10

1.00

0.89

0.81

PEXP

1.25

1.12

1.00

0.88

0.81

LTEX

1.22

1.10

1.00

0.91

0.84

TOOL

1.24

1.12

1.00

0.86

0.72

SITE

1.25

1.10

1.00

0.92

0.84

SCED

1.29

1.10

1.00

1.00

Extra High

1.00

0.78

Table 18: 17 Cost Drivers

116

Schedule estimation
Development time can be calculated using PMadjusted as a key factor and the
desired equation is:

TDEVnominal = [φ × ( PM adjusted )

( 0.28+ 0.2 ( B − 0.091))]

SCED %
∗
100

where Φ = constant, provisionally set to 3.67
TDEVnominal = calendar time in months with a scheduled constraint
B = Scaling factor
PMadjusted = Estimated effort in Person months (after adjustment)

117

Size measurement
Size can be measured in any unit and the model can be calibrated
accordingly. However, COCOMO II details are:
i. Application composition model uses the size in object points.
ii. The other two models use size in KLOC
Early design model uses unadjusted function points. These function points
are converted into KLOC using Table 19. Post architecture model may
compute KLOC after defining LOC counting rules. If function points are
used, then use unadjusted function points and convert it into KLOC using
Table 19.


118

Language

SLOC/UFP

Ada

71

AI Shell

49

APL

32

Assembly

320

Assembly (Macro)

213

ANSI/Quick/Turbo Basic

64

Basic-Compiled

91

Basic-Interpreted

128

C

128

C++

29
Table 19: Converting function points to lines of code


Cont…119

Language

SLOC/UFP

ANSI Cobol 85

91

Fortan 77

105

Forth

64

Jovial

105

Lisp

64

Modula 2

80

Pascal

91

Prolog

64

Report Generator

80

Spreadsheet

6

Table 19: Converting function points to lines of code

120

Example: 4.11
Consider the software project given in example 4.10. Size and scale factor
(B) are the same. The identified 17 Cost drivers are high reliability (RELY),
very high database size (DATA), high execution time constraint (TIME),
very high analyst capability (ACAP), high programmers capability (PCAP).
The other cost drivers are nominal. Calculate the effort in Person-Months for
the development of the project.


121

Solution

Here

B = 1.1129
PMnominal

= 194.41 Person-months

PM adjusted

 17

 i =7

= 194.41 x (1.15 x 1.19 x 1.11 x 0.67 x 0.87)
= 194.41 x 0.885
= 172.05 Person-months


122

Putnam Resource Allocation Model
Norden of IBM
Rayleigh curve
Model for a range of hardware development projects.
Overall Curve
Design and Coding

Persons

Time
Fig.6: The Rayleigh manpower loading curve

123

Putnam observed that this curve was a close
approximation at project level and software subsystem
level.

No. of projects = 150


124

The Norden / Rayleigh Curve
The curve is modeled by differential equation

dy
− at 2
m(t ) =
= 2kate
dt

--------- (1)

dy
dt

= manpower utilization rate per unit time

a

= parameter that affects the shape of the curve

K

= area under curve in the interval [0, ∞ ]

t

= elapsed time

125

On Integration on interval [o, t]
y(t) = K [1-e-at2] -------------(2)
Where y(t): cumulative manpower used upto time t.
y(0) = 0
y(∞) = k
The cumulative manpower is null at the start of the project, and
grows monotonically towards the total effort K (area under the
curve).


126

d2y
− at 2
= 2kae [1 − 2at 2 ] = 0
dt 2
1
2
td =
2a
“td”: time where maximum effort rate occurs
Replace “td” for t in equation (2)


 1 − e 2 t d2
E = y (t ) = k


E = y ( t ) = 0 . 3935 k
2
td

a=


 = K (1 − e − 0 .5 )



1
2
2t d

127

1
Replace “a” with
2 in the Norden/Rayleigh model. By
2t d
making this substitution in equation we have

m(t ) =

2K
2
2t d

−

te

t2
2
2t d

t2

K − 2td2
= 2 te
td

128

a=2

a=0.5
m (t)
Person

a=0.222

a=0.125

Time (years)
Fig.7: Influence of parameter ‘a’ on the manpower
distribution

129

At time t=td, peak manning m (td) is obtained and denoted by mo.

mo =
k
td
m0
e

k
td e

= Total project cost/effort in person-years.
= Delivery time in years
= No. of persons employed at the peak
= 2.71828


130

Example: 4.12
A software development project is planned to cost 95 MY in a period
of 1 year and 9 months. Calculate the peak manning and average rate
of software team build up.


