ISO-BASED MODELS TO MEASURE SOFTWARE

PRODUCT QUALITY

Alain Abran1, Rafa E. Al-Qutaish1,

Jean-Marc Desharnais1 and Naji Habra2
1
École de Technologie Supérieure, 1100 Notre-Dame Ouest, Montréal, Québec H3C1K3, Canada
2
Institut d'Informatique, Universitaire de Namur, Rue Grandgagnage 21, 5000 Namur, Belgium
[email protected], [email protected],
[email protected], [email protected]

ABSTRACT
The ISO is developing a new ISO 25000 series on Software Product Quality Requirements
and Evaluation (SQuaRE) to improve the interpretation and use of quality measures for
software products. This chapter explains how the ISO 19539 Measurement Information Model
can be used to implement the ISO 9126 models for software product quality. It also identifies
some of the harmonization issues arising as a result of the addition of new documents like ISO
25020 and ISO 25021, in particular with respect to previously published measurement
standards for software engineering.

KEYWORDS: ISO 25020, ISO 25021, ISO 9126, ISO 15939, ISO 25000-SQuaRE, Software
Measurement, Software Product Quality.

1. INTRODUCTION
In 1991, the ISO published its first international consensus on terminology for the quality
characteristics for software product evaluation (ISO 9126:1991) [1]. During the period 2001 to
2004, the ISO published an expanded version, containing the ISO quality models and a
consensus on inventories of proposed measures for these models. This version of the ISO 9126
series consists of one International Standard (IS) [2] and three Technical Reports (TR) [3-5]:
- ISO IS 9126-1: Quality Models
- ISO TR 9126-2: External Metrics1
- ISO TR 9126-3: Internal Metrics
- ISO TR 9126-4: Quality in Use Metrics

The ISO has recognized a need for further enhancement of ISO 9126, primarily as a result
of advances in the information technologies (IT field) and changes in the IT environment.
Consequently, the ISO is now working on the next generation of software product quality
standards, which will be referred to as Software Product Quality Requirements and Evaluation
(SQuaRE – ISO 25000). Once completed, this series of standards will replace the current ISO
9126 and ISO 14598 series of standards. The SQuaRE series is made up of five divisions:
- Quality management division (ISO 2500n)
- Quality model division (ISO 2501n)
1
The term ‘metrics’ used in ISO 9126 is replaced by ‘measures’ in the new series of standards, in
accordance with ISO 15939.

1/29
 - Quality measurement division (ISO 2502n)
- Quality requirements division (ISO 2503n)
- Quality evaluation division (ISO 2504n)

One of the main objectives of the SQuaRE series and the current ISO 9126 series (which
also constitutes the difference between them) is the coordination and harmonization of its
contents with the ISO standard on software measurement process – ISO 15939 [6, 7]. In
particular, the quality measurement division (ISO 2502n) will consist of the following five
documents that are scheduled to replace the current four-part ISO 9126 series:
- ISO 25020: Measurement Reference Model and Guide [6]
- ISO 25021: Quality Measure Elements [7]
- ISO 25022: Measurement of Internal Quality
- ISO 25023: Measurement of External Quality
- ISO 25024: Measurement of Quality in Use

This chapter presents some of the harmonization issues arising with respect to previously
published measurement standards for software engineering, including ISO 15939, ISO 14143-
1 and ISO 19761, and proposes ways to address them using the measurement information
model of ISO 15939 on software measurement process.

This chapter is organized as follows: Section 2 presents the ISO Measurement Information
Model adopted in ISO 15939. Section 3 presents our solution for alignment of the ISO models
of software product quality with the measurement information model of ISO 15939. Section 4
presents next the outstanding harmonization issues in terminology and coverage in ISO FDIS
25020 and ISO 25021. Finally, a discussion and a summary are presented in section 5.

2.ISO 15939 MEASUREMENT INFORMATION MODEL

The ISO has produced an information model within ISO 15939 (2002) (see Figure 1) to help
in determining what must be specified during measurement planning, performance and
evaluation.

Figure 1 shows that a specific measurement method is used to collect a base measure for a
specific attribute. The values of two or more base measures can then be used in a
computational formula (by means of a measurement function) to produce and construct a
specific derived measure. These derived measures are in turn used in an analysis model to
arrive at an indicator, which is a value, and to interpret the indicator’s value in order to explain
the relationship between it and the information needed, doing so in the language of the
measurement user, to produce an Information Product for the user’s Information Needs.
 Information Product
Information Needs

Interpretation

Indicator

(Analysis) Model

Derived Measures Derived Measures

Measurable Concept
Measurement Function

Base Measures Base Measures

Measurement Method Measurement Method

Entity Attribute Attribute

Figure 1: Measurement Information Model from ISO 15939 (2002) 2

There already exists a very mature measurement terminology which is well documented in
the ISO International Vocabulary of Basic and General Terms in Metrology (ISO VIM) [8].
This terminology is widely accepted and used in most fields of science, and has been adopted
in ISO 15939 [9] as the agreed-upon measurement terminology for software and system
engineering-related ISO standards.

3.MAPPING THE QUALITY MODEL TO THE MEASUREMENT INFORMATION MODEL

The following two expressions come from the ISO standard on software measurement
process, ISO 15939 [9], which is itself based on the definitions in the ISO International
Vocabulary of Basic and General Terms in Metrology (VIM 1993) [8]:
- Base measure: a measure defined in terms of an attribute and the method for
quantifying it. A base measure is functionally independent of other measures.
- Derived measure: a measure defined as a function of two or more values of base
measures. A transformation of a base measure using a mathematical function can
also be considered as a derived measure.
2
We added the arrows to the ISO 15939 Measurement Information Model to point up the dataflow
relationships. Ovals represent activities, and rectangles represent the input and output of an activity.
 In practice, the data collection associated with a property of an object (or concept), and
quantification of that property, occurs at the base measure level, at which time a measurement
unit is assigned based on the rules of the measurement method used for the quantification.

