Extreme software estimation (xsoft estimation)

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 02 Issue: 12 | Dec-2013, Available @ http://www.ijret.org 366
EXTREME SOFTWARE ESTIMATION (XSOFT ESTIMATION)
E. Karunakaran1
, P.Radhakrishnan2
1, 2
Department of Computer Science & Engineering, Pondicherry Engineering College, Pillaichavady,
Puducherry – 605014, India.
Abstract
The management of software cost, development effort and project planning are the key aspects of software development. Throughout
the sixty-odd years of software development, the industry has gone at least four generations of programming languages and three
major development paradigms. Still the total ability to move consistently from idea to product is yet to be achieved. In fact, recent
studies document that the failure rate for software development has risen almost to 50 percent. There is no magic in managing
software development successfully, but a number of issues related to software development make it unique. The basic problem of
software development is risky. Some example of risk is error in estimation, schedule slips, project cancelled after numerous slips,
high defect rate, system goes sour, business misunderstanding, false feature rich, staff turnover. XSoft Estimation addresses the risks
by accurate measurement. A new methodology to estimate using software COSMIC-Full Function Point and named as EXtreme
Software Estimation (XSoft Estimation). Based on the experience gained on the original XSoft project develpment, this paper
describes what makes XSoft Estimation work from sizing to estimation.
Keywords: -COSMIC function size unit, XSoft Estimation, XSoft Measurement, Cost Estimation.
---------------------------------------------------------------------***----------------------------------------------------------------------
1. INTRODUCTION
Software Measurement : It is fundamental to any
engineering discipline and software engineering is not an
exception. Measurements in the physical world can be
categorized in two ways namely a direct measure (eg. The
length of bolt) and indirect measurement (e.g. the quality of
the bolts). Software metrics can be categorized similarly.
Direct measures of the software engineering process include
cost and effort applied. Direct measures of the product
include lines of code (LOC) produced, execution speed,
memory size and defects reported over some set period of
time. Indirect measures of the product include functionality,
quality, complexity, efficiency, reliability, maintainability and
many other abilities. Throughout the industry, the size
oriented or/and function oriented metrics are used. Size-
oriented metrics make use of the line of codes as a
normalizing factor so that the size is calculated only after
finishing the product. The function point is derived from
measures of the information domain and a subjective
assessment of problem complexity. Since functionality cannot
be measured directly, it must be derived indirectly using other
direct measures. So, the modified function points method
which is simple to understand, easy to evaluate is known as
COSMIC-FULL Function Point (2).
Estimation : In the early days of computing, software cost
comprised a small percentage of overall computer-based
system cost. An order of magnitude error in estimates of
software cost had relatively little impact. Today, software is
the most expensive element in most computer-based systems.
A cost estimation error can make the difference between profit
and loss. Software cost and effort estimation will never be an
exact science. To achieve reliable cost and effort estimates, a
number of options arises namely :
1. Delay estimation until late in the project.
2. Base estimation on similar projects that have already
been completed.
3. Use relatively simple “decomposition techniques” to
generate project cost and effort estimates.
4. Use one or more empirical models for software cost
and effort estimation.
Unfortunately, the first option, however attractive, is not
practical. Cost estimates must be provided “up-front”. The
second option can work reasonably well if the current project
is quite similar to past efforts and other project influences (eg.
The customer, business conditions, deadlines) is equivalent
and also past experience has not always been a good indicator
of future results. The third option decomposition technique
takes a “divide and conquer” approach to software project
estimation. By decomposing a project into major functions
and related software engineering activities, cost and effort
estimation can be performed in a stepwise fashion. The fourth
option “empirical estimation models” can be used to
complement decomposition techniques and offer a potentially
valuable estimation approach in their own right.
2. RELATED WORK
[1] The COSMIC Method defines a standardized measure of
software Functional Size expressed in CFP units. The

