Вы находитесь на странице: 1из 9

Empirical Studies of Requirements Validation

Techniques

Uzair Akbar Raja


Blekinge Tekniska Hgskola, Ronneby, Sweden
uara06@student.bth.se

Abstract Requirements validation ensures that the 1998 which provides guidelines for documenting
software requirements specification contains complete and software requirements specification.
consistent set of requirements which are according to the
customer needs. This paper provides overview of The requirements for the system are subject to change
requirements validation techniques like requirements due to changing stakeholders needs, changes in
inspections, requirements prototyping, requirements testing environment, change in laws and business plan [1].
and viewpoint-oriented requirements validation. This paper Therefore requirements management facilitates to manage
highlights pros and cons of these requirements validation the changes to the system [1].
techniques. In requirements testing special attention is given
to TCD inspections.
The requirement validation process ensures that the
requirements in software requirements specification are
consistent and complete [1]. The requirements validation
1. INTRODUCTION process attempts to identify the errors in software
requirements specification (SRS) before it is used as a
basis for further system development [1].
According to Sommerville and Kotonya requirements
engineering covers all activities involved in discovering, The final phase of requirements engineering is
documenting and maintaining a set of requirements for requirements validation. Requirements validation
computer based systems [1]. Requirements engineering is identifies the ambiguous requirements and resolves
based on requirements engineering process. requirement conflicts.
Requirements engineering process is a structured set According to [3] 56 percent of all the bugs in software
of activities which are carried out to derive, validate and projects are due to the bugs inserted in the requirements
maintain requirements documents [1]. Requirements phase as shown in Figure 1. Half of these bugs are due to
engineering process is divided into five categories or the incomplete and ambiguous requirements and the other
process activities namely requirements elicitation, half are due to the requirements that are omitted. The cost
requirements analysis & negotiation, requirements of fixing a bug is less in requirements engineering phase
specification, requirements management and requirements because few deliverables have been delivered at this
validation. phase.
In the process of requirements elicitation requirements
engineers discover user requirements from domain
knowledge, system documents and market studies [1].
Requirements engineers usually conduct interviews with
stakeholders or customers to articulate their needs.
During this phase requirements are analyzed in detail
and stakeholders negotiate to decide set of agreed
requirements for the proposed system [1]. This process
helps to resolve conflicts among system stakeholders due
to incomplete information or requirements incompatibility
within the available budget [1].
The agreed requirements are documented in
requirements specification document. This document is
also known as software requirements specification (SRS).
IEEE has defined a standard known as IEEE Std 830- FIGURE 1. DISTRIBUTION OF BUGS [3]
2. RESEARCH QUESTIONS 6. REQUIREMENTS INSPECTIONS

Following research questions will be answered during Inspections were introduced by Fagan in 1976 and
this study on requirements validation. they are considered to be the powerful technique to detect
errors [10]. Literature evidence indicates that 50-90 % of
What are the problems with requirements validation in
defects can be caught using inspections [9]. CMMI and
general?
ISO / IEC 15504 also recommend inspections as a part of
What are various requirements validation techniques development process [11], [12].
proposed in literature?
In requirements inspection process we use inspections
What are the benefits and problems with requirements to detect incomplete or ambiguous requirements in
validation techniques? software requirements specification (SRS).
What is the importance in selection of a particular
requirements validation technique in software
development projects? a. Requirements inspection process