At the derived measure level, the base measures have been already collected and are being
assembled according to the rules of combination (e.g. a computational formula) defined within
each derived measure. A derived measured is therefore the product of a set of properly
combined measurement units (through a measurement function). This combination is then
labelled to represent an attribute (of a quality characteristic or subcharacteristic) of a software
product.

Table 1 shows examples of base measures used in the measure definitions documented in
ISO 9126-2, -3 and -4 (see Table A1 in Appendix A for the complete list of base measures).
Table 1 shows the name of each base measure and the unit of measurement that is assigned to
its value. These base measures can be used to calculate each of the derived measures (akin to
metrics) in ISO 9126-4.

Table 1: Examples of Base Measures in ISO 9126-4.

Quality in Use Base Measures
Measure Name Unit of Measurement
1 Task Effectiveness (a given weight)
2 Total Number of Tasks Task (number of)
3 Task Time Minute
4 Cost of the Task Dollar
5 Help Time Second
6 Error Time Second
7 Search Time Second
8 Number of Users User (number of)
9 Total Number of People Potentially Affected by the System Person (number of)
10 Total Number of Usage Situations Situation (number of)

Each of these base measures must be collected individually. They can be used at least
once, or multiple times, to obtain the derived measure required to quantify the software
properties specified in the ISO 9126 quality model. Table 2 provides an example of where
some base measures are used throughout ISO 9126-2. For instance, the base measure ‘number
of inaccurate computations encountered by users’ is used only once in ‘external functionality –
accuracy measures’, while the base measure ‘number of items requiring compliance’ can be
used in 6 subcharacteristics of external quality (ISO 9126-2). The construction of derived
measures is based on a computational formula consisting of two or more base measures (see
Table A2 in Appendix A for the complete cross-reference of base measure usage).
 Table 2: Examples of the Use of Base Measures in ISO 9126-2.
External

Measure Name Functionality Reliability Usability Efficiency Maintainability Portability

F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U E1 E2 E3 M M2 M M4 M P1 P2 P3 P4 P5
5 1 3 5
1 Number of functions ! ! ! ! ! !

2 Operation time ! ! ! ! ! ! ! ! ! ! ! ! ! !
Number of inaccurate
3 computations !
encountered by users
4 Number of data formats !

Number of illegal
5 !
operations

Number of items
6 ! ! ! ! ! !
requiring compliance

Number of interfaces
7 !
requiring compliance

8 Number of faults ! ! !

Such lists of base measures and usage cross-references are currently missing from ISO
9126 and would be helpful to those designing programs for implementing measurement of the
quality of software products using ISO 9126 quality models and related measures. In
particular, these lists can help in:
- Identifying, selecting and collecting a base measure (once), and then using this
base measure to evaluate a number of derived measures;
- Knowledge of which base measures are required to evaluate specific software
quality attributes (characteristics and subcharacteristics).

Below, we present a mapping of both the measures and the quality models in ISO 9126 to
the measurement information model described in ISO 15939.

First, we refer to the bottom section of Figure 2 by the term ‘Data Collection’ (e.g. the
measurement methods and the base measures), the middle section by the term ‘Data
Preparation’, using agreed-upon mathematical formulas and related labels (e.g. measurement
functions and derived measures), and the top section by the term the ‘Data Analysis’ (e.g.
analysis model, indicator and interpretation).
 Information Product

Data Analysis
Interpretation

Indicator

Analysis Model

Data Preparation Derived Measure

Measurement Function

Data Collection
Base Measure

Measurement Method

Attributes

Figure 2: ISO 15939 (2002) Measurement Information Model – three different sections.

As we have already discussed both data collection and data preparation, we now focus on
the ‘Data Analysis’ section. It is in the ‘Analysis Model’ part of the ISO 15939 measurement
information model that the ISO 9126 models of software product quality are to be used.
Figures 3, 4 and 5 present these generic models of ISO 9126 [2].

Process
Software Product Effect of Software Product
influences influences influences
Process Internal External Quality in
Quality Quality Quality use
Attributes Attributes Attributes
depends on depends on depends on
Context of use
Process Internal External Quality in use
Measures Measures Measures Measures
Figure 3: Quality in the life cycle – ISO 9126-1 [2].
 User Quality Needs Quality in use
Use and feedback

Contribute to specifying
Indicates

External Quality Requirement External Quality

Validation

Contribute to specifying
Indicates

Internal Quality Requirement Internal Quality

Verification

Figure 4: Quality in the software life cycle – ISO 9126-1 [2].

External and Internal Quality

Functionality Reliability Usability EfficiencyMaintainabilityPortability

Suitability- Maturity- Understandability- Time- Analyzability- Adaptability

Accuracy- Fault- LearnabilityBehavior- Changeability- Installability
InteroperabilityTolerance- Operability- Resource- Stability- Co-existence
Security- Recoverability- AttractivenessUtilization- Testability- Replaceability
Functionality- Reliability- Usability- Efficiency- Maintainability- Portability ComplianceComplianceComplianceComplianceComplianceCompliance

Figure 5: Quality model for External and Internal Quality – ISO 9126-1 [2].

These generic ISO models (Figures 3, 4 and 5) are to be instantiated in any particular
context of measuring the quality of a specific software product. This is usually performed in a
four-step process:
1. Identification of quality-related requirements, that is, the selection of the parts of the
ISO quality models that are relevant to a particular context of quality evaluation
(Figure 5);
2. Identification of the context of interpretation (Figure 6), that is:
- the selection of reference values, such values being either generic or specific
threshold values, or
- the determination of targets specified for a particular context;
3. Use of the derived measures from the data preparation phase to fill out the
instantiated quality model determined in 1 (Figure 5);
4. Comparison of the results of step 3 with either the set of reference values or targets
determined in step 2 (Figure 6).
This process is summarized in Figure 6, and a number of examples are presented in the side
box.
 Data Analysis Comparison of the Values (Interpretation Model)

Generic Thresholds Values

OR
Or Targets Values

(Indicators)
Identification of the context of interpretation

Identification of quality related requirements

Measurement Values

Data Preparation Derived Measures

Measurement Function

Data Collection Base Measures Base Measures

Measurement Method Measurement Method

Attributes Attributes

Figure 6: Mapping to the Measurement Information Model.