131

Solution
Software development cost
Peak development time
Peak manning

k=95 MY
td = 1.75 years
mo=

k
td e

95
= 32.94 = 33 persons
1.75 × 1.648
Average rate of software team build up

=

m0
td

=

33
= 18.8 persons / year or 1.56 person / month
1.75

132

Example: 4.13
Consider a large-scale project for which the manpower requirement is
K=600 PY and the development time is 3 years 6 months.
(a)Calculate the peak manning and peak time.
(b)What is the manpower cost after 1 year and 2 months?


133

Solution
(a) We know td=3 years and 6 months = 3.5 years
NOW

m0 =

∴ m0 =

K
td e

600/(3.5x1.648) ≅ 104 persons


134

(b) We know

[

y (t ) = K 1 − e

− at 2

]

t = 1 year and 2 months
= 1.17 years
a=

1
1
=
= 0.041
2
2
2t d 2 × (3.5)

[

y (1 .17 ) = 600 1 − e

− 0 . 041 (1 . 17 ) 2

]

= 32.6 PY


135

Difficulty Metric
Slope of manpower distribution curve at start time t=0 has
some useful properties.

d2y
− at 2
m' (t ) = 2 = 2kae (1 − 2at 2 )
dt

Then, for t=0
2K K
m' (0) = 2 Ka = 2 = 2
2t d t d

136

K
The ratio t 2 is called difficulty and denoted by D,
d
which is measured in person/year :

k
D= 2 persons/year
td


137

Project is difficult to develop
if

Manpower demand
is high

When time schedule
is short


138

Peak manning is defined as:

m0 =

k
td e

k m0 e
D= 2 =
td
td

Thus difficult projects tend to have a higher peak
manning for a given development time, which is in line
with Norden’s observations relative to the parameter “a”.

139

Manpower buildup

D is dependent upon “K”. The derivative of D relative to
“K” and “td” are
D' (t d ) =

−2k
3
td

persons / year 2

1
D ' (k ) = 2 year − 2
td

140

D1(K) will always be very much smaller than the absolute value of
D1(td). This difference in sensitivity is shown by considering two
projects
Project A
Project B

: Cost = 20 PY & td = 1 year
: Cost = 120 PY & td = 2.5 years

The derivative values are
Project A
Project B

: D` (td) = -40 & D`(K) = 1
: D` (td) = -15.36 & D`(K) = 0.16

This shows that a given software development is time sensitive.

141

Putnam observed that
Difficulty derivative relative to time
Behavior of s/w development
If project scale is increased, the development time also
increase to such an extent that k remains constant
3
td
around a value which could be 8,15,27.


142

It is represented by D0 and can be expressed as:
k
D0 = 3 person / year 2
td

D0 =8, new s/w with many interfaces & interactions
with other systems.
D0 =15, New standalone system.
D0 =27, The software is rebuild form existing software.


143

Example: 4.14
Consider the example 4.13 and calculate the difficulty and
manpower build up.