3. RESEARCH METHODOLOGY The inspection process comprises steps like planning,


overview, defect detection, defect collection, defect
correction and follow-up. These steps are explained in
Most studies in software engineering are exploratory. detail as follow.
Exploratory studies are based on analyzing a particular
problem or situation and these studies are useful for
determining the current research work or new insights in a 1) Planning: The primary objective of the planning phase
specific domain ( like requirements validation in our case) is to organize an inspection meeting when materials for
[16]. Exploratory studies can be performed by literature inspections are determined [4]. The materials for
reviews, open questionnaires and interviews [16]. In this inspections are normally selected when they pass entry
research study authors objective is to compare and criteria for example when the software requirements
analyze requirements validation techniques. Therefore specification (SRS) is complete [4]. The key activities
literature review is suitable methodology to fulfill this during the planning phase are the selection of inspection
purpose. During the study articles from IEEE, ACM and participants, assigning roles to them, scheduling
Springer will be referenced. inspection meeting and distribution of inspection materials
[4].
4. PROBLEMS WITH REQUIREMENTS VALIDATION
2) Overview: In the overview phase the author or authors
of the software requirements specification (SRS) explains
One of the major problems with requirements it to the inspection team [4]. The objective behind
validation phase is the application of a strategy or overview phase is to make the inspected software
framework for requirements validation. Mostly companies requirements specification (SRS) easier to understand for
practice requirements validation on ad-hoc basis due to the inspection team. There are certain limitations
lack of trained staff or lack of training or exposure to associated with overview phase [4]. Firstly these meetings
requirements validation. More attention is given to consume effort and time. Secondly this phase does not
software testing phase which is usually carried out in the ensure the independent assessment of inspected
end when all modules in software project are integrated. requirements specification. For example, it may happen
that during the overview phase all inspectors focus their
5. REQUIREMENTS VALIDATION TECHNIQUES attentions on a particular aspect of software requirements
specification (SRS) and this conflict may consume more
time before arriving at a conclusion.
Requirement validation techniques ensure the
completeness of the software requirements specification. Researchers like Fagan, Glib and Graham name this
Some of the requirements validation techniques are phase as Kickoff Meeting and categorize this phase as
requirements reviews, prototyping, use case based optional phase [4]. On the other hand there are three
modeling and viewpoint-oriented requirements validation conditions which support this phase [4].
[1].
When the artefact to be inspected is complex and
The requirements validation techniques covered difficult to understand then this phase is justified.
during this study will be requirements inspections,
requirements prototyping, requirements testing and When the artefact to inspected belongs to a large
viewpoint-oriented requirements validation. software system. In this case overview phase
helps the inspectors to understand the In order to make decision regarding reinspection of a
relationship of that artefact to other components software artefact certain statistical models like capture-
recapture models are used [4]. Capture-recapture models
of the system.
are used in biological studies to estimate the size of
This phase facilitates the training process. For animal population [17]. When an animal gets trapped, it is
instance the new team members participate in captured, marked and then released. When marked animal
these meeting. This way they learn about various is recaptured, it provides estimate of the total population
software engineering activities. size based on overlaping samples.
The idea behind capture-recapture models in
requirements inspections is to allow several inspectors
3) Defect detection: Defect detection phase is the core read the same requirements specification [17]. Each
phase of inspection process [4]. In this phase defects are inspector will have a list of requirements with defects.
identified in the software requirements specification Then based on the overlap of defected requirements
(SRS). This phase can be organized in two ways. First is among the inspectors, one can estimate the number of
the individual defect detection where each inspector defects remaining in the software requirements
individually inspects the requirements specification or in specification.
group meeting where two or more inspectors inspects the
requirements specification.
5) Defect correction: In the defect correction phase the
Knight and Meyers [4] have proposed the idea of a author of the software requirements specification (SRS)
Phased Inspection Method. In the Phased Inspection
Method each inspection phase is further divided into mini- corrects the software requirements specification based on
inspections. One or more inspectors are assigned to each the feedback obtained from defect collection phase. This
mini-inspection [4]. phase ensures that the detected defects are completely
For example consider a software requirements removed form the software requirements specifcation
specification (SRS) for library information system. We (SRS).
create four mini-inspections using Phased Index Method
and assign two or more inspectors to each mini-inspection. 6) Follow-up: The objective of follow-up phase is to
Then if any requirement in software requirements
ensure that the author of software requirements
specification (SRS) pertaining to these four activities is
missing, this will be identified. specification (SRS) has resolved all the reported
incomplete or inconsistent requirements [4].
Porter et.al has proved empirically that mini-
inspections are costly [4]. During mini-inspections when a
defect is dectected it is repaired and then the same process b. Requirements inspection roles
of inspection starts again. This increases the project cost.
This is the reason that we often find mini inspections [4].
The people in the inspection team are assigned various
roles to carry out the requirements inspection process
smoothly. Inspection team normally comprises three to six
4) Defect collection: There are three major objectives of people with following roles [5].
defect collection phase [4], the defects detected by the
inspectors during the defect detection phase are collected
and documented, the decision is made that the reported 1) Moderator: Moderator is a key person in the inspection
defect is actually a defect ot not and the decision is made process. Moderator ensures that inspection team is
to reinspect the defected software artefact or not.
selected appropriately, inspection team is trained to carry
In order to fulfill these objectives this phase is out the inspection process and abides by the predefined
normally carried out in group meeting. Since group
rules and regulations for a particular meeting [5].
meetings are time consuming so three alternatives namely
managed meetings, depositions and correspondence are
reported in literature [4]. 2) Producer or author: Producer is normally a person
Managed meeting are formal meetings involving who has developed a software requirements specification
limited number of participants [4]. Deposition is a special under inspection [5]. During the overview phase of
type of meeting involving three persons only [4]. They are requirements inspection process the producer explains the
moderator, inspector and author. In correspondence software requirements specification to other inspection
inspectors and authors use email for communication [4].
team members. In some literature books producers are
They do not need to physically meet at some place.
also called as authors [6]. The producer must not serve as d. Advantages & disadvantages of requirements
moderator, reader or recorder [6]. inspections