Examples
The examples presented here illustrate the process described in Figure 6. We include some of the
ISO 9126 base measures, and show the way in which they are combined to construct a derived
measure using a computational formula (measurement function):
Example 1:
Data Collection:
- Base Measure 1 (B1): Number of inaccurate computations encountered by users.
- Base Measure 2 (B2): Operation time.
Data Preparation:
- Derived Measure: B1 / B2
- Name of Derived Measure: Computational accuracy.
Data Analysis
- Quality group name: External quality measures.
- Characteristic: Functionality. - Subcharacteristic: Accuracy.
- Comparison of values obtained with the indicators (generic thresholds and/or targets).
Example 2:
Data Collection:
- Base Measure 1 (B1): Number of detected failures.
- Base Measure 2 (B2): Number of performed test cases.
 Data preparation:
- Derived Measure: B1 / B2
- Name of Derived Measure: Failure density against test cases.
Data Analysis:
- Quality group name: External quality measures.
- Characteristic: Reliability. - Subcharacteristic: Maturity.
- Comparison of values obtained with the indicators (generic thresholds and/or targets).
Example 3:
Data Collection:
- Base Measure 1 (B1): Task time.
- Base Measure 2 (B2): Help time.
- Base Measure 3 (B3): Error time.
- Base Measure 4 (B4): Search time.
Data Preparation:
- Derived Measure: (B1-B2-B3-B4) / B1
- Name of Derived Measure: Productive proportion.
Data Analysis:
- Quality group name: Quality in Use measures.
- Characteristic: Productivity.
- Comparison of values obtained with the indicators (generic thresholds and/or targets).
Example 4:
Data Collection:
- Base Measure 1 (B1): Number of errors made by user.
- Base Measure 2 (B2): Number of tasks.
Data Preparation:
- Derived Measure: B1 / B2
- Name of Derived Measure: Error frequency.
Data Analysis:
- Quality group name: Quality in Use measures.
- Characteristic: Effectiveness.
- Comparison of values obtained with the indicators (generic thresholds and/or targets).

4. OUTSTANDING HARMONIZATION ISSUES

4.1 Terminology
The ISO 9126 working group has come up with three new expressions which they introduced
in ISO TR 25021, namely: quality measure element categories, quality measure elements
and quality measures [6, 7]. The introduction of these new terms raises the following
concern: either the proper mapping to the set of classic metrology terms has not yet been
completed or there are concepts and related terms missing in the metrology vocabulary. The
latter would be surprising, since metrology is a rather mature domain of knowledge based on
centuries of expertise in the field of measurement and related international standardization. In
this section, we revisit the new documents, ISO FDIS 25020 and ISO 25021, with a view to
recommending a proper mapping of concepts to the related metrology terms, as well as to ISO
15939.

In ISO 25021, it is claimed that a quality measure element is either a base measure or a
derived measure [6, 7]; however, the consensual metrology terms are then ignored in favor of
locally defined WG6 measures, thus bypassing the ISO and SC7 harmonization requirements
on measurement terminology.

Quality measure elements are described as an input for the measurement of the software
quality measures of external quality, internal quality and quality in use [6, 7]. Figure 7 shows
the relationship between the quality measure elements and the software quality measures, and
between the software quality measures and the quality characteristics and subcharacteristics.
In metrology, these would correspond to base measures and derived measures respectively. It
can be observed that these measures, in particular the derived measures, are defined
specifically to measure the subcharacteristics of internal and external quality or the
characteristics of quality in use. None of these is directly related to the top level of software
quality (which is itself broken down into three models, then into 16 characteristics and further
into a large number of subcharacteristics). Therefore, the expression ‘software quality
measures’, which was selected in ISO 25021, is at a level of abstraction that does not represent
the proper mapping of the measures to the concept being measured.

Software Product Quality

indicate Software Quality Measures
composed of
generates
Quality
Measurement
Characteristics
Function
composed of
indicate are applied to

Quality Measure Elements

Quality
Sub-Characteristics
Figure 7: Quality Measure Elements Concept in the Software Product Quality Measurement
Reference Model [6, 7]

Now, there are 15 categories of quality measure elements [7] in ISO FDIS 25021 (see
Table 3)3.

Table 3: The 15 Quality Measure Elements Categories

Data Size Number of I/O Events Number of Test Cases
Number of Data Items Number of Requirements Number of Trials
Number of Failures Number of Restarts Number of User Operations
Number of Faults Number of System Operations Product Size
Number of Functions Number of Tasks Time Duration

It can be observed in Table 3 that a number of the quantities have a label starting with
‘number of’. However, these do not use a reference scale typical of measures in the sciences
or in engineering, but are rather counts of entities. For any of these proposed counts, such as
Number of Functions, no specific method is proposed for identifying this number in a
consistent manner across measurers and organizations; for instance, the definition of the word
‘function’ could differ from one individual to another within a single organization, and more
so across organizations. Therefore, to say in ISO TR 25021 that such numbers are obtained by

3
It is to be noted that in ISO DIS 25021 there are no specific quality measure elements proposed within the
Number of User Operations or Number of System Operations categories.
an ‘objective’ method is an overstatement, since they must be obtained mostly on the basis of
the judgment of the person performing the count.