_______________________________________________________________________________________
measurement is carried out by mapping the FUR of the
software onto the COSMIC Generic Software Model (shown
in Figure 1). The purpose of the measurement and scope of
the software to be measured the level of decomposition and
level of granularity of the software. As can be seen in Fig.1,
there are four different data movement types. Entry type
move data across the boundary from the user to the functional
process. Exit type move data across the boundary to the user.
Read type move data from persistent storage in the functional
process. Write type move data from the functional process to
persistent storage.
Fig 1 : Functional Process
[2] Research work on the quality of the documentation using a
functional size method is recent. In this work documentation
quality rating scale a, b, c, d, e is fixed based on facts, some of
which can be listed as :
• The presence or absence of a data model.
• The presence or absence of information to identify
the data movements (entry, read, write, exit).
• The presence or absence of documentation enabling
identification of each functional process.
In this process three iterations are made for documentation
quality and the quality rating are never decreasing from
iteration to iteration. In the third iteration the documentation
quality of a functional process is always equal to or higher
than in the second and first iteration. At the end of third
iteration only the highest rating like a or b will exists.
[3] A tool is developed for automated estimation of Code Size
based on UML profile. The main functionalities of the tool
are importing information modeled using the UML Profile and
store component data needed for CFP measurement and Code
Size estimation to calculate estimated code size using linear
regression and present estimation results.
In [1] the enter type and exit type of a function process is
considered as one cfsu (cosmic function size unit) irrespective
of number of fields. The development time will vary based
on the software used. So, sizing and estimation will not be
accurate. In [2] based on the quality of the documentation,
the size and estimation are done, but it is not always sure that
the software development will be made according to the
documentation. In the third and final iteration of documenting
all the stories are highly rated, and in general the
documentation is prepared at the end of the software
development. So, the software may not be improved.
In [3] it is required to convert from COSMIC concept to UML
concept. The tool is required to calculate and it is not easy to
use and it is also hard to convince the customer about the size
and estimation.
To make the measurement and estimation more accurate,
some research is applied based on practical experience using
COSMIC FFU and it is named as XSoft Estimation.
3. PROPOSED WORK
XSoft Measurement Process : The purpose and scope is to
size the functionality which corresponds to the effort
interested in. For an application, written partly in Java and in
VB, measurement is made separately for software size and
development effort. First identify the software layers,
boundary, users, trigger, functional processes, data groups and
sub-processes. In each of the above identify the number of
entry, exit, read and write.
XSoft Estimation : A model is based on experience and takes
the form
D = S * V
Where
D is one of the estimated value (eg. Effort, Cost, Project
duration)
S is the size of the project and V is one of the variable values
depends upon effort, cost, time estimation.
The following five different projects were taken to measure
the size of the software and the result are given in the table :
1,2,3,4, and 5
Table: 1 - Project-1 : Mobile Park
Layers Functional Process Cfsu
Telecommunication MobilePark Owner 14
City Entry 14
Zone Entry 15
Policy Entry 15
Loading Credits 15
Effected Parking 4
Client Account 4
Parking Policy Database 4
Town Parking Tax 4
Credit Purchase 4
Upload 4
Download 4
Total 101

_______________________________________________________________________________________
Table : 2 - Project-2 : EX-CARGO
MIS Shipment Details 29
Consignor / Consignee Details 17
Inbound/Outbound 22
Report-Finance 22
Report-Planning 10
Booking Details 13
Shipment Status 10
Planning by Truck 10
Status Report 21
Shipment Sold 19
Cosignee Reverse 14
Account Details 12
Total 199
Table : 3 - Project-3 : L-SALES
MIS Estimation 22
Sales 374
Credit 12
Cash 48
File 67
User 15
Invoice 25
Rate 87
Total 313
Table : 4 - Project-4 : DDE
Tele
Comm.
OPCGROUP
63
EnumOPCItemAttributes 8
OPCServer Object 27
OPCBrowser Object 18
OPCGroups Object 14
OPCGroup Object 12
OPCItems Object 18
OPCItem Object 5
IOPCCommon 15
IOPCEventServer 45
IOPCConnectionPointContainer 5
OPCHDAServer 24
OPCHDABrower Object 15
OPCHDAItems Object 72
Total 341
Table : 5 - Project-5 : FORGING
MIS Receipts 16
Issues 16
Stock 11
Ledge & Opening Balance 20
RM details 16
Queries – Ledger details 38
Inventory Maintenance 9
RM, TR Entry Details 10
TDC & TC Entry 14
Forging Entry & Backup 18
Queries – Forging details 20
Transfer Heatcode, ICIN, DN
Modification & ReIndexing
18
IOPCEventServer 45
Reports 50
Total 260
4. RESULTS AND DISCUSSION
The five different projects were used and find the man days
required to finish the functional process of each project
between the skilled and expert teams. As well as man days
required between various layers like MIS and Tele
Communication. The results are given in the table : 6, 7, 8 &
9.
Table : 6 Man days requirement – Comparison between Expert and Skilled teams
Functional Process Actual Size
(Cfsu)
Allotted Size (Cfsu) Actual Man Days (5 Cfsu)
E-TEAM S-TEAM E-TEAM S-TEAM
MobilePark Owner 14 7 7 7 10
City Entry 14 7 7 7 10
Zone Entry 15 8 7 8 9
Policy Entry 15 8 7 9 9
Loading Credits 15 8 7 8 9
Parking details 20 10 10 11 14
Upload & Download 8 4 4 5 6
AVERAGE 14.42 7.42 7 7.86 9.57
Project Name : Mobile Parking, Layer : Tele. Communication Software Used : Java

