Goal Modeling Betty H.C. Cheng. This presentation is available free for non-commercial use with attribution under a creative commons license.

Transcription

1 Goal Modeling 1

2 We Will Cover What is a goal? Where do goals come from? What is a goal model? When to use goal models? How do goal models relate to UML models? Why use goal models? Capturing the goal model 2

3 What is a goal? A stakeholder objective for the system The system includes the software and its environment Who are stakeholders? Anyone who has an interest in the system Customers, end users, system developers, system maintainers... What is a goal model? A hierarchy of goals Relates the high-level goals to low-level system requirements

4 Goal or Not a Goal? Goal Examples: Credit card information is kept private Credit card information is accurate Safe transportation Highly reliability Non-goal Examples: The system will be implemented in C++ The paint colors for the cars will be yellow, orange, and red 4

5 Goal Exercise Order the list of goals from high-level concern to low-level concern User receives a request for a timetable from a system Collect timetables by system Schedule meeting System collect timetables from user Collect timetables

6 Goal Exercise Order the list of goals from high-level concern to low-level concern Schedule meeting Collect timetables Collect timetables by system System collect timetables from user User receives a request for a timetable

7 Types of Goals Functional (Hard) escribe functions the system will perform Well defined criteria for satisfaction E.g., System collects timetables from user Non-functional (Soft or fuzzy) escribe desired system qualities Hard to define; satisficed rather than satisfied Reliability E.g., System should be reliable Quality E.g., System should be high quality.

8 Where do goals come from? Conveyed by stakeholders isclosed in requirements documents Analysis of similar or current system Elaborating other goals

9 Goal Exercise Identify goals in the following paragraph An Adaptive Regional Forecasting Ecological Observatory (ARFEO) is proposed as means to implement an operational, run-time configurable, and adaptable ecological observatory to be used for regional ecological forecasting. ARFEO has three main dimensions. First, using a holistic perspective of environment, we will use sensors that are analogous to human and organism senses to monitor the environment and its changes. As such, ARFEO can be configured to observe and answer ecological questions specific to one class of sensory input or across multiple sensory inputs. Second, ARFEO will use a cyberinfrastructure comprising smart, heterogeneous sensor networks, small-scale and GRI-scale distributed computing [6]. ARFEO s architecture will be a distributed design which will enable users to perform small and complex computations and transparent integration of data and analysis. ARFEO will enable a user to adapt monitoring capabilities at run-time, thus allowing a user to customize the configuration to specific needs and questions. Finally, ARFEO will emphasize the reuse and synergistic integration of existing analysis and visualization techniques to include in the computational toolkit for processing the sensor and ancillary data, as well as metadata. A key part of ARFEO will be the development of processes to make use of the toolkit element to support the analysis and visualization capabilities.

10 When to use goal models Early requirements engineering Focus on identifying problems Exploring system solutions and alternatives one before UML modeling Early RE Late RRE esign Code Test Why? What? How?

11 Why use goal models? Give rationale for requirements Identify stable information Guide requirement elaboration

12 The Goal Model

13 Running Example Scheduler Assists the initiator in scheduling a meeting should be convenient for participants Participants should be available Modeled using the i* goal notation Example adapted from the RE06 keynote given by John Mylopoulos 13

14 Collect timetables Goal Model Schedule meeting Choose schedule Goals are refined into subgoals that elaborate how the goal is achieved By person By system Manually Automatically Collect from agents Collect from users Send request Receive request Example adapted from the RE06 keynote given by John Mylopoulos

15 AN Refinement Collect timetables AN Schedule meeting AN Choose schedule All of the subgoals must be achieved for the goal to be achieved By person By system Manually Automatically Collect from agents Send request Collect from users AN AN Receive request Example adapted from the RE06 keynote given by John Mylopoulos 15

16 Refinement Collect timetables By person AN By system Schedule meeting AN Choose schedule Manually At least one of the subgoals must be achieved for the goal to be achieved Automatically Collect from agents Send request Collect from users AN AN Receive request Example adapted from the RE06 keynote given by John Mylopoulos 16

17 Interpretations of Refinements Collect timetables By person AN By system Schedule meeting AN Choose schedule Manually Static systems - alternative solutions Adaptive systems points of variation Automatically Collect from agents Send request Collect from users AN AN Receive request Example adapted from the RE06 keynote given by John Mylopoulos 17

18 Softgoals Good Quality schedule Minimal effort AN Minimal conflicts AN AN Good participation AN Matching effort Collection effort Minimal disturbances Accurate constraints Example adapted from the RE06 keynote given by John Mylopoulos

19 Modeling Softgoals Used to evaluate alternatives Helps (+) Makes (++) Hurts (-) Breaks (--) Example adapted from the RE06 keynote given by John Mylopoulos

