Regarding abstract/concrete of the extending class in the «Specify» relationship: Abstract classes are not mapped to YANG container/list. The Mapping Guidelines need to be corrected in some places. The target text in Figure 5.31 of the Modelling Guidelines need to be corrected.
specify-augment.pptx has been corrected during the discussion and will be presented again in the next meeting. Remaining question: Should the UML Modelling Guidelines allow concrete specification classes like TAPI is using it?
The current guidelines shows abstract classes (TopologyContext and CepList)
Only abstract extending classes are allowed
On the previous call Italo had explained that in the yang tapi-topology:topology-context should not be present (as marked below)
Discussion on whether the tool is parsing the path.
Noted that the tool was elimination the green text above (consistently)
AI: Scott Mansfield Need to test to determine whether the tool can construct the path with no content in specify.
Not clear if the tool analyses the string or parses the model structure or some hybrid.
After some discussion it was agreed that the current guidelines were actually not correct and the target should also not include TapiTopology:TopologyContext:_topology (as highlighted below) when the augmenting class is abstract.
Tapi currently augments with concrete classes. In this case the path element "TapiTopology:TopologyContext:_topology" should be included (as below).
The current guidelines do not cover the case of augmentation with a concrete class.
It was agreed that the guidelines should be updated to recommend augmentation with an abstract class but allow augmentation with a concrete class.
Bernd Zeuner Upload corrected slides to cover path without "TapiTopology:TopologyContext:_topology" in target string for augmentation with an abstract class.
Bernd Zeuner Update guidelines to match slide update and to cover allowance for augmentation with concrete class.