Chapter 21

Software quality metrics

1
“You can’t control what you can’t measure”
Tom DeMarco

Software Quality metrics as tool for QA tools

2
 Measurement, Measures, Metrics
 Measurement
◦ is the act of obtaining a measure

 Measure
◦ provides a quantitative indication of the size of some product or
process attribute, E.g., Number of errors

 Metric
◦ is a quantitative measure of the degree to which a system,
component, or process possesses a given attribute (IEEE Software
Engineering Standards 1993) : Software Quality - E.g., Number of
errors found per person hours expended

3
 IEEE definitions of
software quality metrics
A quantitative measure of the degree to )1(
which an item possesses a given quality
.attribute
A function whose inputs are software )2(
data and whose output is a single
numerical value that can be interpreted as
the degree to which the software
.possesses a given quality attribute

4
 Objectives of quality measurement
1. Facilitate management control, planning and
managerial intervention.
Based on:
        ·   Deviations of actual from planned performance.
        ·   Deviations of actual timetable and budget
performance from planned.
2. Identify situations for development or maintenance
process improvement (preventive or corrective
actions). Based on:
        ·  Accumulation of metrics information regarding the
performance of teams, units, etc.

5
 Software quality
metrics — Requirements
General requirements
 Relevant
 Valid
 Reliable
 Comprehensive
 Mutually exclusive
Operative requirements
 Easy and simple
 Does not require independent data collection
 Immune to biased interventions by interested parties

6
 What to measure
• Process
Measure the efficacy of processes. What works, what
doesn't.
• Project
Assess the status of projects. Track risk. Identify problem
areas. Adjust work flow.
• Product
Measure predefined product attributes (generally related
to ISO9126 Software Characteristics)

7
 Classification of software quality
metrics
 Three kinds of Software Quality Metrics
◦ Product Metrics - describe the characteristics of product
 size, complexity, design features, performance, and quality level

◦ Process Metrics - used for improving software

development/maintenance process
 effectiveness of defect removal, pattern of testing defect arrival,
and response time of fixes

◦ Project Metrics - describe the project characteristics and

execution
 number of developers, cost, schedule, productivity, etc.
 fairly straight forward

8
 Software size (volume) measures
 KLOC — classic metric that measures the
size of software by thousands of code lines.
 Number of function points (NFP) — a
measure of the development resources
(human resources) required to develop a
program, based on the functionality
specified for the software system

9
 Process metrics
 S/W development process metrics fall into
one of the following categories:
1-Software process quality metrics
 Error density metrics
 Error severity metrics
 Error removal effectiveness metrics
2- Software process timetable metrics
3- Software process productivity metrics

10
Software process quality metrics
 Software process quality metrics may be
classified into two classes:
 Error density metrics
 Error severity metrics
 Error removal effectiveness metrics

11
 Error density metrics
 Calculation of error density metrics involves two
measures:
1- Software volume measures. Some density metrics use
the number of lines of code while others apply function
points.
2- Errors counted measures. Some relate to the number
of errors(NCE) and others to the weighted number of
errors (WCE).
 this measure is classification of the detected errors into
severity classes, followed by weighting each class.

12
 Example
 A software development department applies two alternative
measures, NCE and WCE, to the code errors detected in
its software development projects. Three classes of error
severity and their relative weights are also defined:
 there were 42 low severity errors, 17 medium severity
errors, and 11 high severity errors