The inspections are useful for large companies as they


3) Reader: The primary responsibility of reader is to require resources in terms of manpower. In large
present the inspected requirements specification and lead companies multiple inspection teams can be formed. Each
the inspection team in a logical fashion [6]. The role of team may be assigned responsibility to inspect particular
reader seems to be conflicting with producer or author software artifact like software requirements specification
role. But reader helps to divert the attention of inspection (SRS), design documents or code.
team from producer to the product under inspcetion [5].
On the other hand inspections cannot be useful for
small companies. Although literature evidences support
inspection team of three people [5]. But in small
4) Recorder (Optional): The recorder helps to reduce the
burden from moderator. The recorder keeps note of all companies with comparatively smaller human resource,
objections and issues raised during the inspection meeting inspections may not be as effective.
and helps the moderator to focus on the leading discussion Inspections require knowledge of reading techniques
[5]. and also some pre-requisite knowledge about software
requirements specification (in case of requirements
inspections). Therefore inspections also require training
5) Inspector: Inspectors responsibilities are to focus on and external consultants can be involved in training [7].
the software requirements specification under inspection So this training also adds to the overall project cost.
If a company is planning to use requirements
and identify the confliciting or incomplete requirements
inspections without any previous experience, they may
[6]. Similarly when the these requirements are removed introduce it as a pilot project with their normal
inspectors are responsible for re-examining the software requirements validation process [7]. This will help to
requirements specification (SRS). judge the suitability of requirements inspections as
effective requirements validation technique.
c. Requirements inspection reading techniques
7. REQUIREMENTS PROTOTYPING
A reading technique provides a series of steps or
guidelines to the inspectors to detect defects in software Requirements prototyping helps requirements
requirements specification (SRS). In other words reading engineers to identify and understand the customer
technique is a strategy for inspector [4]. Among reading requirements. In addition prototyping also facilitates other
techniques ad-hoc reading and checklist-based reading are activities like system designing and user interface
most popular because they are practiced by the number of development. Requirements prototyping serves as a
companies [4]. communication media between requirements engineers
and requirement issuers. It enables customers to
1) Ad-hoc reading: In ad-hoc reading software participate in the requirements validation process. There
requirements specification is given to inspectors without are two types of prototyping [8],
any guidelines or directions to look for incomplete a. Throw-away prototyping.
requirements [4]. The word ad-hoc means that no
technical support is given to inspectors to detect defects in b. Evolutionary prototyping.
software requirements specification [4]. In ad-hoc reading According to [8] throw-away prototyping helps in the
defect detection depends on the experience and identification of poorly understood requirements. It is
knowledge of the inspector. used to resolve inconsistencies between requirements
engineers and requirement issuers. Once both agree to a
2) Checklist based readin: In checklist-based reading set of requirements, the prototype is discarded and these
inspectors are provided with questions to answer while agreed requirements are incorporated in the software
inspecting a software requirements specification (SRS). requirements specification (SRS).
These questions are related to the quality of the software
requirements specification [4]. Evolutionary prototyping is based on the agreed
requirements and is subject to quality constraints as
Checklists provide a structure to the validation process imposed in the software development [8]. The theme
and reduce the chances that inspectors will forget some behind evolutionary prototyping is to develop a functional
aspects of the requirements documents [7]. prototype and then constantly refine it based on the
feedback. The Dynamic Systems Development Method process they may be incorporated into requirements
(DSDM) is one of the agile methods and it uses documents.
prototypes in its framework [18]. These prototypes can be
Sometimes it is necessary to write prototype manual.
either throw away or evolutionary.
This will require resources like a person responsible for
Therefore the companies using agile methods like drafting manual, time to write manual and financial
DSDM can use requirements prototyping for requirements resources. In such cases this cost should be included in the
validation. requirements validation costs [1].

