Use and integrate STMD for individual models

[ClemensLinnhoff]

The content of the STMD files not only represents the Credible Simulation Process, but also the Credible Modeling Process.

For SRMDs there is an explicit specification on how to integrate an SRMD file into an FMU in the documentation.

This is lacking for STMDs. The specification says

Any STMD-based package MUST include an STMD XML file under the extra/org.ssp-standard.ssp-traceability.stmd/SimulationTask.stmd name. The file MUST be a valid STMD file according to the specification below.

In my interpretation a model FMU, developed based on the Credible Modeling Process, is also an "STMD-based package". Therefore, there should be a section about the integration of STMDs into FMUs, analog to the SRMD specification.

[pmai]

An STMD-based package in the context of the spec is always an SSP package, as this is a layered standard on top of SSP, not FMI, and it explicitly defines the package as such (c.f. section 9, as well as section 3.6). Only if packaging in other formats is explicitly mentioned is there an alternative to the SSP package.

As the current spec does not yet address the CMP (only CSP and CDP are handled), there is currently not yet even a decision whether STMD will be the format for the CMP (or whether a new format, e.g. "MTMD") will be defined - as is currently being explored. This is to be handled in the next release after the currently planned one, as current practice in CMP is still fluid.

Also the exchange of information at the full STMD/SMMD level is not always going to be the goal in any case, where usually SRMD-based exchange of information at the level of e.g. MIC-Core is the more likely avenue, and for that, which is currently specified.

And, depending on the use case for CMP, there is also the likelyhood that an SSP package will be preferable, to include the referenced resources and design information in a more clear manner.

So this leaves a number of questions open, which will likely be addressed in the release supporting the CMP, which will be the one after the current initial release.

[hubertus65]

So it is the intent of the ssp-ls-traceability designers, that STMD is for SSP only, and not for FMI? This may be more a question for the Modelica Standards Coordination meetings, but I agree with Clemens that it would be just as useful in FMI. @chrbertsch, what is your take on this?