Clarification and Example about using UML semantic variation points

Question

As I read through UML specification 2.5(still Beta)

Some tools may use the different Images for different purposes: the icon replacing the box, fo r the reduced-size icon inside the box, for icons within tree browsers, etc. Alternatively, depending on the Image format, tools may choose to scale one single Image into different sizes for these different purposes. Some model elements already use an icon for their default presentation. A typical example of this is the Actor model element, which uses the “stickman” icon. When a Stereotype with an icon is applied to such a model element, the Stereotype’s icon replaces the default presentation icon within diagrams.

Also read about it in UML certification(2007,page 155)

UML 2 uses several semantic variation points, which means that UML 2 leaves the semantics intentionally open at this point. The interpretation is left up to the environment in which the model is used.

My questions:

1)Could you please clarify semantic variation points through simple example (not related to tool as provided above)?

2)How should I use semantic variation points(I wonder if there is guidelines or limitations about How and where I can use it)?

Answer 1

Some things are strictly set. Some things are left opened to be set by the tools or modellers. For example,

Semantic Variation Points
- Precise semantics of shared aggregation varies by application area and modeler.
- The order and way in which part instances are created is not defined.

The first point will help you so: you can use the shared aggregation as you choose, only leave the principle understandable and stable for the whole project.

The second one will tell you not to try to show by the means of the class diagram the order of creation - you need other diagrams for it, or simple text.