144

Solution
We know

K
Difficulty D = 2
td
600
=
= 49 person / year
2
(3.5)

Manpower build up can be calculated by following equation

K
D0 = 3
td
=

600
= 14 person / year 2
(3.5)3


145

Productivity Versus Difficulty

Productivity = No. of LOC developed per person-month
P ∞ Dβ
Avg. productivity
P=

LOC produced
cumulative manpower
used to produce code


146

P = S/E
P = φD −2 / 3
S = φD − 2 / 3 E
= φD − 2 / 3 (0.3935 K )
−

2
3

k 
S = φ  2  k (0.3935)
 td 

S = 0.3935 φ K

1/ 3

td

4/3


147

0.39 φ

c
Technology Factor

Hardware
constraints

Experience
Complexity

Programming
environment


148

C

610 – 57314
K : P-Y
T : Years

S = CK

1/ 3 4 / 3
td

−1 / 3 −4 / 3
td

C = S .K
The trade off of time versus cost
1/ 3 4 / 3
d

K t
K

= S /C

1  S 
=

4 
td  C 

3


149

C = 5000
S = 5,00,000 LOC

1
3
K = 4 (100 )
td

td (years)

K (P-Y)

5.0
4.0
3.5
3.0

1600
3906
6664
12346

Table 20: (Manpower versus development time)

150

Development Subcycle
All that has been discussed so far is related to project life cycle as
represented by project curve
Manpower
distribution
Project

Requirements
& Specification

Design code
development

Test &
Validation

Ma

Fig.8: Project life cycle

inte
nan
Time

ce
151

Project life cycle

Project curve is the addition of two curves

Development
Curve

Test &
Validation
Curve


152

md (t) = 2kdbt e-bt2
∴
yd (t) = Kd [1-e-bt2]
An examination of md(t) function shows a non-zero value of md
at time td.
This is because the manpower involved in design & coding is
still completing this activity after td in form of rework due to
the validation of the product.
Nevertheless, for the model, a level of completion has to be
assumed for development.
It is assumed that 95% of the development will be completed
by the time td.

153

−bt 2
yd (t )
= 1 − e = 0.95
Kd

∴ We may say that

1
b= 2
2tod

Tod: time at which development curve exhibits a peak
manning.

t od

td
=
6

154

Relationship between Kd & K must be established.
At the time of origin, both cycles have the same slope.

K K d  dmd 
 dm 

 = 2 = 2 =

 dt  o t d tod  dt  o
Kd=K/6

Kd
K
D = 2 = 2
td
t od

155

This does not apply to the manpower build up D0.

Kd
K
Do = 3 =
3
td
6tod
Conte investigated that
Larger projects

reasonable

Medium & small projects

overestimate


156

Example: 4.15
A software development requires 90 PY during the total development
sub-cycle. The development time is planned for a duration of 3 years
and 5 months
(a)Calculate the manpower cost expended until development time
(b) Determine the development peak time
(c) Calculate the difficulty and manpower build up.


157

Solution
(a) Duration td = 3.41 years
We know from equation

−btd
yd (t )
= 1 − e = 0.95
Kd

yd (t d )
= 0.95
Kd

Yd (t d ) = 0.95 × 90
= 85.5 PY


158

(b) We know from equation

t od

t od

td
=
6

td
=
= 3 . 41 / 2 . 449 = 1 . 39 years
6

≅ 17 months


159

(c) Total Manpower development

K d = yd (t d ) / 0.95
= 85.5 / 0.95 = 90

K = 6 K d = 90 × 6 = 540 PY
2
D = K / t d = 540 /(3.41) 2 = 46

persons/years

K
Do = 3 = 540 /(3.41) 3 = 13.6
td

persons/years2


160

Example:4.16

A software development for avionics has consumed 32 PY
up to development cycle and produced a size of 48000
LOC. The development of project was completed in 25
months. Calculate the development time, total manpower
requirement, development peak time, difficulty,
manpower build up and technology factor.


161

Solution:

Development time td = 25 months = 2.08 years
Yd (t d )
32
=
= 33.7 PY
Total manpower development k d =
0.95 0.95

Development peak time

t od =

(t d )

= 0.85 years = 10 months

6

K = 6Kd = 6 x 33.7 = 202 PY
k
202
D= 2 =
= 46.7 pesons / years
2
t d (2.08)

162

D0 =

k
3
td

=

202
( 2.08)3

= 22.5 Persons / year 2

Technology factor

C = SK

−1 / 3

td

−4 / 3

= 3077


163

Example 4.17

What amount of software can be delivered in 1 year 10 months in an
organization whose technology factor is 2400 if a total of 25 PY is
permitted for development effort.


