Journal of Software Engineering and Applications, 2013, 6, 121-130
http://dx.doi.org/10.4236/jsea.2013.63016 Published Online March 2013 (http://www.scirp.org/journal/jsea)
121
Building Productivity Models for Small Enhancements
Jean-Marc Desharnais1,2, Gülşah Yıldızoğlu1, Alain April2, Alain Abran2
1Boğaziçi University, Istanbul, Turkey; 2Department of Software Engineering and IT, École de Technologie Supérieure, Montréal,
Canada.
Received January 9th, 2013; revised February 10th, 2013; accepted February 18th, 2013
ABSTRACT
Software is in constant evolution and many approaches have been suggested to study software maintenance productivity.
This research reports on a process to design and implement a productivity model of legacy software based on the meas-
urement of small functional enhancements using the COSMIC ISO 19761 international standard. Two motivations in-
fluence this research: 1) understanding the productivity of the software maintenance process to help manage the cost of
maintenance; 2) understanding the cost drivers that affect the software maintenance productivity. This research reports
on an empirical study of a productivity measurement program implemented in a large banking legacy system.
Keywords: Small Enhancements; Productivity; Productivity Models; Maintenance; Cost of Maintenance
1. Introduction
1.1. Definition of Maintenance
The software lifecycle can be divided into two distinct
parts, as presented in ISO 12207 [1]: the initial develop-
ment of the software and its use and ongoing mainte-
nance.
The international standard ISO 14764 [2] on software
maintenance defines four categories to classify the nature
of individual maintenance work requests: adaptive, cor-
rective, preventive, and perfective (see Table 1).
ISO 14764 classifies adaptive and perfective mainte-
nance as enhancements, and the corrective and preven-
tive maintenance as corrections [2]. This research is fo-
cused mainly on adaptive and corrective work requests
where most of the changes to functionality occur.
Table 1. ISO 14764 Software maintenance categories.
Category Description
Adaptive
Modifications to adapt a software product to
change in data requirements and
processing environments
Corrective
Reactive modification of a software product
performed after delivery to correct the faults
discovered. These modifications often repair code
to satisfy functional requirements
Preventive
Modification of a software product after delivery
to detect and correct latent faults before they
become operational faults
Perfective
Modification of a software product after delivery
implementing new or changed user requirements
which concern functional enhancements
to the software
In large organizations, most of the IT personnel are as-
signed to software maintenance—see Table 2 [3]. How-
ever, software maintenance is still a rather neglected ac-
tivity by both IT managers and academic research ac-
cording to a number of authors (Torchiano, Ricca, and
De Lucia [4], Koskinen [5], Kuhlmann [6]). Already in
1996 Basili et al. were reporting that there were not
enough empirical studies and available research data for
software maintenance [4] while Koskinen [5] was report-
ing that “Software maintenance and evolution is a con-
siderably understudied area while taking into account its
cost effect”. Many of the empirical studies on software
maintenance management date back to the pre-2000,
such as Abran [7], Lehman [8], Genuchten [9], Arfa [10],
Desharnais [11] and others.
Table 2 published in 2006 shows that between 2000
and 2005 there was an increase of nearly 4% of the
maintenance personnel in USA [12]. The projection for
the following years is showing that the number of per-
sonnel in maintenance would increase considerably over
Table 2. USA software personnel in software development
and maintenance—2006 [12].
Year Development
personnel
Maintenance
personnel
Total
personnel
Maintenance
percent
2000 750,000 2,000,000 2,750,000 72.73%
2005 775,000 2,500,000 3,275,000 76.34%
2010 800,000 3,000,000 3,800,000 78.95%
2015 1,000,0003,500,000 4,500,000 77.78%
2020 1,100,0003,750,000 4,850,000 77.32%
2025 1,250,0004,250,000 5,500,000 77.27%

Building Productivity Models for Small Enhancements
122
the next 20 years.
1.2. Research Motivation
This paper presents an empirical study of 88 small func-
tional enhancements to a software system from the core
banking ERP legacy system of a large retail bank to fig-
ure out the productivity of a maintenance process. More
specifically, the motivations of this paper are: 1) to un-
derstand the productivity of the software maintenance
process to help manage the cost of maintenance; and 2)
to understand the cost drivers that affect the software
maintenance productivity.
Benestad et al. [13] presents an overview of recent ap-
proaches by researchers for software maintenance im-
provements: maturity models to help with software main-
tenance process improvement initiatives [14,15], estima-
tion of maintenance costs [16], insights into the precon-
ditions for measuring software maintenance productivity
[14,17] and direct analysis of productivity models [4,10].