_______________________________________________________________________________________
Table : 7 Man days requirement – Comparison between Expert and Skilled teams
Functional Process Actual Size
(Cfsu)
Allotted Size (Cfsu) Actual Man Days (5 Cfsu)
E-TEAM S-TEAM E-TEAM S-TEAM
Estimation 22 12 10 1.7 3
Sales 32 17 15 1.8 2.8
Credit 12 7 5 1.2 2
Cash 48 26 22 1.9 2.7
File 67 37 30 2.1 2.6
User 15 9 6 1.3 2.2
Invoice 25 15 10 1.4 2
Rate 87 47 40 1.6 2.5
AVERAGE 38.5 21.25 17.25 1.625 2.475
Project Name : LIPS, Layer : MIS, Software Used : VB6
Fig. 2 : Expert Vs Skilled (Telecommunication) Fig.3 : Expert Vs Skilled (MIS)
Table : 8 Man days requirement – Comparison between Layers
SOFTWARE Actual Size (Cfsu) Actual Man Days (5
Cfsu)
Actual Time Taken (in
Man Days )
MIS T.C. MIS T.C. MIS T.C.
STORY - 1 22 14 2.95 10.71 13 30
STORY - 2 37 14 2.43 10.00 18 28
STORY - 3 12 15 2.92 9.33 7 28
STORY - 4 48 15 2.81 8.67 27 26
STORY - 5 67 15 2.76 8.67 37 26
STORY - 6 15 20 4.33 9.00 13 36
STORY - 7 25 8 2.60 8.75 13 14
STORY - 8 87 2.76 48
AVERAGE 38.5 3.37 9.30 22 26.86
Project Name : LIPS & MOBI LE PARK Layer : MIS & T.C. Software Used : VB6 & Java
0
2
4
6
8
10
12
14
16
1 2 3 4 5 6 7
ACTUALTIMETAKEN(INMANDAYS)
STORIES (AVERAGED TO 5 CFSU)
Expert Skilled
0
0.5
1
1.5
2
2.5
3
3.5
1 2 3 4 5 6 7 8
Actualtimetaken(inmandays)
Stories (averaged to 5 Cfsu)
E-TEAM S-TEAM

