Let's look first at the elicited Types. There are blocks Fox and Dog, and some ValueTypes represented (as a modelling choice) using Enumerations: ColorKind (carries literal 'brown'), SpeedKind (carries literal 'quick), and MoodKind (carries literal 'lazy').

Those enumerations have corresponding value properties: mood:MoodKind (in Dog), speed:SpeedKind and color:ColorKind (in Fox), with defaults that correspond to the description in the pangram.

There also some instances quick brown fox with classifier Fox and lazy dog with classifier Dog, with their Slots populated with the defaults.

What's that abstract base block Animal doing there? Nobody has even told us yet whether dogs and foxes are animals!

In Webel Parsing Analysis if you want to track elicited model elements that have not been explicitly mentioned in source text you can introduce Element-level custom tracking Stereotypes with the keywords «pa:implied» or «pa:assumed».

If we had additional information about animals we might know whether all animals have a mood, speed, and color. For now we'll assume this is correct, and so there are Generalizations from Fox and Dog to the abstract Animal and the value properties have versions in the abstract Animal with no defaults; the versions in Fox and Dog are re-definitions that carry specific defaults.

An Operation jumpOver(jumpee) has also been placed "up" in the abstract Animal from the start, where the 'in' Parameter jumpee represents any Animal that gets jumped over. The Operation is then redefined in Fox to have the default for its jumpee Parameter set to be the lazy dog instance.

What about snakes? Is a snake an animal? And if so, can a snake jump over other animals? And what about 'slithering' as an operation? Can a fox or dog slither like a snake? Can all animals slither()?

Such concerns are a matter for a fully-fledged parsing analysis recipe, and beyond the scope of this trail section on the basic SysML ElementGroup. The quick answer is that as one processes relevant text snippets from more domain source documents such wrinkles in the model do get ironed out well.

In any case, that's 14 elicited model elements already and a nice starting point for a robust model!

Even relatively informal elicitation of model elements using the SysML ElementGroup and a Parsing Analysis approach is extremely powerful.

We'll look next at how the tool represents an ElementGroup in the model tree browser.

Up next

SysMLv1.x: An ElementGroup in the model tree browser of MagicDraw/Cameo

Notes

[OPTION]{OPTION} Do not feel obliged to use the 'name' property of an ElementGroup if you don't feel it adds value.

[OPTION]{OPTION} The default display of all tagged values on the comment symbol for an ElementGroup can make diagrams cluttered, but you can extend it with a custom stereotype and customise the default visible tagged values.

[TIP]{INFORMATIVE} The Webel Parsing Analysis recipe for SysML1.x does not use the ElementGroup directly, it extends it as a user-defined Snippet stereotype with keyword «snippet» and with MagicDraw Customization.

[TOOL]{INFORMATIVE} In MagicDraw/Cameo the dashed line symbol from a UML Comment or SysMLv1.x ElementGroup to an annotatedElement is called an 'anchor'.

[TIP]{INFORMATIVE} An ElementGroup comment symbol may appear on more than one diagram type and may collect different kinds of anchored annotatedElement members from more than diagram.

[ASSERTION]{TIP} Even relatively informal elicitation of model elements using the SysML ElementGroup and a Parsing Analysis approach is extremely powerful.