As suggested by those authors, the main hypothesis in
the design of productivity models specific to software
maintenance is that such models should distinguish be-
tween the product, resources and processes data. More
specifically, Benestad et al. reports that “Investigations
into cost drivers during software maintenance and evolu-
tion have investigated the effects of project properties
such as maintainer skills, team size, development prac-
tices, execution environment, and documentation” [13].
In this organization, IT management was interested in
understanding and updating their existing productivity
models for the following reasons:
- Productivity data on software maintenance was no
longer up to date (older than 10 years);
- Leveraging the new 2nd generation of functional size
measurement method, namely the COSMIC ISO
19761 measurement method [18] independently of the
programming language technology and with a finer
measurement granularity [19];
- A productivity model built from an up to date data
sample more representative of their current software
portfolio;
- A capability to tackle both management and custom-
ers questions about software maintenance productivity.
The rest of this paper is structured as follows. Section
2 presents the data collection and the process to build a
productivity model for software maintenance. Section 3
presents the data set of the case study. Section 4 presents
the data analysis and, Section 5 presents the conclusion
and future work.
2. The Data Collection Process
For this empirical study, the productivity model is built
in two phases:
1) A rigorous data collection process to collect reliable
data—steps 1 to 5 in Figure 1.
2) Implementing a rigorous process to design a pro-
ductivity model—steps 6 and 7 in Figure 1.
The following sub-sections explain the building pro-
cess of the productivity model as sketched in Figure 1.
2.1. Identification of Each Small Enhancement
A change request (CR) is defined in ISO 14764 [2] as
“proposed changes to a product that is being maintained”.
The data was collected through the system change re-
quest (CR) process implemented at the bank. The main-
tenance manager then looks at individual request, priori-
tizes it and, next, assigns it to a maintainer who will con-
duct the impact analysis and address it. A specific team
handlesthe maintenance requests received for one spe-
cific system module.
Figure 1. Steps to design the productivity model.

Building Productivity Models for Small Enhancements 123
Adaptive and corrective maintenance:
 The enhancements with the effort less than 3 weeks
of effort1 were selected for analysis.
 All from the same core banking application, ensuring
the same application domain (i.e., management infor-
mation system—MIS), programmed on an IBM main-
frame environment: e.g., COBOL, PL/1 or tools like
Cool:Genusing DB2 or Oracle databases.
2.2. Measurement of the Size of Each Small
Enhancement
Each 88 small enhancements, being a functional process,
represent enough data points to fulfill good statistical
conditions for the sixth criteria. These small enhance-
ments (adaptive and corrective only) were designed, pro-
grammed and implemented on the same large core bank-
ing ERP application from April 2010 to December 2011.
Each small enhancement had documented and well un-
derstood characteristics, in the change request system of
the bank, such as: the programming language used, types
(i.e. batch vs. online), and tools used—see Appendix A
for more details.
These 88 small functional enhancements were meas-
ured by the same “maintainer” and next verified by an
independent COSMIC certified expert. In this step, the
COSMIC [18] measurement method was used to deter-
mine the size of each small enhancement. To measure the
size of a small enhancement, all of its functional proc-
esses were identified, and then measured in terms of
COSMIC function points (CFP). The maintainer who
applied the measurement method was the one who ana-
lyzed, programmed and implemented each small enhan-
cement. Afterwards, an independent functional measure-
ment expert verified the number of CFP obtained for
each functional process based on the documentation of
each small enhancement (see examples in Appendix A).
Each small enhancement had a functional size of less
than twenty CFP (COSMIC Function Points).
2.3. Determination of the Quality of the
Information Provided for Each Small
Enhancement
The documentation was produced and controlled by the
same maintainer who had implemented the small en-
hancements. The quality of the documentation of each
small enhancement was assessed based on the COSMIC
verification process requirements [19]. This activity
helps verify the quality of the functional size results. The
determination of the quality of the documentation can be
assessed based on: 1) the presence or absence of a data
model; 2) the availability of the description of the data
movements; and 3) the identification of each functional
process in the software. Because the maintainer who had
done the maintenance was present during this exercise, it
was possible to complete the documentation, when need-
ed. The resulting quality of the documentation related to
small enhancements used in this case study is considered
as high (i.e., of very good quality) based on the quality
rating (Appendix A).
2.4. Collection of Effort for Each Small
Enhancement
The maintainers recorded the effort information on a
daily basis.
 Duration between half a day and 3 weeks. Effort and
duration measures were available and reliable.