_______________________________________________________________________________________
Table : 9 Time taken – Comparison between different Software / Layers
SOFTWARE Actual Size (Cfsu) Actual Time Taken (Man Days for 5 Cfsu)
VB JAVA JAVA C++ VB JAVA JAVA C++
MIS MIS T.C. T.C. MIS MIS T.C. T.C.
STORY - 1 22 16 14 16 2.95 5.94 10.71 10.00
STORY - 2 37 27 14 15 2.43 5.19 10.00 9.33
STORY - 3 12 36 15 14 2.92 5.14 9.33 10.00
STORY - 4 48 38 15 14 2.81 4.74 8.67 9.29
STORY - 5 67 19 15 17 2.76 5.53 8.67 10.00
STORY - 6 15 32 20 10 4.33 5.00 9.00 11.00
STORY - 7 25 39 8 16 2.60 4.87 8.75 9.38
STORY - 8 87 50 2.76 4.00 9.31 9.80
AVERAGE 44.71 36.71 14.43 14.57 2.46 4.27 9.31 9.80
Fig. 4 : MIS Vs Telecommunication
The results show that the VB requires half the time to
complete the project of size 5 Cfsu than the time required for
Java. Also it is very clear that the time required for Java in
MIS is almost half of the time required in telecommunication
system. (Fig. 4).
The time effort for 1 Cfsu for different software and different
layers are estimated and furnished with table : 10.
Table : 10 Time effort – Comparison between different
Software / Layers
Layers/ Software VB 6 C++ JAVA
MIS 0.5 1.2 0.9
Tele
Communication
--- 2.0 1.9
Fig. 5 : Between Softwares (VB, Java, C++)
It is evident from the above table that the VB software
requires very less effort time to develop a Management
Information System followed by Java. In case of Tele
Communication C++ may be effective software language to
optimize time effort. (Fig. 5).
Finally, a new formula is proposed to determine the effort
estimation to arrive at optimum cost to develop the software.
Effort Estimation = SS * (SL + SU + DU)
Where SS – Software Size (using COSMIC – FFP)
SL – Software Layer
SU – Software Used
DU – Database Used
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8
ActualTimeTaken(inMandays)
Stories (Averaged to 5 Cfsu's)
MIS T.C.
0
2
4
6
8
10
12
2 3 4 5 6 7 8 9
ActualTimeTaken(inMandays)
Stories (averaged to 5 Cfsu)
VB-MIS JAVA-MIS
JAVA-T.C. C++-T.C.

_______________________________________________________________________________________
The effort value for different software, layers and databases
are furnished in the table : 11. The effort values given in the
table are based on the effort valued of the five projects used
for the present study.
Table : 11 Time effort – Comparison between different
Software / Layers
SL/SU/DU Effort (for 1 Cfsu)
SL ---
MIS
Tele Communication
0.3
1.4
SU ---
VB
Java
C++
.2
.4
.6
DU ---
Oracle
MS-Acess
My-SQL
.100
.050
.075
Cost Estimation = Effort Estimation * Cost per Man day.
The cost per man day will vary from company to company and
the cost estimation for the present study is based on the cost
per man day fixed by this Company is given in table :12.
Table : 12 Cost Estimation Details
Sl.no. Description Hourly Monthly
1 Developer, Tester,
Technical Writer
400 50,000
2 Consultant, System
Analyst, Module Leader,
System Administrator
600 75,000
3 System Architect 800 1,00,000
4 Project Manager, Senior
Consultant
1200 1,50,000
5. IMPLEMENTATION – A CASE STUDY
A Case study is undertaken to explain the implementation
process which is shown in Fig 6. The various stages using
XSOFT Measurement are detailed below
Fig. 6 E-Mail Server Block Diagram
5.1 Purpose and Scope
Presentation Layer
Incoming / Outgoing
Mail Browsing Modules
Mail Editing Modules
Control Layer
System Core Layer
Incoming Mail
Control
Mail Box
Access
Outgoing Mail
Control
POP3
Manager
MAIL BOX
Manager
SMTP
Manager