20 Contributions to Softgoals Good Quality schedule Schedule meeting Minimal conflicts AN AN AN Choose schedule + + Good participation Manually Automatically Example adapted from the RE06 keynote given by John Mylopoulos

21 Minimal effort AN Collection effort Construction Process AN Matching effort Minimal conflicts AN Good Quality schedule AN Good participation Minimal disturbances 1. efine Softgoals Accurate constraints 21

22 Minimal effort AN Collection effort Construction Process AN Matching effort By person Collect timetables AN By system Schedule meeting AN Choose schedule Manually Minimal conflicts AN Automatically Good Quality schedule AN Good participation Minimal disturbances Accurate constraints Collect from agents Send request Collect from users AN AN Receive request Example adapted from the RE06 keynote given by John Mylopoulos 2. efine Hardgoals Steps 1 & 2 may be iterative 22

23 Minimal effort AN Collection effort Construction Process AN + Matching effort By person Collect timetables AN By system Schedule meeting AN + Choose schedule Manually Minimal conflicts + + AN Automatically Good Quality schedule Good participation AN + Minimal disturbances Accurate constraints Collect from agents + Send request Collect from users AN AN Receive request 3. Evaluate hardgoal contribution to softgoals Example adapted from the RE06 keynote given by John Mylopoulos 23

24 Integrating Goals with Other Models 24

25 Integrated Use of Goals KAOS Refining goals into requirements 4 models Goal Agent Operationalization Object

26 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

27 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

28 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

29 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

30 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

31 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

32 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

33 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

34 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

35 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

36 KAOS Goal Model Agent - active system component Object - inactive system component Operation - an action an agent takes to achieve a goal Requirement - a goal for which an automated component is responsible Expectation - a goal for which a human is responsible Safe Transportation Maintain Knowledge of Other Train Locations Manually Radio Other Train Conductors Radio Automatically Sense Other Train Locations Radio Other Train Human Software Sensor Train

37 KAOS Agent Model Objective: depicts agent responsibilities Agent models can be inferred from goal models Maintain Knowledge of Other Train Locations Software Sensor Manually Radio Other Train Conductors Automatically Sense Other Train Locations Automatically Sense Other Train Locations Human Human Software Sensor Manually Radio Other Train Conductors Goal Model Agent Model 27

38 KAOS Agent Model Objective: depicts agent responsibilities Agent models can be inferred from goal models Maintain Knowledge of Other Train Locations Software Sensor Manually Radio Other Train Conductors Manually Radio Other Train Conductors Automatically Sense Other Train Locations Automatically Sense Other Train Locations Human Human Human Software Sensor Manually Radio Other Train Conductors Goal Model Agent Model 27

39 KAOS Agent Model Objective: depicts agent responsibilities Agent models can be inferred from goal models Maintain Knowledge of Other Train Locations Software Sensor Manually Radio Other Train Conductors Automatically Sense Other Train Locations Automatically Sense Other Train Locations Human Human Software Sensor Manually Radio Other Train Conductors Goal Model Agent Model 27

40 KAOS Agent Model Objective: depicts agent responsibilities Agent models can be inferred from goal models Maintain Knowledge of Other Train Alert Locations Passengers to elays Software Sensor Manually Radio Other Train Conductors Human Automatically Sense Other Train Locations Human Automatically Sense Other Train Locations Manually Slow the Train Human Software Manually Radio Other Train Conductors Sensor Manually Radio Other Train Conductors Goal Model Agent Model 27

41 KAOS Agent Model Objective: depicts agent responsibilities Agent models can be inferred from goal models Maintain Knowledge of Other Train Locations Software Sensor Manually Radio Other Train Conductors Automatically Sense Other Train Locations Automatically Sense Other Train Locations Human Human Software Sensor Manually Radio Other Train Conductors Goal Model Agent Model 27

42 KAOS Operationalization Model Objective: specifies the operations that agents must perform to achieve the goals Safe Transportation Train oors Closed While Moving Close oor Request Train Begins to Move Close oor oor 28

43 KAOS Operationalization Model Objective: specifies the operations that agents must perform to achieve the goals Train oors Closed While Moving Safe Transportation Train oors Closed While Moving Train Begins to Move Close oor Request Train Begins to Move Close oor Close oor oor oor 28

44 KAOS Object Model Objective: further specifies objects used in the goal model The syntax is similar to a that of a UML class diagram Following Track On Track Train On TrackSegment hasgate Gate Speed Loc SpeedLimit Loc Status Switch position 29