 A small team of 1 or 2 maintainers executed every
small enhancement.
2.5. Collection of Characteristics of Each Small
Enhancement
The data available to analyze characteristics are: 1) func-
tional size [18]; 2) maintenance categories (adaptive and
corrective maintenance types are handled); 3) develop-
ment tool (Cool:Gen versus PL/1); 4) batch versus online
programs. Cost drivers are used in this empirical study to
analyze their influence on the effort.
2.6. Determination of Unit Cost for Each Small
Enhancement
After the maintenance requests are ordered according to
efforts, the number of functional points per hour is used
to calculate the productivity ratio while the unit cost is
determined by dividing effort (input) required to develop
small enhancements to functional size (output) of each
small enhancement.
Effort
Functional Size
2.7. Constructing Productivity Model(s) for
Small Enhancements
A productivity model is typically built with data from
CRs completed, when all information on a CR is avail-
able and that there is no more uncertainty: all of the
software functions have been delivered and all of the
number of hours for the project have been completed and
measured.
3. Presentation of the Dataset
This section presents a descriptive analysis of the dataset.
Table 3 shows the minimum, maximum, average and
standard deviation of the functional size (in number of
CFP) and effort (in person-hours) of this data set. In
1It was decided by the organization analyzed that if an enhancement
works take more than three weeks it fall in another category.

Building Productivity Models for Small Enhancements
124
Table 3. Distribution of CFP and hours.
Min. Max. Avg. Std. Dev.
CFP 3 13 5.0 2.1
Effort 1 40 9.9 8.4
Table 3, the functional size of the for small enhance-
ments varies from 3 to 13 CFP, while the effort in hours
varies from 1 to 40 hours. The average functional size of
a small enhancement is 5 CFP with a standard deviation
of 2.1 CFP. The average number of hours is 9.9 with a
standard deviation of 8.4 hours.
Table 4 presents the others variables. Those variables
are related to the programming (PL/1 or Cool:Gen), the
access type (Batch or Online) and the type of main-
tenance (adaptive or corrective). Since these variables are
descriptive, they must be transformed into dummy vari-
ables, where eachcandidate value isbinary (either 0 or 1:
presence or absence) for each of those category of vari-
able .
In Tabl e 4, the number of small enhancements pro-
gram in PL/1 is 39, and in Cool:Gen 49. The types of
access for small enhancements are respectively 39 and 49,
and the type of enhancement are 51 for adaptive and 26
for corrective. There is no information for 11 small
enhancements or they are mixed. Also, when looking at
the data (Appendix A) it shows that the PL/1 is always
associated with Batch and Cool:Gen with Online. For
this reason it is not possible to analyze those variables
independently.
4. Analysis of the Data
SPSS was used to analyze the data with regression statis-
tics. The acceptance of regression results hinges on di-
agnostic checking for the acceptance of “classical as-
sumptions” [20]. In software engineering, the classical
assumption is that the size should explain 70% of the
effort 80% of the time. For this analysis, if the regression
analysis is showing a result greater than 0.70, this will
confirm the “classical assumption” [20].
4.1. Regression Analysis Procedure
To proceed for a regression analysis it is necessary to
follow the procedure commended in [20].
A) Steps before starting the regression:
1) Define the methodology to build the regression
model for productivity (Section 2).
2) Decide which variable will be used in the model.
3) Choose the statistical tool and enter the data for sta-
tistical analysis.
4) Use descriptive analysis to find outliers. From the
descriptive statistics no outlier is present for the quantita-
tive variables (functional size and effort) in the data set
Table 4. Others variables.
Variable Category Number of enhancements
PL/1 39
Programming
language Cool:Gen 49
Batch 49
Type of programs Online 39
Adaptive 51
Type of maintenanceCorrective 26
used.
B) Steps for the regression analysis:
1) Choose which variables will be analyzed.
2) Run a regression analysis.
3) Present the results.
4) Interpret the results.
5) Accept or reject the results (in the productivity mo-
del).
6) Start again with new choices of variables.
4.2. Linear Regression Models
It is not possible to present all the results. Only some of
them will be presented following the proposed steps:
1) Choose which variables will be analyzed.
In this regression analysis, the functional size of small
enhancements is selected as the dependent variable and
effort as the independent variable in the model. This
analysis aims to find the relation between functional size
and effort: Effort = function (functional size in CFP)
2) Run a regression analysis.
SPSS was chosen to run a linear regression analysis
with the ANOVA (Table 4).
Table 5 shows the regression analysis and the
ANOVA, and Figure 2 the corresponding regression plot.
