CPSC 3720 - DAY 14
FEBRUARY 17, 2017

================================================================================

CONTROL FLOW
------------

	Procedure-driven control
		Operations wait for input from actor
		Traditional method
		step-by-step
		difficult to follow when objects get involved
			ex. text/command line programs
	Event-driven control
		System waits for external event
		Objects handle each event
		Multi-step sequences are more difficult
	Threads
		Threads are created for each event/procedure
		Rest of the system can still function
		Difficult to work with and debug
			Multicore processors require this for optimum
			performance

BOUNDRY CONDITIONS
------------------

	How is the system started?  Shut Down?
	Crashes?
	Network Outages?
	Power Outages?
	Exception Handling?
		Hardware failure?
		Software fault?
		Changes in operating environment - network issues, power outage
	Create use cases for these conditions
		boundry use cases
	
REVIEWING SYSTEM DESIGN
-----------------------

	The system design model is complete if every requirement and every 
	system design issue has been addressed.
		Have the boundry conditions been handled?
		Was there a walkthrough of use cases to identify missing
		functionality?
		Have all use cases been examined and assigned a control object?
		Have all aspects of system design been addressed?
		Do all subsystems have defintions?
	
	The system design model is correct if the analysis model can be mapped
	to the system design model
		Can every subsystem be traced back to a use case or 
		nonfunctional requirement?
		Can every use case be mapped to a set of subsystems?
		Can every design goal be traced back to a nonfunctional 
		requirement?
		Is every nonfunctional requirement addressed in system design
		model?
		Does each actor have an access policy?
		Is every access policy consistent with the nonfunctional
		security requirement?

	The system design is consistent if it does not contain any
	contradictions
		Are conflicting design goals prioritized?
		Does any design goal violate a nonfunctional requiremnts?
		Are there multiple subsystems or classes with the same name?
		Are collections of objects exchanged among subsystems in a 
		consistent manner?

	The model is realistic if the system can be implemented
	The model is readable if developers not involved in th esystem design
	can understand the model.
		Are subsystem names understandable?
		Do entities with similar names denote similar conceps