L ECTURE 9 – PROCESS MODELLING PART 1 Data Flow Diagrams for Process Modelling Multi-level Data Flow Diagrams Logical Vs Physical DFDs Steps to Construct Data Flow Diagrams Software Project Management Maria Petridou 1

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 2 Maria Petridou Once the system requirements have been defined and refined: Process modelling and Data modelling Process models can be of two types: Logical process models – describe processes with no details about specific implementation. Physical process models – produced in the design phase, provide further information necessary to build the system.

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 3 Maria Petridou Data Flow Diagram (DFD) Structured analysis technique for constructing a graphical representation of processes. One of the main methods available for analyzing data-oriented systems. Simple diagram to model processes and represent flow of information. DFDs emphasize the logic underlying the system. Contains four types of symbols: process, data flow, data store, external entity.

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 4 Maria Petridou Advantages of DFDs: Freedom from committing to the technical implementation too early. Understanding of the interrelationships of systems and subsystems Communicating current system knowledge to users. Analysis of the proposed system.

Software Project Management 5 Maria Petridou

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 6 Maria Petridou Four Basic Symbols: Process: Manual or computerised activity or function that is performed for some specific business reason (e.g. A whole system, a subsystem and an activity). Represented as a rectangle with rounded corners. Always denotes changes in data. Names should be in the form verb-adjective-noun Complex processes may require the use of more formal process specification techniques such as structured English (pseudo-code), decision tables or decision trees.

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 7 Maria Petridou Data Flow: Shows the data about a person, place, or thing that moves through the system. Represented as an arrow labelled with the data name (a noun). Data flows hold processes together and one end of the data flow will always come from or go to a process. Direction of the arrow indicates destination of data.

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 8 Maria Petridou Data Store: Collection of data stored in some way. Denotes long-term storage. Represented as a rectangle with an open right side. Has a descriptive name (noun) and a unique identification number such as D1, D2, D3. Data stores are the starting point for the data model and are the link between the process model and the data model. Must have at least one input data flow and at least one output data flow.

D ATA F LOW D IAGRAMS FOR P ROCESS M ODELLING Software Project Management 9 Maria Petridou External Entity: Person or organisation that is external to the system but interacts with it. Also known as source or destination of data. Represented as an square labelled with the entity name (a noun). People that are part of the system are not considered as external entities because if they execute a process, then they are part of the process (not external to the system). The same entity symbol can be used several times to avoid complex diagrams.

M ULTI - LEVEL D ATA F LOW D IAGRAMS Software Project Management 10 Maria Petridou Most business processes cannot be shown on a single DFD. Then, a hierarchy of data flow diagrams is required. Lower level diagrams (child diagrams) show a portion of an upper level diagram (parent diagram) in more detail. Balancing means that information presented at one level of a DFD is accurately represented in the next level DFD.

11

M ULTI - LEVEL D ATA F LOW D IAGRAMS Software Project Management 12 Maria Petridou Important Issues in Multi-level DFDs: Context diagram shows the overall business process as a single process and shows data flows to and from external entities. There is only one level 0 DFD and it shows the major high-level processes (typically up to 9 processes) and the data stores. There is one level 1 DFD for each process in the level 0 DFD showing more details on how the high-level process operates. If a given process at level n is decomposed into x processes at level n+1, then the x child processes should make up the parent process. Correct numbering at the different levels helps to understand the structure of the business process.

V S L OGICAL V S P HYSICAL DFD S Software Project Management 13 Maria Petridou Logical DFDs Shows how the business operates Processes represent business activities Data stores represent collections of data Not important how the data is stored Permanent collections Controls are rules of the business

V S L OGICAL V S P HYSICAL DFD S Software Project Management 14 Maria Petridou Physical DFDs Shows how the system will be implemented Processes represent programs/functions Data stores represent physical files and databases Processes operating at different times must be connected via a data store Controls are validation of user input, file formats and security measures

V S L OGICAL V S P HYSICAL DFD S Software Project Management 15 Maria Petridou Analyse the current system Add features for the new system Develop best methods for implementing the new system Current Logical DFD Current Logical DFD New Logical DFD New Logical DFD New Physical DFD New Physical DFD

V S L OGICAL V S P HYSICAL DFD S Software Project Management 16 Maria Petridou Logical DFDs – Advantages Better communication with system users Better stability for the system Better business understanding for analysts Better flexibility and maintenance Physical DFDs – Advantages Easier to categorise processes as manual or automatic Better description of processes Better for ordering processes into a sequence Better for imposing controls

S TEPS TO C ONSTRUCT D ATA F LOW D IAGRAMS Software Project Management 17 Maria Petridou 1. Build the context diagram, including all external entities and the major data flow to or from them. 2. Create Diagram Level 0 by analyzing the major activities within the context process - Include the external entities and major data stores. 3. Decompose to a child diagram (Level 1 DFD) for each complex process on Diagram Decompose level 1 processes into level 2 DFDs and decompose further if needed. 5. Balance and validate DFDs to ensure completeness and correctness.

S TEPS TO C ONSTRUCT D ATA F LOW D IAGRAMS Software Project Management 18 Maria Petridou Context Level Data Flow Diagram Contains only one process, representing the entire system The process is given the number zero All external entities are shown on the context diagram as well as major data flow to and from them. The diagram does not contain any data stores

S TEPS TO C ONSTRUCT D ATA F LOW D IAGRAMS Software Project Management 19 Maria Petridou Diagram Level 0 Diagram Level 0 is the explosion of the context level diagram. Should include up to 7 or 9 processes - Any more will result in a messy diagram. Processes are numbered with an integer. The major data stores and all external entities are included on Diagram 0.

S TEPS TO C ONSTRUCT D ATA F LOW D IAGRAMS Software Project Management 20 Maria Petridou Child Diagrams Each process on diagram Level 0 may be exploded (decomposed) to create a child diagram. Each process on a lower-level diagram may be exploded to create another child diagram. Each process is numbered with the parent diagram number, a period, and a unique child diagram number  3.2 on Diagram 3, the child of process 3  On Diagram 3, the processes would be numbered 3.1, 3.2, 3.3 and so on. External entities are usually not shown on the child diagrams below Diagram 0. Reading: (Kendall&Kendall, chapter 7), (Dennis &Wixom, chapter 6).

S TEPS TO C ONSTRUCT D ATA F LOW D IAGRAMS Software Project Management 21 Maria Petridou