In Figure 2 the regression plot is based on 88 small en-
hancements points at the intersection of the size (CFP)
and the effort. There is no obvious outlier that can be
seen.
3) Interpret the results.
The coefficient of determination (i.e. R2) result is 0.75.
This study considers that a R2 of 0.70 is acceptable in
software engineering. For the Anova in Table 5, the last
column Sig. shows the goodness of fit of the model. If
this number is smaller than 0.01 then the model is sig-
nificant at 99%, if it is smaller than 0.05 then the model
is significant at 95%, and if it less than 0.1 the model is
significant at 90% [20]. Significance implies the accep-
tance of the model: the lower this number, the better it
fits. For the Anova in Ta ble 5, the Sig. value of 0.000
confirms the significance of the model at more than 99%.
The result is acceptable from the goodness of the fit.
The number of CFP explains more than 75% of the

Building Productivity Models for Small Enhancements 125
Table 5. Regression analysis and Anova.
Variables Entered/Removeda
Model Variables
Entered
Variables
Removed Method
1 CFPb Enter
aDependent variable: Effort; bAll requested variables entered.
Model Summary
Model R R Square
Adjusted R
Square
Std. Error of the
Estimate
1 0.870a 0.757 0.754 4.155
aPredictors: (Constant), CFP.
ANOVAa
Model Sum of
Squares df Mean Square F Sig.
Regression 1
Residual
Total
4,617,496
1,484,948
6,102,443
1
86
87
4,617,496
17,26
267,420 0,000^b
aDependent variable: Effort, bPredictors: (Constant), CFP.
variance in the effort.
4) Start again from new choices of variables.
4.3. Linear Regression Models
In the previous models, only two variables were used.
The next question is: is it possible to improve the regres-
sion value using some other variables or multiple vari-
ables?
Table 5 presents the multi-regressions results for the
three (3) independent cost drivers analyzed, together with
the independent variable CFP, and the dependent vari-
able effort.
Table 5 shows that the R2 is constantly over 0.75,
which means that independents variables (adaptive/cor-
rective, online/batch) used for regression analysis explain
more than 75% of the variance in the effort variable. The
ANOVA still have a Sig. of 0.000 for all three multi-re-
gressions and the plot is not very useful because of the
binary nature of each variable. However, none of these
additional independent variables adds much to the ex-
planation of the relationship with functional size and
effort, which already had an R2 of 0.75, thereby, the con-
tribution of these additional variables, concurrently, is
minimal.
Average Unit Cost
This research also presents the average unit costs using
a number of cost drivers.
Table 6 shows that the average unit cost for all the
data is 1.81 hours per CFP with a difference between the
lowest and highest average of 0.9 hours for 88 small en-
hancements.
Table 7 shows that that the lowest unit cost is for
Table 6. Multi-regression models.
Independent
Variables
Size & Adaptive/
Corrective
Size &
Online/Batch
Size &
PL1/Cool:Gen
R2 0.773 0.761 0.759
Table 7. Unit cost per variable.
Variables Average hours/CFP
All data 1.8
Batch 1.6
Online 2.0
Adaptive 1.4
Corrective 2.5
Adaptive maintenance (1.4 hrs/CFP), while corrective
maintenance costs almost twice as much (2.5 hrs/CFP).
For Batch and Online the average is 1.6 and 2.0 hours/
CFP respectively.
5. Conclusions and Future Works
The following five criteria were followed in this empiri-
cal study:
1) Use of an internationally recognized functional size
measurement method to measure every small functional
enhancement, that is: the COSMIC ISO 19761 measure-
ment method [17] was used to measure the functional
size of each small enhancement (functional processes
were measured and verified).
2) Assessment of the quality of the documentation
used for the sizing of each functional enhancement.
3) A controlled environment for the maintenance per-
sonnel that worked on the enhancements.
4) Implementation of a reliable effort data collection
mechanism for each small enhancement.
5) Documentation of the individual characteristics of
each small enhancements included in this empirical study.
Enough data points (individual enhancements) are col-
lected to build a valid productivity model.
Using those criteria, it was possible to produce a pro-
ductivity model with this sample using all independent
variables (R2 of more than 0.75) of a specific large retail
bank.
The productivity is 20% better using Online, instead of
Batch mode (Table 6). Unit cost is also better (almost
60% decrease) for adaptive maintenance compare to cor-
rective maintenance.
