Business Modeling – The Domain Model

1. Business Modeling – The Domain Model Source: Use Case Driven Object Modeling with UML – A Practical Approach By Doug Rosenberg ISBN: 0-201-43289-7 Also, pp. 101-110 OOSE Text

2. Background • A key component of Business Modeling (Domain Analysis) - in addition to Business Use Case Model and Business Object Model and other artifacts - is creating the Domain Model • The Domain Model contains Key Abstractions (business / technology) • Is a Visual Model of entities, relationships, multiplicities, and more. • Prior to embarking on gathering requirements and capturing them via Use Cases (system use cases, that is…), we need to understand the key entities in the domain – business or technology domain.

3. Domain Modeling - an Introduction • … • Domain Modeling is the task of discovering “objects” (classes, actually) that represent those entities and concepts. •  Recognize that the domain model will be a superset of your entity classes needed for your application development. • Your class diagrams – later on - will likely contain some of the entities found in your domain model. • Primarily ‘data’ – attributes of domain entities. • Similar to ERDs, but not the same.

4. Domain Modeling - continued • Domain Models sometimes considered ‘informal’ class diagrams. • Developed as part of domain analysis (business modeling) • The ‘classes’ (entities) represent what you have learned about various ‘things’ (entities) and relationships between them in the domain itself. • As a Visual Model, they will help in understanding the domain. • Also serve as a Glossary, in that the entities will contain attributes and other defining qualities.

5. Domain Modeling - continued • The Domain Model: not a class diagram • Does NOT wholly support requirements analysis (ahead). • Addresses entities in the domain of the organization – apart from computer applications that may use them. • Domain model contains entities, relationships and attributes • Many entities may well be outside the application scope to be developed. • Domain models are normallynot concerned with representations of inheritance, polymorphism, etc, (as we would be when embarking on development of an application.)

6. Domain Modeling - continued • Later, during requirements analysis, we will be developing class diagrams that will contain classes taken from the domain model. • But the class diagrams (for development) will represent data that will be stored (persistent objects) and manipulated by the application. • Application = real software modules – likely stored in a database • Instances (objects) will be loaded from and stored back into a database as the application runs.

7. Domain Modeling - continued • Models produced as part of requirements analysis (ahead lectures) will contain: • Entity classes - some may be derived from Domain Model, • Boundary classes - used for the user interface, and • Control Classes used for controlling logic and business rules, and more.

8. So, what do we do? • Doing domain modeling is very important. • But, we don’t want to spend too much time and try to model everything!!! • Yet we need to have a good starting point for requirements analysis to solve the ‘problem.’ • So, our domain modeling approach is to develop an ‘initial’ set of classes.

9. Domain Modeling (cont.) •  The domain model can serve as a glossary of terms, which is very useful to use case developers • Some approaches specify creation of a domain model OR a Glossary. • The domain model consists of entities – groups of objects with similar properties, common behaviors, common relationships, and common semantics. • Need to develop a static model of the domain by finding appropriate entities that represent the real key abstractions in the domain. • This serves to underpin system development later.

10. Domain Modeling (cont.) • Carefully Review Sources of Domain Knowledge. Here are a few: • Interviews • Questionnaires • Quarterly Reports • Mission Statements • Subject Matter Experts (SMEs) • Newsletters • Web pages • A company’s e-business…. • Look for nouns – these are candidate classes in the domain model. • Examples: Menu, customer, food order, Payment, On-line order… • Customer (class with attributes /behaviors) ‘orders’ (relationship) Food (class with attributes) – captured graphically. • Customer and Food are entities related by Orders…. • What does the organization DO? What is it all about? • (We will do this again later when we undertake requirements analysis…but we will use Use Cases as the primary input, if we already have a domain model / glossary)

11. Developing Domain Classes • Read sources very closely to capture these ‘nouns’ and noun phrases. • Verbs and verb phrases become candidate operations (methods later) • Possessive phrases generally indicate that the nouns should be attributes rather than objects. Try to identify attributes / operations. • Build associations (and name them) between the domain classes • Add multiplicities carefully • Don’t worry about aggregations and association classes and much more unless these relationships are clearly evident. • Model will undergo a refinement later. Try to capture all the domain info you can; model it; and verify it. Do in Rose or other suitable tool

12. Generalization Relationships and Associations • If clear from your info gathering, build generalization relationships • Parents, subclasses…. Inheritance of attributes, methods, and associations! • ‘is a’ relationships…. • Associations are static relationships between classes. • Show dependencies if needed.