a. Advantages & disadvantages of requirements 8. REQUIREMENTS TESTING


prototyping
The objective behind requirements testing is to
The advantages and disadvantages of requirements validate the requirements in the software requirements
prototyping are summarized below. specification (SRS) rather than the system [7]. In
Prototyping helps customers to visualize the requirements testing test-case are defined for each
developing software system. Customers may find difficult requirement in the software requirements specification
to visualize the requirements which are written in the (SRS). This helps to identify incomplete or ambiguous
software requirements specification (SRS) [7]. They may requirements because if for a particular requirement it is
find difficult to understand how these written difficult to derive a test-case then it indicates some
requirements will be transformed into a working software problem [7].
system. So they may not be able to identify the problems.
But on the other hand prototyping helps them to
understand the requirements and identify the problems,
issues or need for some additional / changed requirements.
The prototypes developed during the requirements
validation phase may also be used for system testing later
in the system validation process [7]. This way prototyping
ensures the consistency and correctness of final developed
system.
In some companies executable prototypes are used as
stop-gap system. The stop-gap systems are those systems
which can be delivered to the customer in case of delays
in the implementation of the final system [7]. Meanwhile
when the final system is developed and ready for
installation, it replaces the executable prototype. The
prototype developed during the requirements validation
phase can be further transformed into executable
prototype.
Some of the disadvantages of the prototyping
approach are listed below. FIGURE 2 . REQUIREMENTS TESTING [3]