There were a number of homogeneous empirical con-
ditions to construct this productivity model (Ta bl e 4 and
Figure 2): functional enhancements to the same major
software banking application within a single organiza-
tion, each distinct functional enhancement designed,

Building Productivity Models for Small Enhancements
126
[8] M. M. Lehman, “System Maintenance and Evolution in an
Era of Reuse, COTS, and Component-Based Systems,”
International Conference on Software Maintenance (ICSM),
Oxford, 30 August 1999.
[9] M. Van Genuchten, G. Brethouwer, T. Van den Boomen
and F. J. Heemstra, “An Empirical Study of Software
Maintenance,” Information and Software Technology,
Vol. 34, No. 8, 1992, pp. 507-512.
doi:10.1016/0950-5849(92)90144-E
[10] L. B. Arfa, A. Mili and L. Sekhri, “An Empirical Study of
Software Maintenance,” Proceedings of Conference on
Software Maintenance, Sorrento, 15-17 October 1991, pp.
52-58.
[11] J. M. Desharnais, F. Pare, M. Maya and D. St-Pierre, “Im-
plementing a Measurement Program in Software Mainte-
nance: An Experience Report Based on Basili’s App-
roach,” IFPUG Spring Conference, Cincinnati, 1997.
Figure 2. Regression plot with size & effort.
programmed and implemented by the same person, docu-
mented by the maintainer, measured within a controlled
environment and verified by a measurement expert.
[12] C. Jones, “The Economics of Software Maintenance in the
Tweenty First Century,” 2006.
[13] H. C. Benestad, B. Anda and E. Arisholm, “Understanding
Software Maintenance and Evolution by Analyzing Indi-
vidual Changes: A Literature Review,” Journal of Soft-
ware Maintenance and Evolution: Research and Practice,
Vol. 21, No. 6, 2009, pp. 349-378.
doi:10.1002/smr.412
While this type of situation is common in practice,
availability of such data for empirical analysis is scarce.
On the other hand, such homogeneity limits the gener-
alization of the results to other contexts, such as different
software applications. Availability of additional data sets
is therefore necessary for further research work. [14] A. April and A. Abran, “Software Maintenance Mana-
gement: Evaluation and Continuous Improvement,” Wi-
ley-IEEE Computer Society Press, Honoken, 2008.
doi:10.1002/9780470258033 REFERENCES
[15] M. Kajko-Mattsson, “Corrective Maintenance Maturity
Model (CM3): Maintainer’s Education and Training,”
Proceedings of the 23rd International Conference on So-
ftware Engineering, Toronto, 12 May 2001, pp. 610-619.
[1] ISO/IEC 12207, Systems and Software Engineering—
Software Life Cycle Processes, International Organization
for Standardization, Geneva, 2008.
[2] ISO/IEC 14764, Software Engineering—Software Life
Cycle Processes—Maintenance, International Organiza-
tion for Standardization, Geneva, 2006.
[16] A. Abran, “Estimation Models for Software Maintenance
Based on Functional Size,” Journal of Software Tech-
nology, Vol. 9, No. 3, 2006, pp. 18-25.
[3] M. Maya, A. Abran and P. Bourque, “Measuring the Size
of Small Functional to Enhancements Software,” The 6th
International Workshop on Software Measurement, Re-
gensburg, 19-20 September 1996.
[17] A. April, A. Abran and R. R. Dumke, “Software Mainte-
nance Productivity Measurement: How to Assess the
Readiness of Your Organization, Software Maintenance
Productivity Measurement,” IWSM/Metrikon, 2004.
[4] J. Koskinen, “Software Maintenance Costs,” University of
Jykäskylä, Finland, 2010. [18] Measurement Manual v3.0.1 (The COSMIC Implementa-
tion Guide for ISO/IEC 19761: 2003), 2009, The Com-
mon Software Measurement International Consortium
(COSMIC), 2012.
[5] M. Torchiano, F. Ricca and A. De Lucia, “Empirical
Studies in Software Maintenance and Evolution,” IEEE
International Conference on Software Maintenance, Paris,
2-5 October 2007, pp. 491-494. [19] The COSMIC Functional Size Measurement Method Ver-
sion 3.0.1 Guideline for Assuring the Accuracy of Meas-
urements Version 0.92, Common Software Measurement
International Consortium, 2011.
[6] U. Kuhlmann, “Maintenance Activities in Software Proc-
ess Models: Theory and Case Study Practice,” Master
Thesis, University of Koblenz Landau, Koblenz, 2003, pp.
