SAF with SysML v2

[bleckenmarvin]

Are there currently efforts to transfer the SAF specification to SysML v2 or is the transfer to SysML v2 planned for the near future? Some tool providers already support/implement SysML v2 or will release new versions in the next one to two years, so it makes sense/is worthwhile to transfer the SAF specification to SysML v2.

[odungern]

In the CASCaDE team we are working on a transformation from Cameo to PIG. It should be applicable to any Cameo Profile and Project ... but is isn't complete, yet. The difficult part is to find and assemble all the UML/SysML artefacts to build a graph. Once it is a graph, transforming it to a SysML v2 Library shouldn't be a major task.

To get there, we would need Cameo example (better test) projects and sparring partners to discuss the results. Especially a systematic set of test-cases would speed up things a lot. Any help is appreciated.

[haarer]

I think we need more than a transformation. SAF should make use of native SysML2 concepts.
IMO a plain transformation would not be sufficient.

[odungern]

Then, it is a new implementation in the SysML v2 context. Will potentially yield a better result than a mere transformation.

In any case we need (and will have) a general transformation from SysML v1 and SysML v2 to PIG. Will it produce the same result ... very probably not. Interesting topic ... not easy.

[sonofmbse]

We have developed a concept for implementing SAF as a library extension in SysML V2 based on textual notation (see image). It is very interesting that we can use the stakeholder/concern/viewpoint/view concept from SysML V2 to create the entire framework grid in SysML V2. However, moving and linking the so-called “stereotypes” in SysML V1 is a major step that most graphical tool manufacturers do not yet support. The first step would be to implement that concept in the textual notation.