164

Solution:

td = 1.8 years
Kd = 25 PY
K = 25 x 6 = 150 PY
C = 2400

We know

S = CK

1/3

td

4/3

= 2400 x 5.313 x 2.18 = 27920 LOC


165

Example 4.18

The software development organization developing real time
software has been assessed at technology factor of 2200. The
maximum value of manpower build up for this type of
software is Do=7.5. The estimated size to be developed is
S=55000 LOC.
(a) Determine the total development time, the total
development manpower cost, the difficulty and the
development peak manning.
(b) The development time determined in (a) is considered too
long. It is recommended that it be reduced by two months.
What would happen?

166

Solution
We have

1/ 3

S = CK t d

4/3

3

s
4
  = ktd
c
3

S
7
which is also equivalent to   = Do t d
 
C 

then

 1 S
td =   
 D0  C 


3 1/ 7







167

Since

S
= 25
C

td = 3 years
3
K = D0t d = 7.5 × 27 = 202 PY

202
Total development manpower cost Kd =
= 33.75PY
06

D = D0td = 22.5 persons / year
td
3
tod =
=
= 1.2 years
6
6

168

Md(t) = 2kd
Yd(t) = kd
Here

-bt2
bte

-bt2
(1-e

)

t = tod

Peak manning

= mod = Dtod e

−1 / 2

= 22.5 x 1.2 x .606 = 16 persons


169

III. If development time is reduced by 2 months

Developing
s/w at higher
manpower
build-up

Producing
less software


170

(i) Increase Manpower Build-up

1S
Do = 7  
td  C 

3

Now td = 3 years – 2 months = 2.8 years
Do = ( 25)3 /( 2.8) 7 = 11.6 persons / years
3
k = D0t d = 254 PY

254
Kd =
= 42.4 PY
6

171

D = D0td = 32.5 persons / year
The peak time is tod = 1.14 years
Peak manning
mod = Dtod e-0.5
= 32.5 x 1.14 x 0.6
= 22 persons
Note the huge increase in peak manning & manpower
cost.


172

(ii) Produce Less Software
3

S
7
= D0t d = 7.5 × (2.8) 7 = 10119.696
 
C 
3

S
  = 21.62989
C 

Then for

C=2200
S=47586 LOC


173

Productivity versus difficult
Example 4.19

A stand alone project for which the size is estimated at 12500
LOC is to be developed in an environment such that the
technology factor is 1200. Choosing a manpower build up
Do=15, Calculate the minimum development time, total
development man power cost, the difficulty, the peak manning,
the development peak time, and the development productivity.


174

Solution
Size (S)

= 12500 LOC

Technology factor (C)

= 1200

Manpower buildup (Do)

= 15

Now

4
S = CK 1/ 3t d / 3

S
4
= K 1 / 3t d / 3
C
3

S
4
  = Kt d
C 

175

K
Also we know Do = 3
td
3
3
K = Dot d = Dot d
3

Hence

S
7
  = Dotd
C
3

 12500 
7
 = 15t d
Substituting the values, we get 
 1200 
 (10.416) 
td = 

15


3

1/ 7

t d = 1.85 years

176

(i) Hence Minimum development time (td)=1.85 years
K
(ii) Total development manpower cost K d =
6

Hence

3
K =15td

=15(1.85)3=94.97 PY
K 94.97
Kd = =
= 15.83 PY
6
6

(iii) Difficulty

D=

K
2
td

=

94.97
(1.85)

2

= 27.75 Persons / year


177

(iv) Peak Manning

m0 =

K
td e

94.97
=
= 31.15Persons
1.85×1.648

(v) Development Peak time

tod

td
=
6

1.85
=
= 0.755 years
2.449

178

(vi) Development Productivity
No .of lines of code ( S )
=
effort ( K d )

12500
=
= 789.6 LOC / PY
15.83


179

Software Risk Management

We Software developers are extremely optimists.
We assume, everything will go exactly as planned.
Other view
not possible to predict what is going to happen ?
Software surprises
Never good news

180

Risk management is required to reduce this surprise
factor
Dealing with concern before it becomes a crisis.
Quantify probability of failure & consequences of failure.