1-135. [20] Regression Explained in Simpler Terms, A Vijay Gupta
Publication, SPSS for Beginners, 2000.
https://mywebspace.wisc.edu/rlbrown3/web/library/regres
sion_explained.pdf
[7] A. Abran and H. Nguyenkim, “Analysis of Maintenance
Work Categories through Measurement,” IEEE Interna-
tional Conference on Software Maintenance, Sorrento,
15-17 October 1991, pp. 104-113.

Building Productivity Models for Small Enhancements 127
Appendix A. List of the Small Enhancements with Measures
No Identification E R W XCFPEffortTypeRatioToolQualityBatch Online Write ExitModule
1 Changing General Ledger
Numbers of Retail Credit 1 1 1 1 44 M 1.00PL/IA Y N 0 0 1
2 Changing General Ledger
Numbers of Commercial Credits1 1 1 1 44 M 1.00PL/IA Y N 0 0 1
3 Changing General Ledger
Numbers of Overdue Credits 1 1 1 1 44 M 1.00PL/IA Y N 0 0 1
4 Calculation of Effective
Interest Rate of Retail Credits 1 2 2 1 622 N 3.67PL/IA Y N 1 0 1
5 Calculating Adjustment Amount of
Retail Credits 1 2 2 1 620 N 3.33PL/IA Y N 1 0 1
6 Daily Adjustment Accounting 1 5 1 1 818 N 2.25PL/IA Y N 0 0 0
7
Changing Commission and
Income General Numbers of
Amortization Batch
1 1 0 1 34 M 1.33PL/IA Y N 0 0 0
8
Transfer of Opening Commissions
related to Commercial Credits with
payment plan number to the
income system
1 1 1 1 44 M 1.00Cool:
Gen A N Y 0 0 0
9
Transfer of Opening Commissions
related to Commercial Credits with
no payment plan number to the
income system (PL/I batch is used)
1 1 0 1 33 M 1.00PL/IA Y N 0 0 0
10
Cancel amortization of opening
commissions related to commercial
credits with payment plan number
to the income system
1 1 1 1 44 M 1.00Cool:
Gen A N Y 0 0 0
11
Transfer of Periodic Commissions
related to Commercial Credits with
payment plan number to the
income system
1 1 0 1 34 M 1.33Cool:
Gen A N Y 0 0 0
12
Cancel amortization of periodic
commissions related to
commercial credits with payment
plan number to the income system
1 1 1 1 44 M 1.00PL/IA Y N 0 0 0
13
Transfer of Opening Commissions
related to Retail Credits with
payment plan number to the
income system
1 1 1 1 44 M 1.00Cool:
Gen A N Y 0 0 0
14
Cancel amortization of opening
commissions related to retail
credits with payment plan number
to the income system
1 1 1 1 44 M 1.00Cool:
Gen A N Y 0 0 0
15
Transfer of Periodic Commissions
related to Retail Credits with
payment plan to the income system
1 1 1 1 44 M 1.00PL/IA Y N 0 0 0
16
Cancel amortization of periodic
commissions related to commercial
retail income system
1 1 1 1 44 M 1.00PL/IA Y N 0 0 0
17 Commercial Credits Commission
Rediscount Create 1 1 1 1 42 M 0.50PL/IA Y N 0 0 0
18 Commercial Credits Commission
Rediscount Function 2 2 2 1 710 N 1.43PL/IA Y N 1 0 0
19
Calling IFRS general link to
calculate EIR, Adjustment and
Effective amount in Commercial
Credits Investment Rediscount
1 2 2 1 610 N 1.67Cool:
Gen A N Y 1 0 0
20
Calling IFRS general link to
calculate EIR, Adjustment and
Effective amount in Retail Credits
Investment Rediscount
1 2 2 1 610 N 1.67Cool:
Gen A N Y 1 0 0

Building Productivity Models for Small Enhancements
128
No Identification E R W X CFPEffortTypeRatioTool QualityBatch Online Write ExitModule
21
Calling IFRS general link to
calculate EIR, Adjustment and
Effective amount in Retail
Credits Commission Rediscount
1 2 2 1 6 8 N 1.33Cool:GenA N Y 1 0 0
22
Cancel of calling amortization
create from periodic commission
collection of retail credits
1 1 1 1 4 3 M 0.75PL/I A Y N 0 0 0
23
Cancel of calling amortization
create from opening commission
collection of retail credits
1 1 1 1 4 3 M 0.75Cool:GenA N Y 0 0 0
24
Cancel of calling amortization
