UML modeling of an augmentation containing directly a "list" statement

[bzeuner]

The «Specify» abstraction association in UML is mapped to an "augment" statement in YANG. Since the abstraction association does not have a multiplicity, it is not possible to define that the specifying object class has a multiplicity of e.g., [*].

See following example from ieee802-dot1q-cfm-bridge:

The corresponding YANG specification is:

[nigel-r-davis]

On the IISOMI call 19 Dec 2018, it was suggested that the presence of a "partOfObjectKey" key stereotypes against at least one attribute in the class indicates a list. If there is no "partOfObjectKey present in the class then this indicates a container. This was not agreed and will be discussed further on the next IISOMI call.

[karthik-sethuraman]

All classes first gets mapped to a grouping and then a specify gets mapped to a container with “uses grouping” sub-statement. Using “partOfObjectKey” may run into problems where same class is used sometimes as list attribute and sometimes as container attribute. The immediate way to solve this is to add multiplicity attribute to the specify stereotype. The correct semantic change would be that specify is no longer an abstraction relationship, but an composition association (variation of extended composite)

[XingZhao-CATR]

Karthik's solution: in UML OpenModel_profile, rename <> to <> stereotype to add the mutiplicity and no longer use abstraction.

[kamlam]

Xing supposes to type:
Karthik's solution: in UML OpenModel_profile, rename the "specify" stereotype to "augment" to add the mutiplicity and no longer use abstraction. Somehow, text within the double angle bracket << and >> got lost.