  means-ends link types
The SR model also provides for several means-ends links. The "end" can be a goal, task, resource, or softgoal, whereas the "means" is usually a task.(p34)
A means-ends link is an application of a rule in a context in which the agent believes the applicability coindition to hold. (p36)
For example, MakeMedicalAssessentLocally can be a GT-Rule [Goal-Task rule] which has been applied as the means-ends link linking the goal node MedicallyAssessed and the task node MakeMedicalAssessement, with the applicability condition stating that the actor doing the assessement must have the requisite knowledge for doing medical assessements.
» See also: rules
 Goal-Task Link
In a Goal-Task Link (GTLink), the end is specified as a goal, and the means is specified as a task. This task specifies the "how" through its decomposition into components.
Resource-Task Link
In a Resource-Task Link (RTLink), the end is specified as a resource, and the means is specified as a task.(p34)
Softgoal links
There are two common link types involving softgoals.(p34)
- Softgoal-Task Link
A Softgoal-Task Link (STLink) have a softgoal as the end, and a task as the means. (p34)
For example, to achieve FastTurnaround (the softgoal), a claims manager can MakeMedicalAssessement herself (a task). On the other hand, letting a medical assessor do it (the task) leads to a negative contributions towards FastTurnaround.(p34)
- Softgoal-Softgoal Link
The Softgoal-Softgoal Link (SSLink) permits the development of a means ends hierarchy of softgoals, until eventually some softgoals are addressed by linking to tasks.(p35)
- contribution types
Links involve softgoals require an extra attribute to indicate the type of contribution - positive or negative, enough or not enough.(p34)
[Chung93] provides a framework for qualitative reasoning using these concepts, using a multi-valued evaluation scheme (satisficed, denied, conflict, unknown, etc.) (p34)
» See also: notion of non-functional requirements
- positive
- negative
- enough (sup)
- not enough (sub)
at actor boundary incoming link
At an actor boundary, an incoming dependency link [the actor being a dependee] is also an implicit means-ends link, with the dependum being the "end"(p35)
» See also: at actor boundary outgoing link
other means-ends link types
Other means-ends link types are possible as a result of the combinations of elements types for the means and for the ends. (p35)
[As example lets pick out two ...]
- Task-Task Link
In a Task-Task (TTLink), both end and the means are tasks. This is allowed due to the inherent openness (incompleteness) assumed by the modelling framework.
Even though a task specifies the "how", it is still possible to hjave alternatives arising from it becaues further components can be added.(p35)
An example of this is a specialization (IS-A) relationship between two tasks. Thus a task-task link can still be seen as a means-ends relationship. There can be multiple alternatives of tasks (the means) which accomplish the "end" task. This is unlike the modelling of the decomposition of non-intentional activities into more detailed none-intentional sub-activities (cf. RML [Greenspan84]).(p35)
- Goal-Goal Link
Links ususally have task as the means specified. However, it is sometimes desirable to have restricted form of a task as the "how". For example, goal reduction: given a goal as the end, the means could be specified as a conjunction of subgoals (with no other types of components). This would be a Goal-Goal Link (GGLink)(p35)
|
