- 1.1 Definition of a Requirement and a Stakeholder
- 1.2 Requirements Pyramid
- 1.3 Traceability between Requirements
- 1.4 Characteristics of a Good Requirement
- 1.5 An Overview of the Requirements Management Process
- 1.6 Summary
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
1.5 An Overview of the Requirements Management Process
A simplified description of the requirements management process contains the following major steps:
- Establishing a requirements management plan
- Requirements elicitation
- Developing the Vision document
- Creating use cases
- Supplementary specification
- Creating test cases from use cases
- Creating test cases from the supplementary specification
- System design
The first step (requirements management plan) defines the requirements pyramid. In each of the next seven steps, one of the elements of the pyramid is built. Table 1.1 describes which requirement types and what documents are created in each step. As you can see, the process navigates through the pyramid from the top down and from left to right.
Table 1.1 Requirements and Documents Created in Each Phase
|
Step |
Requirement Types |
Documents |
|
Requirements elicitation |
Stakeholder needs |
Stakeholder requests |
|
Developing the Vision document |
Features |
Vision |
|
Creating use cases |
Use cases, scenarios |
Use case specifications |
|
Supplementary specification |
Supplementary requirements |
Supplementary specification |
|
Creating test cases from use cases |
Test cases |
Test cases |
|
Creating test cases from the supplementary specification |
Test cases |
Test cases |
|
System design |
Class diagrams, interaction diagrams |
UML diagrams |
Requirements management is an interactive process. In a typical iteration, a full pass through the pyramid is performed. Even in the same iteration, we can go back a few steps and repeat the activity. For example, during the creation of a test case, we can discover that some information is missing, and we need more input from a stakeholder, so we go back a step to “gathering requirements.” To maintain the model’s integrity, it is important to update all affected requirements. In initial iterations the emphasis is placed on the first few steps (the top of the pyramid), and in later iterations more time is spent on the lower part of the pyramid.
This description is simplified because only major steps are described. For example, some activities defined by the Rational Unified Process (RUP) are more granular. (For example, our step of creating use cases contains the following RUP activities: find actors and use cases, structure the use case model, prioritize use cases, and detail use cases.)
Let’s briefly look through all the steps.
Establishing a Requirements Management Plan
One of the first tasks in the project is developing a Requirements Management Plan (RMP). The RMP describes the overall approach to managing requirements in the project. The document details how requirements are created, organized, modified, and traced during the project lifecycle. It also describes all requirement types and their attributes used in the project.
Here are some questions that can be answered in the RMP:
- Will any RM tool be used?
- What requirement types will be tracked in the project?
- What are the attributes of these requirements?
- Where will the requirements be created—in the database only or in the documents?
- Between which requirements do we need to implement traceability?
- What documents are required?
- Which requirements and documents will be used as a contract with customers?
- If part of the project is outsourced, what requirements and documents will be used as a contract with a vendor?
- Will we follow the RUP or some other methodology?
- Does the customer need any specific documents to comply with his development process?
- How will change management be implemented?
- Assuming that RequisitePro is used, will the whole system be stored in one RequisitePro project or spread among many projects?
- What process will guarantee that all requirements were implemented and tested?
- Which requirements or views do we need to generate reports?
Chapter 3, “Establishing a Requirements Management Plan,” describes all these decisions in detail.
Requirements Elicitation
At the top level of the pyramid are stakeholder needs, as shown in Figure 1.3.
)
Figure 1.3 Needs (stakeholder requests) are at the top of the pyramid.
This book uses the terms “stakeholder needs” and “stakeholder requests” as synonyms. However, if project specifics require, it is possible to define them as two separate requirement types. In that case, needs would be very high-level requirements, such as “The system shall have the capability to book a flight.” Usually there are no more than five high-level needs per stakeholder and no more than 15 needs per project. All detailed requirements would be captured as stakeholder requests. However, in many projects it is easier to capture all input from the stakeholders in the same type of requirement, so in the example used in this book, stakeholder needs represent all input from the stakeholders, regardless of granularity. In some projects there may be a need to distinguish between “stakeholder needs” describing initial requirements and “stakeholder requests” that may include subsequent change requests.
Requirements elicitation, also called requirements gathering, is a very important step. Missing or misinterpreting a requirement at this stage will propagate the problem through the development lifecycle.
Here are some of the techniques used to elicit requirements from stakeholders:
- Interviews
- Questionnaires
- Workshops
- Storyboards
- Role-playing
- Brainstorming sessions
- Affinity diagrams
- Prototyping
- Analysis of existing documents
- Use cases
- Analysis of existing systems
These techniques are described in Chapter 5, “Requirements Elicitation.”