181

What is risk ?
Tomorrow’s problems are today’s risks.

“Risk is a problem that may cause some loss or
threaten the success of the project, but which has
not happened yet”.


182

Risk management is the process of identifying addressing
and eliminating these problems before they can damage
the project.
Current problems &

Potential Problems


183

Typical Software Risk

Capers Jones has identified the top five risk factors that
threaten projects in different applications.
1.

Dependencies on outside agencies or factors.
•

Availability of trained, experienced persons

•

Inter group dependencies

•

Customer-Furnished items or information

•

Internal & external subcontractor relationships

184

2.

Requirement issues
Uncertain requirements

Wrong product
or
Right product badly
Either situation results in unpleasant surprises and
unhappy customers.

185

• Lack of clear product vision
• Lack of agreement on product requirements
• Unprioritized requirements
• New market with uncertain needs
• Rapidly changing requirements
• Inadequate Impact analysis of requirements changes


186

Management Issues

3.

Project managers usually write the risk management
plans, and most people do not wish to air their
weaknesses in public.
•
•

Inadequate planning
Inadequate visibility into actual project status

•

Unclear project ownership and decision making

•

Staff personality conflicts

•

Unrealistic expectation

•

Poor communication

187

4.

Lack of knowledge
•

Inadequate training

•

Poor understanding of methods, tools, and
techniques

•

Inadequate application domain experience

•

New Technologies

•

Ineffective, poorly documented or neglected
processes

188

5.

Other risk categories
•

Unavailability of adequate testing facilities

•

Turnover of essential personnel

•

Unachievable performance requirements

•

Technical approaches that may not work


189

Risk Management Activities

Risk Identification
Risk
Assessment
Risk
Management

Risk Analysis
Risk Prioritization
Risk Management
Planning

Risk Control
Fig. 9: Risk Management
Activities

Risk Monitoring
Risk Resolution


190

Risk Assessment
Identification of risks
Risk analysis involves examining how project outcomes
might change with modification of risk input variables.
Risk prioritization focus for severe risks.
Risk exposure: It is the product of the probability of incurring
a loss due to the risk and the potential magnitude of that loss.


191

Another way of handling risk is the risk avoidance. Do not do
the risky things! We may avoid risks by not undertaking
certain projects, or by relying on proven rather than cutting
edge technologies.


192

Risk Control
Risk Management Planning produces a plan for dealing with
each significant risks.
Record decision in the plan.
Risk resolution is the execution of the plans of dealing with
each risk.


193

Multiple Choice Questions
Note: Choose most appropriate answer of the following questions:
4.1 After the finalization of SRS, we may like to estimate
(a) Size
(b) Cost
(c) Development time
(d) All of the above.
4.2 Which one is not a size measure for software
(a) LOC
(b) Function Count
(c) Cyclomatic Complexity
(d) Halstead’s program length
4.3 Function count method was developed by
(a) B.Beizer
(b) B.Boehm
(c) M.halstead
(d) Alan Albrecht
4.4 Function point analysis (FPA) method decomposes the system into functional
units. The total number of functional units are
(a) 2
(b) 5
(c) 4
(d) 1


194

4.5 IFPUG stand for
(a) Initial function point uniform group
(b) International function point uniform group
(c) International function point user group
(d) Initial function point user group
4.6 Function point can be calculated by
(a) UFP * CAF
(b) UFP * FAC
(c) UFP * Cost
(d) UFP * Productivity
4.7 Putnam resource allocation model is based on
(a) Function points
(b) Norden/ Rayleigh curve
(c) Putnam theory of software management
(d) Boehm’s observation on manpower utilisation rate
4.8 Manpower buildup for Putnam resource allocation model is
2
( a ) K / t d persons / year 2