Of the 15 categories, only ‘time’ comes from a classic base measure, in that it uses the
international standard unit of the second (or a multiple or submultiple of it) as its reference
scale. There are also measuring instruments to ensure that time measurements are indeed
obtained in an objective manner.

It can also be observed that, of the 15 categories in Table 3, at most four are directly
related to the quality of software: number of faults, number of failures, number of restarts and
number of trials. None of the other 11 measures is directly or indirectly related to the quality
of software. In fact, they are strictly independent of it, as they are often used for normalization
purposes, for instance.

For the Product Size category, ISO TR 25021 lists only one way to measure this
dimension, that is, in non-comment lines of code. There are also other ways to measure
product size, such as function points, modules, classes and visual structures. Furthermore,
there are various methods for counting lines of code and for measuring function points.
Therefore, this quality measure element category could be further split into different quality
measure elements (base measures). Moreover, the ISO has specified mandatory requirements
for functional size measurement methods (ISO 14143-1) [10], and has recognized four of
these as ISO standards which meet these requirements, such as COSMIC-FFP [11]. None of
these existing ISO software engineering standards, which are referenced in ISO 90003 [12],
has been mentioned or referenced in ISO TR 25021. Also, the various methods available to
obtain those numbers have their strengths and weaknesses, from a measurement perspective,
in terms of repeatability, reproducibility, software domains of applicability and accuracy.

In summary, from our point of view, it is not necessary to introduce new terms such as
quality measure element categories , quality measure elements or quality measures: the
terminology and concepts currently in ISO VIM [8] and ISO 15939 [9] are sufficient.

4.2 Limited coverage of the ISO quality models and corresponding measures
ISO TR 9126, parts 2 to 4, presents the ISO inventory of measures for the full coverage of the
ISO software product quality models (internal quality, external quality and quality in use) for
measuring any of their quality characteristics and subcharacteristics. The full sets of base
measures in these three parts of ISO 9126 is presented in Appendix A, and includes 82 base
measures.

Of these 82 base measures, only 57 are included in ISO 25021; this means that the
coverage in this new ISO document is limited, and the reasons for this are not obvious. In
addition, out of the 197 measures in ISO 9126, only 51 were selected for ISO 25021 as quality
measures. The content coverage of this subset of quality measures (derived measures) is
limited, and no specific criteria are provided on how they were selected. Some generic
information is provided in this standard to suggest that they were derived from a
questionnaire-based survey; however, no information is provided about the criteria for
selection, the size and representativeness of the sample in the countries where the data were
collected, nor about the representativeness of this sample outside these countries. Another
claim, that “they represent a default kernel of quality measures which are proven to be used in
common practice” [7], is not supported by documented evidence, nor is there a discussion of
their generalizability outside their data collection context.

Appendix B (Tables B1, B2 and B3) presents a detailed analysis of the coverage of the
quality measures in ISO 25021, together with their corresponding availability in ISO 9126.
Table B1 specifically illustrates that 34 measures for the external quality of a software product
were selected in ISO 25021 out of an inventory of 112 in the corresponding ISO 9126-2, while
78 measures were excluded, again without a documented rationale.

Table B2 provides similar information on measures for the internal quality of a software
product, as selected for ISO 25021; out of 70 measures, only 15 were selected, and they cover
only 4 of the 6 quality characteristics of the ISO model of internal quality, and only 9 of 27
subcharacteristics; again, there is no documented rationale for excluding any characteristic or
subcharacteristic. The same is true of Table B3 for the Quality in Use quality measures:
- Included: only 2 measures of the 15 already available in ISO 9126-4;
- Excluded: 2 quality-in-use characteristics, which are ‘safety’ and ‘satisfaction’;
- Also excluded: any quality measure elements related to the Number of User
Operations and Number of System Operations.

4.3 Redundancy issues

Some additional information included in ISO 25021 has already been covered in ISO 9126
documents, and will be included in the ISO 25000 series; for instance, information about scale
types is covered through rephrasing information contained in other documents, once again
increasing synchronization and harmonization right away and over the long term.

The situation is similar for commentaries on the measures of internal software quality,
external software quality and software quality in use, as well as for commentaries on the
software measurement methods. This is contrary to the ISO practice of avoiding duplication,
redundancy or the rephrasing of information across ISO documents, and increases the
possibility of inconsistencies arising across documents, all of which could lead to significant
effort over the long term in maintaining synchronization of documents covering similar
subsets of information.

These examples point to configuration management issues over the long term which will
represent additional cost to the purchasers of these ISO documents, since they will be required
to pay twice for the same information, which is, in fact, a subset of the full inventory. This
could lead to some confusion for standards users as to which of these documents is most
valuable to a standard purchaser, and under what circumstances.

We have illustrated in this section how the issue of ambiguity and redundancy in ISO
FDIS 25020 and ISO 25021 with respect to the use of the new terms quality measure
elements categories, quality measure elements and quality measures, which can be avoided
through the use of the corresponding metrology concepts and terms.
5.SUMMARY & DISCUSSION
Within their mandate to upgrade its set of technical reports on the measurement of the quality
of software products (ISO 9126), the ISO has come up with a new structure for upgrading the
current series of ISO 9126 documents for the measurement of the quality of software products.
This new structure is referred to as Software Product Quality Requirements and Evaluation, or
SQuaRE. In this chapter, we have presented practitioners with an alignment of the ISO models
of software product quality with the measurement information model of ISO 15939, and
explained how to use them for data collection, data preparation and data analysis. Some
examples have also been provided.