Prototyping is usually costly and time consuming


activity. If a company wants to implement prototyping The test-case derived for requirements testing can also
then they need to purchase softwares for prototyping and be used during the final testing of the system. Figure 2
train people. In addition a requirement engineer will have represents the requirements testing.
to do additional activity of building a prototype along with
gathering customer requirements, prioritizing them and In requirements testing the verification and validation
documenting them in software requirements specification department or testing department is involved early in the
(SRS). requirements validation phase to derive test-cases. This
normally requires human resource in terms of testers along
Paper prototyping is relatively cheap to apply [7]. In with requirements engineers. Similarly the testers should
paper prototyping, prototypes are drawn on paper. But it be experienced professional and in case of new people
requires special record logs to be preserved for future proper training is needed. This normally increases the cost
reference. Similarly paper prototypes cannot be of project.
transformed into executables ones, so after validation
Test-case driven inspections commonly known as
TCD inspections are well known in requirements testing
and are based on the concept of reuse. Test cases step 2. These requirements are specified according to a
produced during TCD inspections are also used in other template and contains attribute such as ID, title,
phases of software development [13]. description.
During step 2 TCD inspection itself is performed by
a. Test case driven inspections (TCD inspections) creating test-cases with the requirements [13].
Inappropriate and low priority requirements are removed
Tony Gorschek and Nina Fogelstrm have proposed during this phase.
the technique of test-case driven inspection of pre-project
In step 3 inspected requirements are prioritized and in
requirements. This technique uses test-case as an
addition project planning is performed [13]. Like step 1
inspection tool by involving testers in the early stage of
and step 2 some requirements may also be discarded
development process known as pre-project stage [13].
during this phase. These three steps come under pre-
The TCD Inspection process was used in Danaher project phases and after step 3 requirements are selected
Motion Sr AB (DHR) Sweden and it is a three step for implementation.
process [13] as shown in Figure 3.
TCD Inspections are characterized by involving testers
in the inspection process, reducing the educational and
training cost of inspectors [13]. The most prominent
advantage of TCD inspections is for the product
managers. Product managers can select better
requirements for requirements selection and cost
estimation in the pre-project phase [13].
During the pilot study of TCD inspections at Danaher
Motion Sr AB (DHR) Sweden, the project managers
realized that the refined requirements were of vey high
quality [13]. Similarly completeness and understanding of
requirements improves during TCD inspections which
help in deciding to postpone or discard requirements.

b. Advantages & disadvantages of requirements testing

The main advantage of requirements testing is that it


ensures elimination of unwanted requirements and the
test-cases produced during it can also be used in the final
testing of the system.
The requirement testing is very useful for larger IT
companies having a well established testing department
with experienced testing professionals. Similarly
companies using this approach can also provide training
facilities to other companies planning to use requirements
testing.
The disadvantage of requirements testing is that it
involves costs. For small companies with relatively less
number of people it may not be useful. Similarly
requirements testing require experienced testers and
requirements engineers. Small companies may lack such
FIGURE 3. TCD INSPECTION PROCESS [13]
requirements in terms of experienced IT force. Moreover
small companies may not provide on job training to new
According to [13] during the step 1 product manager people for requirements testing.
selects and reviews requirements to be included in the
initial formulation / specification. Some requirements are 9. VIEWPOINT-ORIENTED REQUIREMENTS VALIDATION
also discarded. This is mostly ad-hoc process where
product manager utilizes his personal as well as
colleagues knowledge and experience for initial trial of Some of the terms used to understand the concept of
requirements. Then the selected requirements known as viewpoint oriented requirements validation are defined
formulated or specified requirements are passed on to the below.
Universe of discourse: The universe of discourse is VWPL: The VWPL stands for viewpoint language. It
the overall context in which software is developed is used to represents the viewpoints [1].
including all information and people involved in software
The method of viewpoint oriented requirements
development [15].
validation requires two or more analysts (viewpoints)
Actors: The people involved in the universe of using VWPL to model the universe of discourse. In order
discourse are known as actors [15]. The actors are of two to understand viewpoint oriented requirements validation
types, users on the demand size and software engineers on consider two analysts, analysts 1 and analyst 2 as shown in
the supply side [15]. the Figure 4.
Users on the demand size represent customers. Users The analyst 1 and analyst 2 model the universe of
on supply side include project managers, requirements discourse using three prospective and two hierarchies. The
engineers, programmers, testers. three prospective used are data, process and actor [1].
These perspectives and hierarchies are analyzed to
Viewpoint: Viewpoint is defined as the standing or produce list of discrepancies and types of discrepancies
mental position of an individual (actor) when examining [15]. Then the perspectives are integrated in views shown
or observing the universe of discourse [15]. It is identified as view1 and view2 in figure 4. These two views are
by individuals name and his role in the universe of compared for correctness and completeness. In figure 4
discourse. circles represents processes and boxes are the inputs and
Perspective: A perspective is the set of facts which outputs of processes [1].
are observed and modeled according to a particular The author searched electronic resources like ACM
modeling aspect and viewpoint [15]. digital library, IEEE Xplorer, Springer Link and Inspec.
View: View is defined as the integration of But the author was unable to find much of the research
perspectives [15]. work done in requirements validation using viewpoints.

a. Advantages & disadvantages of viewpoint-oriented


requirements validation

The viewpoint-oriented requirements validation helps


to identify conflicting requirements based on viewpoints.
This approach recognizes that requirements should not be
considered from a single prospective. The major
advantage of this approach is that it incorporates multiple
views.
As there is no much work available in this area. This
may hinder requirements engineers in industry to use this
approach for requirements validation.