3
(b) K / t d persons / year 2

2
(c ) K / t d persons / year

3
( d ) K / t d persons / year


195

4.9 COCOMO was developed initially by
(a) B.W.Bohem
(c) B.Beizer

(b) Gregg Rothermal
(d) Rajiv Gupta

4.10 A COCOMO model is
(a) Common Cost estimation model
(b) Constructive cost Estimation model
(c) Complete cost estimation model
(d) Comprehensive Cost estimation model
4.11 Estimation of software development effort for organic software is COCOMO is
(a) E=2.4(KLOC)1.05PM
(b) E=3.4(KLOC)1.06PM
(c) E=2.0(KLOC)1.05PM
(d) E-2.4(KLOC)1.07PM
4.12 Estimation of size for a project is dependent on
(a) Cost
(b) Schedule
(c) Time
(d) None of the above
4.13 In function point analysis, number of Complexity adjustment factor are
(a) 10
(b) 20
(c) 14
(d) 12

196

4.14 COCOMO-II estimation model is based on
(a) Complex approach
(b) Algorithm approach
(c) Bottom up approach
(d) Top down approach
4.15 Cost estimation for a project may include
(a) Software Cost
(b) Hardware Cost
(c) Personnel Costs
(d) All of the above
4.16 In COCOMO model, if project size is typically 2-50 KLOC, then which mode
is to be selected?
(a) Organic
(b) Semidetached
(c) Embedded
4.17 COCOMO-II was developed at
(a) University of Maryland
(c) IBM

(b) University of Southern California
(d) AT & T Bell labs

4.18 Which one is not a Category of COCOMO-II
(a) End User Programming
(b) Infrastructure Sector
(c) Requirement Sector
(d) System Integration

197

4.19 Which one is not an infrastructure software?
(a) Operating system
(b) Database management system
(c) Compilers
(d) Result management system
4.20 How many stages are in COCOMO-II?
(a) 2
(b) 3
(c) 4
(d) 5
4.21 Which one is not a stage of COCOMO-II?
(a) Application Composition estimation model
(b) Early design estimation model
(c) Post architecture estimation model
(d) Comprehensive cost estimation model
4.22 In Putnam resource allocation model, Rayleigh curve is modeled by the equation
2

(a ) m(t ) = 2at e − at
− at 2
(c) m(t ) = 2 Kat e

2

(b) m(t ) = 2 Kt e − at
− at 2
( d ) m(t ) = 2 Kbt e


198

4.23 In Putnam resource allocation model, technology factor ‘C’ is defined as
−
(a) C = SK −1/ 3t d 4 / 3

4
(b) C = SK 1/ 3t d / 3

−
(c) C = SK 1/ 3t d 4 / 3

4
(d ) C = SK −1/ 3t d / 3

4.24 Risk management activities are divided in
(a) 3 Categories
(b) 2 Categories
(c) 5 Categories
(d) 10 Categories
4.25 Which one is not a risk management activity?
(a) Risk assessment
(b) Risk control
(c) Risk generation


199

Exercises
4.1 What are various activities during software project planning?
4.2 Describe any two software size estimation techniques.
4.3 A proposal is made to count the size of ‘C’ programs by number of
semicolons, except those occurring with literal strings. Discuss the
strengths and weaknesses to this size measure when compared with the
lines of code count.
4.4 Design a LOC counter for counting LOC automatically. Is it language
dependent? What are the limitations of such a counter?
4.5 Compute the function point value for a project with the following
information domain characteristics.
Number of user inputs = 30
Number of user outputs = 42
Number of user enquiries = 08
Number of files = 07
Number of external interfaces = 6
Assume that all complexity adjustment values are moderate.

200

Exercises
4.6 Explain the concept of function points. Why FPs are becoming
acceptable in industry?
4.7 What are the size metrics? How is function point metric advantageous
over LOC metric? Explain.
4.8 Is it possible to estimate software size before coding? Justify your answer
with suitable example.
4.9 Describe the Albrecht’s function count method with a suitable example.
4.10 Compute the function point FP for a payroll program that reads a file of
employee and a file of information for the current month and prints
cheque for all the employees. The program is capable of handling an
interactive command to print an individually requested cheque
immediately.