13
Calculation of NCE and WCE

14
 Calculation formula Name Code
NCE
CED = ----------- Code Error Density CED
KLOC
NDE
DED = ----------- Development Error Density DED
KLOC
WCE
WCDE = --------- Weighted Code Error Density WCED
KLOC
WDE Weighted Development Error Density WDED
WDED = ---------
KLOC
WCE Weighted Code Errors per Function
WCEF = ---------- Point
WCEF
NFP
WDE Weighted Development Errors per
WDEF = ---------- Function Point
WDEF
NFP
NCE = The number of code errors detected by code inspections and testing.
NDE = total number of development (design and code) errors) detected in the development
process.
WCE = weighted total code errors detected by code inspections and testing.
WDE = total weighted development (design and code) errors detected in development process.
 Calculation Of
CED and WCED Example
•The unit determined the following indicators for unacceptable software
quality: CED > 2 and WCED > 4.
•The software system size is 40 KLOC. Calculation of the two metrics
resulted in the following:
Error severity metrics
 The metrics belonging to this group are
used to detect adverse situations of
increasing numbers of severe errors in
situations where errors and weighted
errors, as measured by error density
metrics, are generally decreasing.
 Two error severity metrics are presented as
follows:
Calculation formula Name Code
WCE Average Severity of Code
ASCE = ----------- ASCE
NCE Errors
WDE Average Severity of
ASDE = ----------- ASDE
NDE Development Errors

NCE = The number of code errors detected by code inspections and testing.
NDE = total number of development (design and code) errors detected in
the development process.
WCE = weighted total code errors detected by code inspections and testing.
WDE = total weighted development (design and code) errors detected in
development process.
 Error removal
effectiveness metrics
• Software developers can measure the effectiveness of error
removal by the software quality assurance system after a
period of regular operation (usually 6 or 12 months) of the
system.
• The metrics combine the error records of the development
stage with the failures records compiled during the first year
(or any defined period) of regular operation. Two error
removal effectiveness metrics are as follows.
Calculation formula Name Code
NDE Development Errors Removal DERE
DERE = ----------------
NDE + NYF Effectiveness
WDE
DWERE = ------------------ Development Weighted DWERE
WDE+WYF Errors Removal Effectiveness

NDE = total number of development (design and code) errors) detected in the
development process.
WCE = weighted total code errors detected by code inspections and testing.
WDE = total weighted development (design and code) errors detected in
development process.
NYF = number software failures detected during a year of maintenance service.
WYF = weighted number of software failures detected during a year of maintenance
service.
 Software process
timetable metrics
• Software process timetable metrics may be based on
accounts of success (completion of milestones per schedule)
in addition to failure events (non-completion per schedule).
•An alternative approach calculates the average delay in
completion of milestones
•TTO and ADMC metrics are based on data for all relevant
milestones scheduled in the project plan. In other words, only
milestones that were designated for completion in the project
plan stage are considered in the metrics’ computation.
Therefore, these metrics can be applied throughout
development and need not wait for the project’s completion.
Calculation formula Name Code
MSOT Time Table Observance
TTO = ----------- TTO
MS
TCDAM Average Delay of Milestone
ADMC = ----------- ADMC
MS Completion

MSOT = Milestones completed on time.

