Data Flow Diagram Guidelines Flashcard
Flashcards for Data Flow Diagram (DFD) Rules
Front: The input to a process must be different from the output of that process.
Back: Inputs and outputs of a process should not be identical to ensure proper data transformation.
Front: Every process must have a unique name.
Back: A process name should clearly describe its function to avoid confusion and ensure accurate documentation.
Front: Can a process have only outputs?
Back: No. If an object has only outputs, it must be a source, not a process.
Front: Can a process have only inputs?
Back: No. If an object has only inputs, it must be a sink, not a process.
Front: How is a process labeled?
Back: A process is labeled using a verb phrase (e.g., "Process Order").
Front: Can data move directly from one data store to another?
Back: No. Data must be moved by a process.
Front: Can data move directly from an outside source to a data store?
Back: No. Data must be moved by a process.
Front: Can data move directly from a data store to an outside sink?
Back: No. Data must be moved by a process.
Front: How is a data store labeled?
Back: A data store has a noun phrase label (e.g., "Customer Database").
Front: Can data move directly from a source to a sink?
Back: No. It must be moved by a process if the data is relevant to the system; otherwise, it is not shown on the DFD.
Front: How is a source/sink labeled?
Back: A source or sink is labeled with a noun phrase (e.g., "Supplier").
Front: How does data flow direction work in a DFD?
Back: Data flows in only one direction between symbols. For actions like "read before update," use two separate arrows to indicate different timings.
Front: What is a fork in data flow?
Back: A fork means the same data is sent from a common point to multiple processes, data stores, or sources/sinks, indicating copies of data going to different locations.
Front: What is a join in data flow?
Back: A join means identical data is received at a common location from multiple sources (processes, data stores, or external entities).
Front: Can data flow directly back to the process it originated from?
Back: No. At least one intermediary process must handle the flow before it loops back.
Front: What does a data flow to a data store represent?
Back: It represents an update (delete or change).
Front: What does a data flow from a data store represent?
Back: It represents a retrieval or use of data.
Front: How is a data flow labeled?
Back: A data flow should have a noun phrase label. Multiple noun phrases can appear on one arrow if they move together as a single unit.
Front: How does time work in DFDs?
Back: Time is not well-represented in DFDs. It is best to treat DFDs as perpetual systems with no defined start or stop points.
Front: How should DFD development be approached?
Back: DFD development is iterative. Expect to redraw diagrams multiple times to achieve the closest approximation of the system.