201

Exercises
4.11 Assume that the previous payroll program is expected to read a file
containing information about all the cheques that have been printed. The
file is supposed to be printed and also used by the program next time it is
run, to produce a report that compares payroll expenses of the current
month with those of the previous month. Compute functions points for
this program. Justify the difference between the function points of this
program and previous one by considering how the complexity of the
program is affected by adding the requirement of interfacing with
another application (in this case, itself).
4.12 Explain the Walson & Felix model and compare with the SEL model.
4.13 The size of a software product to be developed has been estimated to be
22000 LOC. Predict the manpower cost (effort) by Walston-Felix Model
and SEL model.
4.14 A database system is to be developed. The effort has been estimated to
be 100 Persons-Months. Calculate the number of lines of code and
productivity in LOC/Person-Month.

202

Exercises
4.15 Discuss various types of COCOMO mode. Explain the phase wise
distribution of effort.
4.16 Explain all the levels of COCOMO model. Assume that the size of an
organic software product has been estimated to be 32,000 lines of code.
Determine the effort required to developed the software product and the
nominal development time.
4.17 Using the basic COCOMO model, under all three operating modes,
determine the performance relation for the ratio of delivered source code
lines per person-month of effort. Determine the reasonableness of this
relation for several types of software projects.
4.18 The effort distribution for a 240 KLOC organic mode software
development project is: product design 12%, detailed design 24%, code
and unit test 36%, integrate and test 28%. How would the following
changes, from low to high, affect the phase distribution of effort and the
total effort: analyst capability, use of modern programming languages,
required reliability, requirements volatility?

203

Exercises
4.19 Specify, design, and develop a program that implements COCOMO.
Using reference as a guide, extend the program so that it can be used as a
planning tool.
4.20 Suppose a system for office automation is to be designed. It is clear
from requirements that there will be five modules of size 0.5 KLOC, 1.5
KLOC, 2.0 KLOC, 1.0 KLOC and 2.0 KLOC respectively. Complexity,
and reliability requirements are high. Programmer’s capability and
experience is low. All other factors are of nominal rating. Use COCOMO
model to determine overall cost and schedule estimates. Also calculate
the cost and schedule estimates for different phases.
4.21 Suppose that a project was estimated to be 600 KLOC. Calculate the
effort and development time for each of the three modes i.e., organic,
semidetached and embedded.
4.22 Explain the COCOMO-II in detail. What types of categories of projects
are identified?


204

Exercises
4.23 Discuss the Infrastructure Sector of COCOMO-II.
4.24 Describe various stages of COCOMO-II. Which stage is more popular
and why?
4.25 A software project of application generator category with estimated size
of 100 KLOC has to be developed. The scale factor (B) has high
percedentness, high development flexibility. Other factors are nominal.
The cost drivers are high reliability, medium database size, high
Personnel capability, high analyst capability. The other cost drivers are
nominal. Calculate the effort in Person-Months for the development of
the project.
4.26 Explain the Putnam resource allocation model. What are the limitations
of this model?
4.27 Describe the trade-off between time versus cost in Putnam resource
allocation model.
4.28 Discuss the Putnam resources allocation model. Derive the time and
effort equations.

205

Exercises
4.29 Assuming the Putnam model, with S=100,000 , C=5000, Do=15,
Compute development time td and manpower development Kd.
4.30 Obtain software productivity data for two or three software development
programs. Use several cost estimating models discussed in this chapter.
How to the results compare with actual project results?
4.31 It seems odd that cost and size estimates are developed during software
project planning-before detailed software requirements analysis or design
has been conducted. Why do we think this is done? Are there
circumstances when it should not be done?
4.32 Discuss typical software risks. How staff turnover problem affects
software projects?
4.33 What are risk management activities? Is it possible to prioritize the risk?


206

Exercises
4.34 What is risk exposure? What techniques can be used to control each
risk?
4.35 What is risk? Is it economical to do risk management? What is the effect
of this activity on the overall cost of the project?
4.36 There are significant risks even in student projects. Analyze a student
project and list all the risk.


207

Chapter 4 software project planning

More Related Content

What's hot

Viewers also liked

Similar to Chapter 4 software project planning

Recently uploaded

Chapter 4 software project planning