create from periodic commission
collection of commercial credits
1 1 1 1 4 3 M 0.75PL/I A Y N 0 0 0
25
Calling income system link from
periodic commission collection
of retail credits
1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
26
Calling income system link from
opening commission collection
of retail credits
1 1 1 1 4 4 M 1.00Cool:GenA N Y 0 0 0
27
Calling income system link from
periodic commission collection
of commercial credit
1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
28 Cancel of calling amortization
create from accrual system 1 1 1 1 4 3 M 0.75PL/I A Y N 0 0 0
29
Calling commercial credits
commission rediscount create
link from accrual system
1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
30
Connection between income
system and amortization system
for opening commission records
of commercial credits with no
payment plan
1 0 1 1 3 4 M 1.33Cool:GenA N Y 0 0 1
31 Extra Commission Collection
Facility for Commercial Credits1 6 5 1 13 35 N 2.69Cool:GenA N Y 1 0 1
32 Extra Commission Collection
Facility for Retail Credits 1 5 4 1 11 28 N 2.55Cool:GenA N Y 1 0 1
33
Discarding new Transactions
from the First Level of
Accounting Unload
1 1 1 3 1 M 0.33PL/I A Y N 0 0 0
34
Adding new Transactions to the
Second Level of Accounting
Unload
1 1 1 3 1 M 0.33PL/I A Y N 0 0 0
35
Loading accounting from excel
file to the system without using
general parametric accounting
link
1 3 1 1 6 14 N 2.33PL/I A Y N 0 0 1
36
Comparison of accounting
movement and account
movement to control trial
balance (in two ways: from
accounting to account movement
and from account movement to
accounting movement)
1 2 1 4 12 M 3,00PL/I A Y N 0 0 1
37 Facility to insert general numbers
according to product 1 3 1 1 6 10 N 1.67Cool:GenA N Y 0 0 0
38
Facility to insert old product
codes and new products codes
into product change parameter
table
1 2 1 1 5 6 M 1.20Cool:GenA N Y 0 0 0
39 NBR Opening Commission
Report of Retail Credits 1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
40 NBR Periodic Commission
Report of Retail Credits 1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
41 NBR Extra Commission Report
of Retail Credits 1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
42 NBR Opening Commission
Report of Commercial Credits 1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0

Building Productivity Models for Small Enhancements 129
No Identification E R W X CFPEffortTypeRatio Tool QualityBatch Online Write ExitModule
43 NBR Periodic Commission
Report of Commercial Credits1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
44 NBR Extra Commission Report
of Commercial Credits 1 1 1 1 4 4 M 1.00PL/I A Y N 0 0 0
45 Facility to Relate Income Codes
with Product Codes 1 1 1 1 4 4 M 1.00Cool:GenA N Y 0 0 0
46 List of Income Codes Related to
Product Codes 1 1 1 1 4 8 M 2.0 Cool:GenA N Y 0 0 0
47 Impair Flag Update for Risky
Credits 1 5 2 1 9 16 N 1.78PL/I A Y N 1 0 1
48 Calculating Effective Interest
Amount of Commercial Credits1 6 2 1 1026 N 2.60PL/I A Y N 1 0 1
49
Calculation Effective Interest
Rate(EIR) of Commercial
Credits
1 2 2 1 6 12 N 2.00PL/I A Y N 1 0 1
50 Calculating Adjustment Amount
of Commercial Credits 1 2 2 1 6 13 N 2.17PL/I A Y N 1 0 1
51 Customer Account Report
General Ledger Number Change1 1 1 3 1 M 0.33PL/I A Y N 0 0 0
52 Recover Table Lock Escalation
Problem 1 1 1 1 4 10 M 2.50PL/I A Y N 0 0 0
53 Income Delivery Operations 1 4 3 2 1040 N 4.00PL/I A Y N 1 0 1
54 Pricing Service List 1 1 2 4 16 M 4.00Cool:GenA N Y 0 0 0
55 Service Definition Operation 1 2 1 3 7 16 N 2.29Cool:GenA N Y 0 1 0
56 Record Priority Screen 1 1 1 1 4 8 M 2.00Cool:GenA N Y 0 0 0