In addition, some issues have been raised in this chapter concerning three new concepts
proposed in SQuaRE; that is, quality measure element categories , quality measure
elements and quality measures The following is a summary of the harmonization issues
identified:
A) Terminology in ISO 25021:
- what is referred to as a ‘quality measure element’ corresponds to the classic concept of
the ‘base measure’ in ISO 15939;
- what is referred to as ‘software quality measure’:
- corresponds to the classic concept of the ‘derived measure’ in ISO 15939;
- is not at the proper level of abstraction for the concept being measured when
mapped to the hierarchy of concepts for software product quality adopted by the
ISO.
- In both ISO FDIS 25020 and ISO 25021, the ‘measurement method’ is defined as “a
logical sequence of operations, described generically, used in quantifying an attribute
with respect to a specific scale.” But, in ISO 25021, it is used in the ‘set of quality
measure elements’ to represent the type of the measurement method (objective or
subjective). In contrast, a new field called ‘detail’ is used to represent the measurement
method.
B) Harmonization with the Information Model of ISO 15939:
- Unless the terminology is harmonized with ISO International Vocabulary of Basic and
General Terms in Metrology, then it is challenging to align the older versions of ISO
9126 and ISO 14598, and it will be even more challenging with the ISO 25000 updates.
- Should the proposed terminology harmonization be accepted, it will become easier to
map each of these ISO 9126 and ISO 14598 series to the ISO 15939 Information Model.
C) Description harmonization:
- A large number of the base measures proposed in ISO 25021 are counts of entities rather
than measures per se with the required metrological characteristics, such as unit, scale,
dimension, measurement method, measurement procedures, etc.
- In ISO 25021, in some instances, like Product Size for example, there is no reference to
other existing ISO standards for software size, such as ISO 19761, etc.
- There are a number of claims that the proposed base measures are objective, while they
are obviously derived from a manual process without precisely documented
measurement procedures, thereby leaving much to the measurer’s judgment.
D) Coverage harmonization in ISO 25021:
- The documented set of base measures represents only a limited subset of the base
measures within ISO 9126, parts 2 to 4; the rationale for inclusion or exclusion is not
documented.
 - The set of base measures does not allow coverage of the full spectrum of quality
characteristics and subcharacteristics in ISO 9126, parts 2 to 4; again, the rationale for
inclusion or exclusion is not documented.

These concerns can be summarized as follows:

- Quality measure element categories and quality measure elements: non alignment
with the classical terminology on measurement is puzzling.
- Quality measures: inconsistency in the terminology used, and ambiguity about
which level of the ISO 9126 multi-level standard is being applied.

From the above analysis, we propose the following recommendations:

- Ensure that the terminology on software product quality measurement is fully
aligned with the classical measurement terminology in the sciences and in
engineering;
- Provide full coverage of the base measures for all three ISO models of software
quality;
- Provide improved documentation of the base measures using the criteria from
metrology;
- Provide clear mapping and traceability of the new ISO 25000 documents to the
ISO 15939 Information Model.

We have analyzed some of weaknesses of the new terms, and have proposed ways to
address them by using the ISO 15939 measurement information model on software
measurement process. In summary, the previously defined terms such as ‘base measure’ and
‘derived measure’, as well as the proper mapping to the Measurement Information Model in
well-developed standards like ISO 15939, and the International Vocabulary of Basic and
General Terms in Metrology (ISO VIM) are more useful than the weakly defined terms that
have recently been introduced.

REFERENCES
[1] ISO, ISO/IEC IS 9126: Software Product Evaluation - Quality Characteristics and
Guidelines for Their Use. Geneva, Switzerland: International Organization for
Standardization, 1991.

[2] ISO, ISO/IEC 9126-1: Software Engineering - Product Quality - Part 1: Quality Model .
Geneva, Switzerland: International Organization for Standardization, 2001.

[3] ISO, ISO/IEC TR 9126-2: Software Engineering - Product Quality - Part 2: External
Metrics. Geneva, Switzerland: International Organization for Standardization, 2003.

[4] ISO, ISO/IEC TR 9126-3: Software Engineering - Product Quality - Part 3: Internal
Metrics. Geneva, Switzerland: International Organization for Standardization, 2003.

[5] ISO, ISO/IEC TR 9126-4: Software Engineering - Product Quality - Part 4: Quality in
Use Metrics. Geneva, Switzerland: International Organization for Standardization, 2004.
[6] ISO, ISO/IEC FDIS 25020: Software Engineering - Software Product Quality
Requirements and Evaluation (SQuaRE) - Measurement Reference Model and Guide .
Geneva, Switzerland: International Organization for Standardization, 2007.

[7] ISO, ISO/IEC TR 25021: Software Engineering - Software Product Quality

Requirements and Evaluation (SQuaRE) - Quality Measure Elements . Geneva,
Switzerland: International Organization for Standardization, 2007.

[8] ISO, International Vocabulary of Basic and General Terms in Metrology (VIM) . Geneva,
Switzerland: International Organization for Standardization, 1993.

[9] ISO, ISO/IEC IS 15939: Software Engineering - Software Measurement Process .

Geneva, Switzerland: International Organization for Standardization, 2002.

[10] ISO/IEC, ISO/IEC 14143-1: Software Measurement - Functional Size Measurement -

Part 1: Definition of Concepts . Geneva, Switzerland: International Organization for
Standardization, 1998.

[11] ISO/IEC, ISO/IEC 19761: Software Engineering - COSMIC-FFP - A Functional Size

Measurement Method. Geneva, Switzerland: International Organization for
Standardization, 2003.

[12] ISO/IEC, ISO/IEC 90003: Software Engineering - Guidelines for the Application of ISO
9001:2000 to Computer Software. Geneva, Switzerland: International Organization for
Standardization, 2004.
 APPENDIX A

Table A1: The List of ISO 9126 Base Measures and their Measurement Units.