10. DISCUSSION

Requirements validation is one of the most crucial


activities in requirements engineering. There are number
of requirements validation techniques available in
literature each having its pros and cons. The software
development companies should implement a strategy to
develop a framework of requirements validation. This
strategy should clearly specify one or more requirements
validation techniques to be used. This will helps to ensure
FIGURE 4. VIEWPOINT-ORIENTED REQUIREMENTS VALIDATION [1] that their developed software systems are according to the
stakeholders needs.
Hierarchies: The two types of hierarchies used in the The importance of requirements validation techniques
universe of discourse are is-a hierarchy of concepts and is determined by the fact that they help to identify defects
parts-of hierarchy of concepts [15]. that can be amplified in later stages of software
development. If these defects are identified in the
requirements validation phase, then overall cost of the specification. These test cases can be reused later in the
project will also be decreased and project will be integration testing or system testing. On the other hand
deployed to the customer in time. The cost is decreased in requirements prototyping can be reused further to develop
terms of less effort in software regression testing phase, fully functional prototype.
test-case reuse (if a company is using requirements
Requirements prototyping involves customers. This
testing) and less development time.
technique helps customers to visualize the system in
software requirements specification. Similarly in
viewpoint-oriented requirements validation customers can
TABLE 1. COMPARISON OF REQUIREMENT VALIDATION TECHNIQUES be involved to have their view of the proposed system.
Requirements inspections and requirements testing lack
this feature of customer involvement.

11. FUTURE WORK

In this research study literature review is carried out


regarding requirements validation techniques. This could
be extended to conducting an industrial survey to gather
information about practiced requirements validation
techniques. During surveys people working in the process
area of requirements engineering can be interviewed. The
interview format can be semi-structured as it combines
both structured and unstructured interviews. This way the
interview questionnaire will consist of set of questions
arranged in if-then structure and list of open topics
which will help to carry our comprehensive interview.

Table 1 shows the comparison of requirement


TABLE 2. DATA INPUT FOR INDUSTRIAL SURVEY
validation techniques discussed in this paper.
It is evident form Table 1 that requirements
inspections and requirements testing techniques require
large teams as compared to requirements prototyping and
viewpoint-oriented requirements validation. These large
teams can be of three to six people or multiple teams
working together.
Requirements inspections and requirements testing are
more costly as compared to requirements prototyping and
viewpoint-oriented requirements validation. They require
inspection or testing professionals, physical meeting place
for teams and training session in case of new
professionals.
All the requirements validation techniques can be used
by large companies. But requirements inspection and
requirements testing are more suitable for large companies
as they require resources. Viewpoint-oriented
requirements validation can be used by both small and
large companies as it does not require specialized teams
for requirements validation. Prototyping is mostly used by
small companies but it can also be equally useful for large
companies. Large companies can use requirements The interview data can be collected in a form given in
prototyping for stop-gap systems. Table 2. In Table 2 third row will provide requirement
validation techniques used in industry. These techniques
In requirements testing test cases are written for
could be used in an experiment to identify some of the
verifying requirements in software requirements
effective requirements validation techniques being
practiced in industry. If the experiment is performed in
academic setting then students can be used as subjects. [8] J. Siddiqi, I. Morrey, R. Hibberd, G. Buckberry, Towards a
System for the Construction, Clarification, Discovery and
Formalisation of Requirements, proceedings of first
12. CONCLUSION international conference on Requirements Engineering, IEEE,
1994, pp.230-238