MS = Total number of milestones.
TCDAM = Total Completion Delays (days, weeks, etc.) for all milestones.
Software process productivity metrics
 This group of metrics includes “direct”
metrics that deal with a project’s human
resources productivity as well as “indirect”
metrics that focus on the extent of software
reuse.
 Software reuse substantially affects
productivity and effectiveness.
 Calculation Name Code
formula
DevH
---------- = DevP Development Productivity DevP
KLOC
DevH Function point Development FDevP
---------- = FDevP Productivity
NFP
ReKLOC
-------------- = Cre Code Reuse CRe
KLOC
ReDoc
----------- = DocRe Documentation Reuse DocRe
NDoc
DevH = Total working hours invested in the development of the software system.
ReKLOC = Number of thousands of reused lines of code.
ReDoc = Number of reused pages of documentation.
NDoc = Number of pages of documentation.
 Product metrics
 Product metrics refer to the software system’s operational
phase – years of regular use of the software system by
customers, whether “internal” or “external” customers,
 In most cases, the software developer is required to provide
customer service during the software’s operational phase.
Customer services are of two main types:
1-Help desk services (HD) – software support by instructing
customers regarding the method of application of the software
and solution of customer implementation problems.
2- Corrective maintenance services – correction of software
failures identified by customers/users or detected by the
customer service team prior to their discovery by customers. The
number of software failures and their density are directly related
to software development quality.
 Product metrics
 HD metrics are based on all customer calls while corrective
maintenance metrics are based on failure reports.
 Product metrics generally rely on performance records compiled
during one year (or any other specified period of time). This policy
enables comparisons of successive years in addition to comparisons
between different units and software systems.
 The array of software product metrics presented here is
classified as follows:
■ HD quality metrics
■ HD productivity and effectiveness metrics
■ Corrective maintenance quality metrics
■ Corrective maintenance productivity and effectiveness metrics.
 Product metrics
 It should be remembered that software maintenance activities
include:
■ Corrective maintenance – correction of software failures
detected during regular operation of the software.
■ Adaptive maintenance – adaptation of existing software to new
customers or new requirements.
■ Functional improvement maintenance – addition of new
functions to the existing software, improvement of reliability, etc.
 In the metrics presented here we limit our selection to those that
deal with corrective maintenance. For other components of
software maintenance, the metrics suggested for the software
development process (process metrics) can be used as is or
with minor adaptations.
 HD quality metrics
 The types of HD quality metrics discussed here
deal with:
■ HD calls density metrics – the extent of customer
requests for HD services as measured by the
number of calls.
■ Metrics of the severity of the HD issues raised.
■ HD success metrics – the level of success in
responding to these calls. A success is achieved
by completing the required service within the time
determined in the service contract.
HD calls density metrics
 For the HD call density there are six different
types of metrics. Some relate to the number of the
errors and others to a weighted number of errors.
As for size/volume measures of the software,
some use number of lines of code while others
apply function points.
 The sources of data for these and the other
metrics in this group are HD reports. Three HD
calls density metrics for HD performance are
presented as follows:
Calculation Formula Name Code
NHYC
-------------- = HDD HD calls density HDD
KLMC
WHYC
------------ = WHYC Weighted HD calls density WHDD
KLMC
WHYC Weighted HD calls per WHDF
------------ = WHDF function point
NMFP

NHYC = the number of HD calls during a year of service.

KLMC = Thousands of lines of maintained software code.
WHYC = weighted HD calls received during one year of service.
NMFP = number of function points to be maintained.
Severity of HD calls metrics
 The metrics belonging to this group of measures
aim at detecting one type of adverse situation:
increasingly severe HD calls.
 The computed results may contribute to
improvements in all or parts of the user interface
(its “user friendliness”) as well as the user manual
and integrated help menus.
 The Average Severity of HD Calls (ASHC):
refers to failures detected during a period of one
year.
Calculation Formula Name Code
WHYC
-------------- = ASHC Average severity of HD calls ASHC
NHYC

NHYC = the number of HD calls during a year of service.

WHYC = weighted HD calls received during one year of service.
Success of the HD services
 The most common metric for the success of HD services is the
capacity to solve problems raised by customer calls within the
time determined in the service contract (availability). Thus, the
metric for success of HD services compares the actual with the
designated time for provision of these services.
 For example, the availability of help desk (HD) services for an
inventory management software package is defined as follows:
■ The HD service undertakes to solve any HD call within one
hour.
■ The probability that HD call solution time exceeds one hour will
not exceed 2%.
■ The probability that HD call solution time exceeds four working
hours will not exceed 0.5%.
 Calculation Formula Name Code
NHYOT
-------------- = HDS HD service success HDS
NHYC