Developing the Vision Document
Section 1.3 discussed attributes of a good requirement. However, information that comes from stakeholders does not necessarily have these attributes. It is especially the case that requirements coming from different sources may be conflicting or redundant.
During development of the Vision document, one of the main goals of business analysts is deriving features from stakeholder needs (see Figure 1.4). Features should have all the attributes of a good requirement. They should be testable, nonredundant, clear, and so on.
)
Figure 1.4 Features are derived from needs.
The Vision document should contain essential information about the system being developed. Besides listing all the features, it should contain a product overview, a user description, a summary of the system’s capabilities, and other information that may be required to understand the system’s purpose. It may also list all stakeholder needs in case they were not captured in separate documents.
Some parts of the Vision are created at the very beginning, before we even start eliciting requirements from stakeholders. Examples of these sections are
- A description of the problem being solved.
- Identification of users/customers/stakeholders.
Creating Use Cases
Functional requirements are best described in the form of use cases. They are derived from features, as shown in Figure 1.5.
)
Figure 1.5 Use cases are derived from features that describe the system’s functionality. Scenarios are derived from use cases.
A use case is a description of a system in terms of a sequence of actions. It should yield an observable result or value for the actor (an actor is someone or something that interacts with the system). The use cases
- Are initiated by an actor.
- Model an interaction between an actor and the system.
- Describe a sequence of actions.
- Capture functional requirements.
- Should provide some value to an actor.
- Represent a complete and meaningful flow of events.
Here is a fragment of a use case:
- The user enters required flight information: departure airport and date, arrival airport and date.
- The system displays all outbound flights matching the search criteria.
- The user selects an outbound flight.
- The system displays a list of available return flights.
The purpose of a use case is to facilitate agreement between developers, customers, and users about what the system should do. A use case becomes sort of a contract between developers and customers. It’s also a basis for use case realizations, which play a major role in design. In addition, you can produce sequence diagrams, communication diagrams, and class diagrams from use cases. Furthermore, you can derive user documentation from use cases. Use cases may also be useful in planning the technical content of iterations and give system developers a better understanding of the system’s purpose. Finally, you can use them as an input for test cases.
While designing use cases we will also define scenarios—specific paths through the use case. We usually implement systems scenario by scenario, not the whole use case at once. Scenarios are required when we derive test cases from use cases. On the requirements pyramid, scenarios are one level below use cases (see Figure 1.5).
Supplementary Specification
Supplementary specification captures nonfunctional requirements (usability, reliability, performance, supportability) and some functional requirements that are spread across the system, so it is tough to capture them in the use cases. These requirements are called supplementary requirements and are derived from features, as shown in Figure 1.6.
)
Figure 1.6 Supplementary requirements are derived from features that cannot be captured in the use cases.
Chapter 8, “Supplementary Specification,” discusses this type of requirement in detail.
Creating Test Cases from Use Cases
As soon as all the requirements are captured, we should design a way to check whether they are properly implemented in the final product. Test cases will show the testers what steps should be performed to test all requirements. In this step we will concentrate on creating test cases from use cases. If we did not create scenarios while generating use cases, we need to define them now. Test cases are at the lowest level of the pyramid, as shown in Figure 1.7.
)
Figure 1.7 Test cases for testing use cases.
This process is described in detail in Chapter 9, “Creating Test Cases from Use Cases.”
Creating Test Cases from the Supplementary Specification
The approach used in the preceding step does not apply to testing supplementary requirements. Because these requirements are not expressed as a sequence of actions, the concept of scenarios does not apply to them. An individual approach should be applied to each of the supplementary requirements because techniques used to test performance requirements are different from techniques used to test usability requirements. In this step we also design testing infrastructure and platform-related issues.
Sometimes we need to “borrow” one scenario that was created to test use cases (see Figure 1.8). For example, to test the requirement “The system should run using the Internet Explorer (IE) browser and using the Netscape browser,” we should select one scenario (preferably basic flow of the most popular use case) and test the full scenario in the IE browser. Then we should test the same scenario again in the Netscape browser. There is no need to test all test cases created in the preceding step in both browsers. Just select those that contain some functionality that may be browser-specific.
)
Figure 1.8 Test cases for testing supplementary requirements.
Some of the supplementary requirements can be tested using automated testing tools such as Rational Robot.
System Design
Requirements are the basis for system design, which is often facilitated by use of the Unified Modeling Language (UML) [BOO98]. Many tools, such as Rational Rose, Rational Software Architect, Rational Data Architect, and Rational Software Modeler, can significantly facilitate the creation of all required diagrams.
One approach is to create interaction diagrams from scenarios and, at the same time, assign functionality to the classes (see Figure 1.9). This topic is discussed in detail in Chapter 11, “Object-Oriented Design.”
)
Figure 1.9 System design.