The area of requirements validation has opened a new [9] O. Laitenberger, T. Beil, T. Schwinn, "An industrial case
horizon for researchers to explore. This will enable study to examine a non-traditional inspection implementation
companies to adopt techniques and develop systems for requirements specifications," presented at Proceedings of the
according to stakeholders needs in cost effective way. In Eighth IEEE Symposium on Software Metrics, Los Alamitos
order to solve the problems with requirements validation CA, 2002.
application of strategy for requirements validation is
required. Requirements prototyping can be easily used by [10] M. E. Fagan, "Design and Code Inspection to Reduce
both small and big companies. While requirements Errors in Program Development," IBM Systems Journal, vol. 15,
1976, pp. 182-211.
inspections and requirements testing are suitable for large
companies as they are performed in teams. On the other
[11] CMMI-PDT, "Capability Maturity Model Integration
hand there is not much research work done in viewpoint- (CMMI), Version 1.1," in CMMI for Systems Engineering,
oriented requirements validation. But viewpoint-oriented Software Engineering, Integrated Product and Process
requirements validation provides possibility to identify Development, and Supplier Sourcing Version 1.1 (CMMI-
conflicting requirements based on viewpoints of various SE/SW/IPPD/SS, V1.1). Pittsburgh, 2002.
stakeholders. These requirements validation techniques
will help to detect errors before they move on to the other [12] ISO/IEC, "Software Process Assessment TR 15504:1998,"
phase of software development life cycle and become vol. 2004. http://www.sei.cmu.edu/iso-15504/ : ISO/IEC, 1998.
amplified.
[13] Tony Gorschek, Nina Dzamashvili Fogelstrm, Test-
13. REFERENCES. case Driven Inspection of Pre-project Requirements-Process
Proposal and Industry Experience Report, in proceedings of the
Requirements Engineering Decision Support Workshop held in
[1] Gerald Kotonya, Ian Sommerville, Requirements conjunction with the 13th IEEE International Conference on
Engineering, John Wiley & Sons, New York, 1998. Requirements Engineering, 2005.

[2] Thayer, R. & Dorfman, M. Software Requirements [14] Carolyn B.Seaman, Qualitative Methods in Empirical
Engineering 2nd Edition, IEEE Computer Society Press, Studies of Software Engineering, IEEE transactions on
California, 1997. Software Engineering, Vol 25, No 4, July/August 1999.

[3] Gary E. Mogyorodi, What is Requirements- Based [15] J.C.S.P. Leite, P.A. Freeman, Requirements validation
Testing?, CROSS TALK The Journal of Defence Software through viewpoint resolution, IEEE Transactions on Software
Engineering, March 2003. Engineering, Vol 17, 1991, pp.1253-1269

[4] Oliver Laitenberger, A Survey of Software Inspection [16] Christian W.Dawson, Projects in Computing and
Technologies, Handbook on Software Engineering and Information Systems, Pearson, 2005, ISBN 0321263553.
Knowledge Engineering, Citeseer, 2002.
[17] L.C. Briand, K. El Emam, B.G. Freimut, O. Laitenberger ,
[5] Steven R. Rakitin. Software Verification and Validation for A comprehensive evaluation of capture-recapture models for
Practitioners and Managers, Artech House Publishers, 2nd estimating software defect content, IEEE Transactions on
edition (August 1, 2001). Software Engineering, Vol 26, No 6, 2000, pp. 518-54

[6] Ackerman, A. F., Buchwald, L. S., and Lewsky, F. H., [18] Dot Tudor, George A. Walter, Using an Agile Approach in
Software Inspections: An Effective Verification Process, a Large, Traditional Organization proceedings of agile 2006
IEEE Software, 1989, pp. 31-36. conference, IEEE, 2006
[7] Ian Sommerville, Pete Sawyer, Requirements Engineering,
John Wiley & Sons, New York, 1997.

Вам также может понравиться