45 Integrated Use of Goals KAOS Refining goals into requirements 4 models Goal Agent Operationalization Object i* Relates goals to the organization context 2 models Actor (Strategic) ependency Model Actor (Strategic) Rationale Model

46 i* Actor ependency Model ependencies between actors - An actor is a black box Attends (p,m) Initiator X epender O Be Scheduled(m) Enter aterange (m) LEGEN ependee Resource ependency Open (uncommitted) Task ependency Goal ependency Softgoal ependency X Scheduler Critical Attends (ip,m) Proposed ate(m) Assured (Attends (ip,m)) Agreement (m,p) iagram from Towards Modelling and Reasoning Support for Early-Phase Requirements Engineering by Eric Yu Enter Availates (m) Participant ISA Important Participant 31

47 i* Actor ependency Model ependencies between actors - An actor is a black box Initiator X epender O Be Scheduled(m) Enter aterange (m) LEGEN ependee Resource ependency Open (uncommitted) epender O X Task ependency Goal ependency Softgoal ependency Scheduler Critical Attends (p,m) Attends (ip,m) LEGEN ependee Open (uncommitted) Proposed ate(m) Assured (Attends X (ip,m)) Resource ependency Task ependency Goal ependency Enter Availates (m) Participant Agreement (m,p) Softgoal ependency Critical ISA Important Participant iagram from Towards Modelling and Reasoning Support for Early-Phase Requirements Engineering by Eric Yu 31

48 i* Actor ependency Model ependencies between actors - An actor is a black box Attends (p,m) Initiator X epender O Be Scheduled(m) Enter aterange (m) LEGEN ependee Resource ependency Open (uncommitted) Task ependency Goal ependency Softgoal ependency X Scheduler Critical Attends (ip,m) Proposed ate(m) Assured (Attends (ip,m)) Agreement (m,p) iagram from Towards Modelling and Reasoning Support for Early-Phase Requirements Engineering by Eric Yu Enter Availates (m) Participant ISA Important Participant 31

49 i* Actor Rationale Model Internal actor relationships - An actor is a white box Quick Organize Be Scheduled Schedule Initiator! +! + +! LEGEN Goal Task Resource Soft!Goal Task!ecompo! sition link Means!ends link Contribution to softgoals Actor Actor Boundary LetSceduler Schedule Low Effort Enter aterange Be Scheduled Find Agreeable Slot Scheduler Schedule Obtain Agreement Merge Availates Attends Obtain Availates Enter Availates Proposed ate Agreement Participant Attend Quality (Proposedate) + Richer Medium! + + Arrange Convenient (,ate) AgreeTo ate Participate In FindAgreeable ateusing Scheduler Agreeable (,ate)! Low Effort + User Friendly + FindAgreeable atebytalking ToInitiator iagram from Towards Modelling and Reasoning Support for Early-Phase Requirements Engineering by Eric Yu 32

50 i* Actor Rationale Model Internal actor relationships - An actor is a white box Quick Organize Be Scheduled Schedule Initiator! +! + +! LEGEN Goal Task Resource Soft!Goal Task!ecompo! sition link Means!ends link Contribution to softgoals Actor LetSceduler Schedule Low Effort Enter aterange Be Scheduled +! Find Agreeable Slot Scheduler Schedule Obtain Agreement Merge Availates LEGEN Goal Attends Task Resource Soft!Goal Task!ecompo! sition link Means!ends link Contribution Obtain to Availates softgoals Actor Actor Boundary Enter Availates Proposed ate Agreement Participant Attend Quality (Proposedate) + Richer Medium! + + Arrange Convenient (,ate) AgreeTo ate Participate In FindAgreeable ateusing Scheduler Agreeable (,ate)! Low Effort + User Friendly + FindAgreeable atebytalking ToInitiator Actor Boundary iagram from Towards Modelling and Reasoning Support for Early-Phase Requirements Engineering by Eric Yu 32

51 i* Actor Rationale Model Internal actor relationships - An actor is a white box Quick Organize Be Scheduled Schedule Initiator! +! + +! LEGEN Goal Task Resource Soft!Goal Task!ecompo! sition link Means!ends link Contribution to softgoals Actor Actor Boundary LetSceduler Schedule Low Effort Enter aterange Be Scheduled Find Agreeable Slot Scheduler Schedule Obtain Agreement Merge Availates Attends Obtain Availates Enter Availates Proposed ate Agreement Participant Attend Quality (Proposedate) + Richer Medium! + + Arrange Convenient (,ate) AgreeTo ate Participate In FindAgreeable ateusing Scheduler Agreeable (,ate)! Low Effort + User Friendly + FindAgreeable atebytalking ToInitiator iagram from Towards Modelling and Reasoning Support for Early-Phase Requirements Engineering by Eric Yu 32