External Base Measures

Measure Name Unit of Measurement
1 Number of Functions Function (number of)
2 Operation Time Minute
3 Number of Inaccurate Computations Encountered by Users Case (number of)
4 Total Number of Data Formats Format (number of)
5 Number of Illegal Operations Operation (number of)
6 Number of Items Requiring Compliance Item (number of)
7 Number of Interfaces Requiring Compliance Interface (number of)
8 Number of Faults Fault (number of)
9 Number of Failures Failure (number of)
10 Product Size Byte
11 Number of Test Cases Case (number of)
12 Number of Breakdowns Breakdown (number of)
13 Time to Repair Minute
14 Down Time Minute
15 Number of Restarts Restart (number of)
16 Number of Restoration Required Restoration (number of)
17 Number of Tutorials Tutorial (number of)
18 Number of I/O Data Items Item (number of)
19 Ease of Function Learning Minute
20 Number of Tasks Task (number of)
21 Help Frequency Access (number of)
22 Error Correction Minute
23 Number of Screens or Forms Screens (number of)
24 Number of User Errors or Changes Error (number of)
25 Number of Attempts to Customize Attempt (number of)
26 Total Number of Usability Compliance Items Specified Item (number of)
27 Response Time Second or Millisecond
28 Number of Evaluations Evaluation (number of)
29 Turnaround Time Second or Millisecond
30 Task Time Minute
31 Number of I/O Related Errors Error (number of)
32 User Waiting Time of I/O Device Utilization Second or Millisecond
33 Number of Memory Related Errors Error (number of)
34 Number of Transmission Related Errors Error (number of)
35 Transmission Capacity Byte
36 Number of Revised Versions Version (number of)

16/2
37 Number of Resolved Failures Failure (number of)
38 Porting User Friendliness Minute

Internal Base Measures

Measure Name Unit of Measurement
1 Number of Functions Function (number of)
2 Number of Data Items Item (number of)
3 Number of Data Formats Formats (number of)
4 Number of Interface Protocols Protocol (number of)
5 Number of Access Types Access-Type (number of)
6 Number of Access Controllability Requirements Requirement (number of)
7 Number of Instances of Data Corruption Instance (number of)
8 Number of Compliance Items Item (number of)
9 Number of Interface Requiring Compliance Interface (number of)
10 Number of Faults Fault (number of)
11 Number of Test Cases Test-Case (number of)
12 Number of Restoration Requirement (number of)
13 Number of Input Items Which Could Check for Valid Data Item (number of)
14 Number of Operations Operation (number of)
15 Number of Messages Implemented Message (number of)
16 Number of Interface Elements Element (number of)
17 Response Time Second or Millisecond
18 Turnaround Time Second or Millisecond
19 I/O Utilization (Number of Buffers) Buffer (number of)
20 Memory Utilization Byte
21 Number of Lines of Code Directly Related to System Calls Line (number of)
22 Number of I/O Related Errors Error (number of)
23 Number of Memory Related Errors Error (number of)
24 Number of Items Required to be Logged Item (number of)
25 Number of Modifications Made Modification (number of)
26 Number of Variables Variable (number of)
27 Number of Diagnostic Functions Required Function (number of)
28 Number of Entities Entity (number of)
29 Number of Built-in Test Function Required Function (number of)
30 Number of Test Dependencies on Other System Dependency (number of)
31 Number of Diagnostic Checkpoints Checkpoint (number of)
32 Number of Data Structures Data-Structure (number of)
33 Total Number of Setup Operations Operation (number of)
34 Number of Installation Steps Step (number of)

Quality in Use Base Measures

Measure Name Unit of Measurement
1 Task Effectiveness (a given weight)
2 Total Number of Tasks Task (number of)
3 Task Time Minute
4 Cost of the Task Dollar
5 Help Time Second
6 Error Time Second
7 Search Time Second
8 Number of Users User (number of)
9 Total Number of People Potentially Affected by the System Person (number of)
10 Total Number of Usage Situations Situation (number of)
 Basic measure to quantify
Table A2: The Cross-Reference Table of ISO 9126 Base Measure Usage.

Quality in
Measure Name Uni External Internal
use
t
Functionality Reliability Usability Efficiency Maintainability Portability Functionality Reliability Usability Efficiency Maintainability Portability Q1 Q2 Q3 Q4
F1 F2
F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5 F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5
1 Number of
Functions
Function ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
2 Operation Time
Number of Minute ! ! ! ! ! ! ! ! ! ! ! ! !
Inaccurate !
3 Computations
Encountered Case
by Users !
4 Number of
Data Formats
Number of Format !
5 Illegal !
Operations Operatio
!
Number n
!
of
6 Items Requiring
Item ! ! ! ! ! ! ! ! ! ! ! !
Compliance
Number of
Interfaces
7 Requiring Interface ! !
Compliance
Number of
8 Faults Fault ! ! !
9 Number of
Failures
10 Product Size Failure ! ! ! ! ! !
11 Number of
Test Cases Byte !
12 Number of
Breakdowns Case ! ! ! !
13 Time to Repair !
14 Down Time Breakdo ! !
15 Number wn
of Restarts
16 Number of Minute !
Restoration !
17 Number of
Tutorials Restart !
18 Number of I/O
Restorati
! !
on
Tutorial !
Item !

19/29
 Quality in
Measure Name Unit External Internal
use
Functionality Reliability Usability Efficiency Maintainability Portability Functionality Reliability Usability Efficiency Maintainability Portability Q1 Q2 Q3 Q4
F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5 F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5
Data Items
Ease of
19 Function Minute !
Learning
20 Number
of Tasks
Task ! ! ! !
21 Help
Frequency Freq. !
22 Error
Correction Minute !
23 Number of
Screens/Forms
24 Number of Screen !
User
Errors/Changes
Error ! !
Number of
25 Attempts to
Customize Attempt !
Number of
26 Corrected
Faults Fault !
27 Response Time
28 Number of Second ! !
Evaluations
29 Turnaround Evaluatio ! !
Time n
30 Task Time Second ! !
31 Number of
I/O Related Minute ! !
Errors
User Waiting Error ! ! !
32 Time of I/O
Device
Utilization
Number of Second !
33 Memory
Related
Errors Error ! !
Number of
34 Transmission
Related Error !
Error
Transmission
35 Capacity Byte !