57 Pricing Service code Product
Code Relation 1 1 1 1 4 8 M 2.00Cool:GenA N Y 0 0 0
58 Service Pricing 3 1 3 1 8 24 N 3.00Cool:GenA N Y 1 0 0
59 Business of Industry Price
Detail Entry 2 3 1 2 8 22 N 2.75Cool:GenA N Y 0 0 0
60 Reference Price Detail Entry 2 3 1 2 8 18 N 2.25Cool:GenA N Y 0 0 0
61 Special Price Detail Entry 2 3 1 2 8 19 N 2.38Cool:GenA N Y 0 0 0
62 Charge Commission Amount
Querying 2 6 4 1238 N 3.17Cool:GenA N Y 0 1 1
63 Calculation of Profitability of
Customer for Company 1 2 2 5 20 M 4.00PL/I A Y N 0 0 1
64 Extract of Account Report 1 3 2 6 24 N 4.00PL/I A Y N 0 0 1
65 Revenue List 1 3 2 6 22 N 3.67PL/I A Y N 0 0 1
66 Campaign Parameter Entry
Screen Change 1 1 1 2 5 8 M 1.60Cool:GenA N Y 0 0 0
67 Closing Accounts According to
Criteria 1 4 1 2 8 24 N 3.00PL/I A Y N 0 0 1
68 General Ledger Number Update1 3 1 2 7 11 N 1.57PL/I A Y N 0 0 0
69
Adding New Fields to the
Campaign Parameter Entry
Screen
1 1 1 1 4 10 M 2.5 PL/I A Y Y 0 0 0
70
Adding New Fields to the
Campaign Product Parameter
Entry Screen
1 1 1 3 8 M 2.67PL/I A Y Y 0 0 0
71 Customer Report Interest
Amount Setting 1 2 1 4 6 M 1.5 PL/I A Y N 0 0 0
72 Adding new Accounting Case
To the Income Return Process1 2 1 1 5 10 M 2 Cool:GenA N Y 0 0 1
73 Retail Credits Interest Discount
Control Link 1 1 1 3 8 M 2.67Cool:GenA N Y 0 0 0
74 Customer Transfer Check Link1 1 1 3 8 M 2.67Cool:GenA N Y 0 0 0
75 Cost Matrix Special Price
Definition Log List 1 1 2 4 6 M 1.5 PL/I A Y N 0 0 0

Building Productivity Models for Small Enhancements
130
No Identification E R W X CFPEffortTypeRatioTool Quality Batch Online Write Exit Module
76 Accounting Transaction Group
List Service 1 1 1 3 8 M 2.67Cool:GenA N Y 0 0 0
77 Account Plan Service 1 1 1 3 8 M 2.67Cool:GenA N Y 0 0 0
78 Accounting Transaction List
Performance Enhancement 1 1 1 3 7 M 2.33Cool:GenA N Y 0 0 0
79
Simulation facility to calculate
EIR, adjustment and effective
amount
1 1 2 4 8 2 Cool:GenA N Y 0 0 0
80 Valuable Fund Tax Transfer 1 3 1 1 6 10 1.67PL/I A Y N 0 0 1
81 General Ledger Number
Update for Unit 1 2 1 4 4 1 PL/I A Y N 0 0 0
82 Credit Read Service 1 1 1 3 8 2.67Cool:GenA N Y 0 0 0
83 Retail Credits Master
Information Read Service 1 1 1 3 8 2.67Cool:GenA N Y 0 0 0
84 Parametric Accounting Detail
Read Service 1 2 2 5 10 2 Cool:GenA N Y 0 0 0
85 Rediscount Information Update
Service 1 1 1 1 4 8 2 Cool:GenA N Y 0 0 0
86 Amortization Information
Update Service 1 1 1 1 4 8 2 Cool:GenA N Y 0 0 0
87 Commission Report new fields
request for Retail Credits 1 1 2 4 3 0.75PL/I A Y N 0 0 0
88 Commission Report new fields
request for Commercial Credits1 2 2 5 5 1 PL/I A Y N 0 0 0
Appendix B. Small Enhancements Documentation (2 Examples)
1) Changing General Ledger Numbers of Retail Credits
There is a ledger number for each product code. If the product code is changed, ledger number must be changed.
Trigger: Product code change
Entry: Product Code
Read: General Ledger (1 Read)
Write: General Ledger (1Write)
Error Message: 1 Exit (General Ledger number is not defined related to product code)
1(E) + 1(R) + 1(W) + 1(X) = 4CFP
Quality: A
Effort Enh. = 4 hours
Data Group: General Ledger
2) Changing General Ledger Numbers of Commercial Credits
There is a ledger number for each product code. If the product code is changed, ledger number must be changed.
Trigger: Product code change
Entry: Product Code
Read: General Ledger (1 Read)
Write: General Ledger (1Write)
Error Message: 1 Exit (General Ledger number is not defined related to product code)
1(E) + 1(R) + 1(W) + 1(X) = 4CFP
Quality: A
Effort Enh. = 4 hours
Data Group: General Ledger