_______________________________________________________________________________________
E-mail component is the core of all Linux-based e-mail
applications. The main functionality of the component is to
process conversations with remote mail servers by
implementing the most popular email protocols, POP3, SMTP,
MIME, etc. in internet. The component also contains the
functionality to store local mail boxes for both incoming and
outgoing messages.
MAPPING PHASE I
Triggering Events & Functional Processes
SMTPObject
POP3 Object, Message List Object, Recipient List Object,
Attachment List Object & Mail Box Object
Data Group
Senders Details & Receivers Details
MAPPING PHASE II
Sub-Processes
SMTPObject - Add Attachment, Add Recipient, Cancel &
Send
POP3 Object - Cancel, Connect, Disconnect, Download
Header, GetDownloadableCount, GetDownloadableSize
& Download Messages.
Message List Object - Get, Delete & DeleteAll.
Recipient List Object -Get Recipient, Attachment List Object
& Get Attachment.
Mail Box Object - Open, Close, GetHeaders, GetMessage,
GetCount , Append & Delete.
Table : 13 Aggregate Result (Measurement)
Layers Functional Process Total Cfsu
Tele
Communication
SMTPObject 6
POP3 Object 21
Message List Object 8
Recipient List Object 2
Attachment List Object 2
Mail Box Object 13
Total 52
5.2 Estimation
The size of the Software = 52 Cfsu
Total Man days = 52 * 2 =104 (since 1 Cfsu = 2 man days)
Developers = 2 Skilled and 2 Beginners
Human Resource = 2 * 1 + 2 * .5 = 3
Estimated time to finish the project with 2 teams
(2 developer in one team) = 104 / 3 = 35 Working days.
Cost Estimation = 104 * 3200 = 3,32,800
CONCLUSIONS AND FUTURE WORK
In the present study an attempt has been made to address the
risk involved in software development with the help if the
proposed methodology which is developed with the features of
COSMIC FFP. Based on the extensive studies carried out
during the present, the following conclusion are drawn. It is
evident from the results that the man days required for the
expert team to complete either the MIS or Tele
Communication project are significantly less than the required
by the skilled team.
The software layer is also one the factors influencing the man
days required to complete the project. The man days required
to complete MIS project is almost three times less than the
telecommunication project. Another factor influencing the
man days is the type of software used for the project. VB
requires half the time to complete the project than the time
required for Java.
The effort time required to complete the MIS using VB is
significantly less than Java and C++. In case of
telecommunication systems C++ may be effective to optimize
time effort than VB and Java. A formula was developed to
determine the effort estimation to arrive at the optimum cost to
develop software.
Using the previous information about the developers ( i.e.. the
man days to complete one Cfsu in various Layers / Software)
the company can estimate the accurate cost estimation and
time taken to finish the project.
REFERENCES
[1]. ISO/IEC. 2003 19761: Software Engineering COSMIC
FFP: A Functional Size Measurement Method, 2003.
[2]. Jean-Marc Desharnais, Computer Engineening
Department, Bogazici University, Istanbul, Turkey, Using the
COSMIC Method to Evaluate the Quality of the
Documentation of Agile User Stories : 2011 Joint Conference
of the 21st
International Workshop on Software Measurement,
pp. 269-272.
[3]. Kenneth Lind, Electrical Systems Engineering, Saab
Automobile AB, Trollhattan, Sweden, CompSize : Automated
Size Estimation of Embedded Software Components, 2011
Joint Conference of the 21st
International Workshop on
Software Measurement, pp. 86-95.
[4]. Joseph C.Giarratano, Gary D.Riley, Expert Systems
Principles and Programming. Beijing: China Machine Press,
2006.
[5]. Z. Li, L.X.Lin, Y.B. Cao, Method and Practice of
Function Point Analysis. Beijing: Tsinghua University Press,
2005.
[6]. J.D. Zhu, C.H. Qi, X.M. Hou, “Estimating software cost
based on grey theory”, Computer Engineering and
Applications, vol. 46, no. 16, pp.71-73, 2010.

_______________________________________________________________________________________
Albrecht, A.: Measuring application development
productivity. Proc. of the IBM Applications Development
Symposium, Monterey, CA, 83-92, Oct. 1979.
[7]. Albrecht, A, and Gaffney, J.: Software function, source
lines of code, and development effort prediction: A software
science validation. IEEE Trans. Softw. Eng. SE-9, 6, 639–648,
1983.
[8]. A. Abran, J.M. Desharnais , A. Lesterhuis, B. Londeix, R.
Meli, P.M.Jorgensen, “A Review of Studies on Expert
Estimation of Software Development Effort”, Journal of
Systems and Software, vol. 70, no.1/2, pp. 58-65, 2004.
[9]. E. Miranda, Sizing User Stories Using Paired
Comparisons, Information and Software Technology (IST)
Journal, vol. 51, issue 9, September 2009, pp. 1327-1337.
[10]. Haugen N.C., An Empirical Study of Using Planning
Poker for User Story Estimation, AGILE 2006 Conference
(AGILE 2006), 23-28 July, 2006, Minneapolis, Minnesota,
USA. IEEE Computer Society 2006, pp. 23-34.
[11]. Extreme Programming on PWAP (ficticious) a wireless
application, Paul Rodrigues, E.Karunakaran, Naturesoft
Esolution Provider, Chennai, India. Agile International
Conference, Itally, 2001, pp. 73.
[12]. COSMIC – Common Software Measurement
International Consortium 2007. The COSMIC Funtional Size
Measurement Method – version 3.0 Measurement Manual
(The COSMIC Implementation Guide for ISO/IEC
19761:2003), September 2007.

Extreme software estimation (xsoft estimation)

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Extreme software estimation (xsoft estimation) (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Extreme software estimation (xsoft estimation)