20/29
 Quality in
Measure Name Unit External Internal
use
Functionality Reliability Usability Efficiency Maintainability Portability Functionality Reliability Usability Efficiency Maintainability Portability Q1 Q2 Q3 Q4
F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5 F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5
Number of
36 Revised Version !
Versions
Number
of Failure ! ! ! !
37 Resolved
Failures
Porting User
38 Friendliness Minute !
39 Number of
Data Items Item ! ! ! ! ! !
Number of
40 Interface
Protocol !
Protocols
41 Number of
Access Types
Type !
Number of
42 Controllability
Requirements Req. !
Number of
43 Instances
of Data
Corruption Instance ! !
Number of
Items Which
44 Could Check
for Valid Data Item !
45 Number of
Operations
Number of Operatio
46 Messages ! !
n
Implemented
Number of Message !
47 Interface
Elements
I/O Utilization Element ! ! !
48 (Number of
Buffers)
Number of Buffer !
Line
49 of Code
Directly Line !

21/29
 Quality in
Measure Name Unit External Internal
use
Functionality Reliability Usability Efficiency Maintainability Portability Functionality Reliability Usability Efficiency Maintainability Portability Q1 Q2 Q3 Q4
F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5 F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5
Related to
System Calls
Number of
50 Items Item !
required to be
Logged
Modifica
Number of !
tion
51 Modifications
Made
52 Number of
Variables Variable !
Number of
Diagnostic
53 Functions
Required Function !

54 Number
of Entities
Entity !
Number of
55 Built-in Test
Functions
Required Function !
Number of Test
Dependencies
56 on Other
Depende
Systems !
Number of ncy
57 Checkpoints
58 Number of Checkpoi
data Structures !
nt
Number of Data-
59 Setup Structure !
Operations
Number Operatio
of ! !
n
60 Installation
Steps Step !
61 Task
Effectiveness
Cost of the (a Given
! !
62 Task Weight)
63 Help Time Dollar !
Second !

22/29
 Quality in
Measure Name Unit External Internal
use
FunctionalityReliabilityUsabilityEfficiency MaintainabilityPortabilityFunctionalityReliabilityUsabilityEfficiency Maintainability PortabilityQ1 Q2 Q3 Q4
F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5 F1 F2 F3 F4 F5 R1 R2 R3 R4 U1 U2 U3 U4 U5 E1 E2 E3 M1 M2 M3 M4 M5 P1 P2 P3 P4 P5

Error Time Search Time Number

64 of Second Second !! !
Users 65 User !
Number of People Potentially
66Affected by the System Number of Usage Situations
!

67 Person !!

68 Situation !

Legend of the Quality in Use characteristics Legend of the External and Internal sub-characteristics
F1 Suitability E1 Time Behavior
F2 Accuracy E2 Resource Utilization
F3 Interoperability E3 Efficiency Compliance
F4 Security M1 Analyzability
F5 Functionality Compliance M2 Changeability
R1 Maturity (Hardware/Software/Data) M3 Stability
R2 Fault Tolerance M4 Testability
R3 Recoverability (Data/Process/Technology) M5 Maintainability Compliance
R4 Reliability Compliance P1 Adaptability
U1 Understandability P2 Instability
U2 Learnability P3 Co-existence
Q Effectiveness U3 Operability P4 Replaceability
1 U4 Attractiveness P5 Portability Compliance
Q2 Productivity
Q3 Safety U5 Usability Compliance
Q4 Satisfaction

23/29
 Appendix B

Table B1: External Quality Measures in ISO 9126-2 and ISO 25021.

Quality Quality ISO DTR ISO

Measure Names
Characteristics Subcharacteristics 25021 9126-2
Functionality Accuracy 1 Computational accuracy  
2 Precision  
3 Accuracy relative to expectations 
Interoperability 4 Data exchangeability (data format-based)  
Data exchangeability (user’s success,
5 attempt-based) 
Security 6 Access controllability  
7 Access auditability 
8 Data corruption prevention 
Suitability 9 Funcional implementation completeness  
10 Functional adequacy  
11 Functional implementation coverage  
Functional specification stability
12
(volatility) 
Functionality 13 Functional compliance 
Compliance 14 Interface standard compliance 
Reliability Maturity 15 Failure density against test cases  
16 Failure resolution  
17 Fault removal  
18 Mean time between failures (MTBF)  
19 Test maturity  
20 Estimated latent fault density  
21 Fault density  
Test coverage (specified operation
22 scenario testing coverage) 
Recoverability 23 Restartability  
24 Availability 
25 Mean down time 
26 Mean recovery time 
27 Restorability 
28 Restore effectiveness 
Fault Tolerance 29 Breakdown avoidance 
30 Failure avoidance 
31 Incorrect operation avoidance 
Reliability
Compliance
32 Reliability compliance 
Usability Learnability Effectiveness of the user documentation
33 and/or help system  
34 Help accessibility  
35 Ease of function learning 
36 Ease of learning to perform a task in use 
Effectiveness of user documentation
37 and/or help system in use 
38 Help frequency 