13. Associations and Multiplicity • Label the associations as best you can. • Try to identify multiplicities, but don’t spend too much time. • Aggregations – identify classes such that one class is ‘made up’ from smaller pieces… ‘has a’ or ‘is a kind of’. • Also, there is composition – where one piece is ‘owned’ by another – later….. • Focus on simple aggregations now. • Don’t stress on relationships that are not obvious at this time.

14. Association Classes • Identify classes that particularly address the many-to-many relationships that link classes • These ‘associations’ typically have properties independent of classes they are linking. • Most domain models have at least one or two link (sometimes called ‘bridge’) classes. • Don’t overuse these….

15. Following slide is an example (has a few errors in it) that you may use as a guide.

16. Domain Model MEMBER Member_ID Member_Type_Number Member_First_Name Member_Middle_Initial Member_Last_Name Member_Address Member_City Member_State Member_Zip_Code Member_Phone_Number Member_Email University_ID_Number SYSTEM_USER Member_ID System_User_Password System_User_Title UNIVERSITY University_ID_Number University_Name University_Address University_City University_Zip_Code Is an authorized Belob to MEMBER_TYPE Member_Type_Number Member_Type_Description Is categorized as manages FINANCE Financial_ID_Number Financial_Date Member_ID Financial_Amount Financial_Desc Payment_Type_ID places VENDOR Vendor_Number Vendor_Name Vendor_Address Vendor_City Vendor_State Vendor_Zip_Code Vendor_Phone SALE_ORDER SO_Order_Number SL_Line_Number SO_Order_Date Member_ID MEMORABILIA_INVENTORY Item_Number Item_Description Cost_To_Member tracks PAYMENT_TYPE Payment_Type_ID Payment_Type_Desc provides Is generated for contains references SUPPLIES Supply_Number Vendor_Number Item_Number Cost_To_UPE REPLENISH_LINE RL_Line_Number Supply_Number RL_Line_Quantity SALE_LINE SL_Line_Number Item_Number REPLENISH_ORDER RO_Order_Number RL_Line_Number RO_Order_Date Is generated for identifies

17. 8 Top Domain Modeling Errors • 8. Start assigning multiplicities to associations right off the bat. Make sure that every association has an explicit multiplicity. • 7. Do noun and verb analysis so exhaustive that you pass out along the way. • 6. Debate whether to use aggregation or composition for each of your part-of associations • 5. Presume a specific implementation strategy without modeling the problem space. • 4. use hard-to-understand names for your classes – like cPortMgrIntf – instead of intuitively obvious ones, such as PortfolioManager.

18. Continuing… • 3. Jump directly to implementation constructs such as friend relationships and parameterized classes • 2. Create a one-for-one mapping between domain classes and relational database tables. • 1. Perform “premature patternization,” which involves building cool solutions, from patterns, that have little or no connection to user problems.

19. Transition to: The Problem Space!!! • Area within which your application is to exist!

20. The Problem and the Scope • The term “problem domain” refers to the area that encompasses real-world persons, places, and/or things and concepts related to the problem that the system to be developed / enhanced, etc. is being ‘required’ to solve. • A ‘problem’ may be considered a ‘difficulty’ (inadequacy of current system) or ‘opportunity’ for benefit, or more

21. Problem Statement (‘Vision’ of the Application to be Developed) • The problem statement should be a simple sentence or two. • Usually then found in a Vision Document • VERY IMPORTANT • Basis to answer the ultimate question: • Have we solved the problem? • Be careful what you write!!! • Wild inferences can be made.

22. Sample Problem Statement(Student Registration System) • “The system will allow students to register for courses, and change their registration, as simply and rapidly as possible. It will help students achieve their personal goals of obtaining their degree in the shortest reasonable time while taking courses that they find most interesting and fulfilling.” (p. 107, OOSE)

23. Scope and Its Bounds • Broader the scope, the more complex the application. • Along with the problem statement, include a list of features to be accommodated. • This will narrow features. • These can simply be ‘bulleted’ items • Scope, hence commitment, is clarified by listing features or sub-problems. • Determine that some of these are ‘out of scope’ or will be accommodated by a different application. • Good to have a comprehensive list citing features that are explicitly OUT of scope as well as those IN scope.

24. Scope and Bounds • Problem Statement (Vision) ‘has’ scope – that is, what the developed / enhanced application will accommodate. • Problem Statement, via plain English text, actually at a high level, contains the scope. • This is why the Vision Statement (or Problem Statement) should be followed by a list of ‘features.’ • More coming…