NHYNOT = Number of yearly HD calls completed on time during one year of service.
NHYC = the number of HD calls during a year of service.
HD productivity and effectiveness
metrics
 Productivity metrics relate to the total of
resources invested during a specified
period, while effectiveness metrics relate to
the resources invested in responding to a
HD customer call.
HD productivity metrics
 HD productivity metrics makes use of the easy-to-
apply KLMC measure of maintained software
system’s size or according to function point
evaluation of the software system. Two HD
productivity metrics are presented as follows.
 HD effectiveness metrics
 The metrics in this group refer to the resources
invested in responding to customers’ HD calls.
Calculation Formula Name Code
HDYH
-------------- =HDP HD Productivity HDP
KLMC
HDYH Function Point HD FHDP
---------- = FHDP Productivity
NMFP
HDYH
-------------- = HDE HD effectiveness HDE
NHYC

HDYH = Total yearly working hours invested in HD servicing of the software system.
KLMC = Thousands of lines of maintained software code.
NMFP = number of function points to be maintained.
NHYC = the number of HD calls during a year of service.
Corrective maintenance quality
metrics
 Software corrective maintenance metrics deal
with several aspects of the quality of maintenance
services. A distinction is needed between
software system failures treated by the
maintenance teams and failures of the
maintenance service that refer to cases where the
maintenance failed to provide a repair that meets
the designated standards or contract
requirements. Thus, software maintenance
metrics are classified as follows:
Corrective maintenance quality
metrics
■ Software system failures density metrics – deal with the extent of
demand for corrective maintenance, based on the records of
failures identified during regular operation of the software system.
■ Software system failures severity metrics – deal with the severity
of software system failures attended to by the corrective
maintenance team.
■ Failures of maintenance services metrics – deal with cases
where maintenance services were unable to complete the failure
correction on time or that the correction performed failed.
■ Software system availability metrics – deal with the extent of
disturbances caused to the customer as realized by periods of time
where the services of the software system are unavailable or only
partly available.
Calculation Formula Name Code
NYF Software System Failure Density SSFD
SSFD = --------------
KLMC
WYF Weighted Software System Failure WSSFD
WFFFD = --------- Density
KLMC
WYF Weighted Software System Failures WSSFF
WSSFF = ---------- per Function point
NMFP

NYF = number of software failures detected during a year of maintenance service.

WYF = weighted number of yearly software failures detected during one year of
maintenance service.
NMFP = number of function points designated for the maintained software.
KLMC = Thousands of lines of maintained software code.
 Calculation Formula Name Code
WYF Average Severity of ASSS
-------------- = ASSSF Software System Failures
NYF F

NYF = number of software failures detected during a year of maintenance service.

WYF = weighted number of yearly software failures detected during one year.
 Calculation Formula Name Code
RepYF Maintenance Repeated MRepF
-------------- = MRepF - repair Failure metric
NYF

         NYF = number of software failures detected during a year of maintenance

service.
RepYF = Number of repeated software failure calls (service failures).
Software system availability metrics
 User metrics distinguish between:
■ Full availability – where all software system
functions perform properly
■ Vital availability – where no vital functions
fail (but non-vital functions may fail)
■ Total unavailability – where all software
system functions fail.
 Calculation Formula Name Code
NYSerH - NYFH
----------------------- = FA Full Availability FA
NYSerH
NYSerH - NYVitFH
----------------------------- = VitA Vital Availability VitA
NYSerH
NYTFH
------------ = TUA Total Unavailability TUA
NYSerH

 NYSerH = Number of hours software system is in service during one year.  

NYFH = Number of hours where at least one function is unavailable (failed) during one year,
including total failure of the software system.
 NYVitFH = Number of hours when at least one vital function is unavailable (failed) during
one year, including total failure of the software system.
 NYTFH = Number of hours of total failure (all system functions failed) during one year.
NYFH ≥ NYVitFH ≥ NYTFH.
1 – TUA ≥ VitA ≥FA
Limitations of software metrics
 Application of quality metrics is strewn with
obstacles. These can be grouped as follows:
■ Budget constraints in allocating the necessary
resources (manpower, funds, etc.) for
development of a quality metrics system and its
regular application.
■ Human factors, especially opposition of employees
to evaluation of their activities.
■ Uncertainty regarding the data’s validity, rooted in
partial and biased reporting.