24/2
 Quality Quality ISO DTR ISO
Measure Names
Characteristics Subcharacteristics 25021 9126-2
Operability 39 Physical accessibility  
40 Operational consistency in use 
41 Error correction 
42 Error correction in use 
43 Default value availability in use 
44 Message understandability in use 
45 Self-explanatory error messages 
46 Operational error recoverability in use 
Time between human error operations in
47
use 
48 Undoability (user error correction) 
49 Customizability 
50 Operation procedure reduction 
Understandability 51 Completeness of description  
52 Function understandability  
53 Understandable input and output  
54 Demonstration accessibility 
55 Demonstration accessibility in use 
56 Demonstration effectiveness 
57 Evident functions 
Attractiveness 58 Attractive interaction 
59 Interface appearance customizability 
Usability
Compliance 60 Usability compliance 
Efficiency Resource 61 I/O loading limits  
Utilization 62 Maximum memory utilization  
63 Maximum transmission utilization  
64 Mean occurrence of transmission error  
65 I/O device utilization 
66 I/O-related errors 
67 Mean I/O fulfillment ratio 
User waiting time of I/O device
68 utilization 
69 Mean occurrence of memory errors 
70 Ratio of memory error/time 
71 Media device utilization balancing 
72 Mean transmission error per time 
73 Transmission capacity utilization 
Time Behavior 74 Response time (mean time to respond)  
75 Throughput (mean amount of throughput)  
Turnaround time (mean time for
76 turnaround)  
77 Response time 
Response time (worst-case response time
78 ratio) 
79 Throughput 
Throughput (worst-case throughput time
80 ratio) 
 Quality Quality ISO DTR ISO
Measure Names
Characteristics Subcharacteristics 25021 9126-2
81 Turnaround time 
Turnaround time (worst-case turnaround
82
time ratio) 
83 Waiting time 
Efficiency
Compliance 84 Efficiency compliance 
Maintainability Analyzability 85 Audit trail capability  
86 Diagnostic function support 
87 Failure analysis capability 
88 Failure analysis efficiency 
89 Status monitoring capability 
Changeability 90 Software change control capability  
91 Change cycle efficiency 
92 Change implementation elapsed time 
93 Modification complexity 
94 Parameterized modifiability 
Stability 95 Change success ratio 
Modification impact localization
96
(emerging failure after change) 
Testability 97 Availability of built-in test function 
98 Retest efficiency 
99 Test restartability 
Maintainability
Compliance 100 Maintainability compliance 
Portability Adaptability 101 Adaptability of data structures  
Hardware environmental adaptability
102 (adaptability to hardware devices and  
network facilities)
System software environmental
adaptability (adaptability to OS, network
103  
software and cooperated application
software)
Organizational environment adaptability
104 (organization‘s adaptability to its 
infrastructure)
105 Porting user-friendliness 
Installability 106 Ease of installation  
107 Ease of setup retry 
Coexistance
Replaceability 108 Availability coexistence 
Replaceability 109 Continued use of data 
110 Function inclusiveness 
111 User support functional consistency 
Portability
Compliance 112 Portability Compliance 
 Table B2: Internal Quality Measures in ISO 9126-3 and ISO 25021.

Quality Quality ISO DTR ISO

Measure Names
Characteristics Subcharacteristics 25021 9126-3
Functionality Accuracy 1 Computational accuracy  
2 Precision  
Interoperability 3 Data exchangeability (data format-based)  
4 Interface consistency (protocol) 
Security 5 Access controllability  
6 Access auditability 
7 Data corruption prevention 
8 Data encryption 
Suitability 9 Funcional implementation completeness  
10 Functional adequacy  
11 Functional implementation coverage  
Functional specification stability
12
(volatility) 
Functionality 13 Functional compliance 
Compliance 14 Intersystem standard compliance 
Reliability Maturity 15 Fault removal  
16 Fault detection 
17 Test adequacy 
Recoverability 18 Restorability 
19 Restoration effictiveness 
Fault Tolerance 20 Failure avoidance 
21 Incorrect operation avoidance 
Reliability
Compliance 22 Reliability compliance 
Usability Learnability Completeness of user documentation
23
and/or help facility  
Operability 24 Physical accessibility  
25 Input validity checking 
26 User operation cancellability 
27 User operation undoability 
28 Customizability 
29 Operation status monitoring capability 
30 Operational consistency 
31 Message clarity 
32 Interface element clarity 
33 Operational error recoverability 
Understandability 34 Completeness of description  
35 Function understandability  
36 Demonstration capability 
37 Evident functions 
Attractiveness 38 Attractive interaction 
39 User interface appearance customizability 
Usability
Compliance
40 Usability compliance 
Efficiency Resource 41 I/O utilization 
Utilization 42 I/O utilization message density 
 Quality Quality ISO DTR ISO
Measure Names
Characteristics Subcharacteristics 25021 9126-3
43 Memory utilization 
44 Memory utilization message density 
45 Transmission utilization 
Time Behavior 46 Response time 
47 Throughput time 
48 Turnaround time 
Efficiency
Compliance
49 Efficiency compliance 
Maintainability Analyzability 50 Activity recording 
51 Readiness of diagnostic function 
Changeability 52 Change recordability 
Stability 53 Change impact 
54 Modification impact localization 
Testability 55 Completeness of built-in test function 
56 Autonomy of testability 
57 Test progress observability 
Maintainability
Compliance 58 Maintainability compliance 
Portability Adaptability 59 Adaptability of data structures  
Hardware environmental adaptability
60 (adaptability to hardware devices and  
network facilities)
System software environmental
adaptability (adaptability to OS, network
61  
software and cooperated application
software)
62 Organizational environment adaptability 
63 Porting user-friendliness 
Installability 64 Ease of setup retry 
65 Installation effort 
66 Installation flexibility 
Co-existence 67 Availability of coexistence 
Replaceability 68 Continued use of data 
69 Functional inclusiveness 
Portability
Compliance 70 Portability compliance 
 Table B3: Quality in Use Measures in ISO/IEC TR 9126-4 and ISO/IEC TR 25021.

ISO DTR ISO

Quality Characteristics Measure Names
25021 9126-4
Effectiveness 1 Task effectiveness 
2 Task completion  
3 Error frequency 
Productivity 4 Task time  
5 Task efficiency 
6 Economic productivity 
7 Productive proportion 
8 Relative user efficiency 
Safety 9 User health and safety 
10 Safety of people affected by use of the system 
11 Economic damage 
12 Software damage 
Satisfaction 13 Satisfaction scale 
14 Satisfaction questionnaire 
15 Discretionary usage 