The 15th of November 2022 I hosted the first day of Capella Days 2022. 3 talks were presented that day by Thales Defense Mission Systems, TNO-ESI & Thermo Fisher Scientific, and Siemens.
Some questions for Thales DMS were not answered during the live session. Here is a wrap up of all the questions with the answers by the speakers Guillaume Journeaux and Karine Pellen.
Q What do you mean when you talk about data packages? You generate formal documentation extracting from capella? What do you include: requirements, interface requirements specs. Behaviour, others…
Data packages produced at the end of an incremental engineer phase is a set of engineering data (Capella model, Doors database, existing interface specification documents, etc … ). It includes all data required by the next engineering level. We produce/generate documents only if required or useful.
Q Do you use all four levels in Capella, or just one for e.g. S-SSDD design?
System Analysis level is part of the “SSS engineering data level”
For ARCHANGE, we’ve used Physical Architecture for the “S-SSDD engineering data level” (but we could use Logical Architecture instead). We use only 1 architecture level to define system architecture.
Q How to verify the corresponding non-functional requirements based-on related architecture and functional requirements?
Non-functional requirements can be verified in test sheets driven by functional chains or in dedicated test sheets (it depends on non-functional topics)
Q How do you link between the different levels?
The main way to link SSS and S-SSDD levels is through requirements (formal links between 2 Doors database). Since requirements are linked to functions and functional exchanges, you can deduce all other kind of artefacts links.
We do not use Capella internal links between engineering levels
Q: Do you have a specific tool to link capella functions with Doors requirements?
Yes, we’ve developed a specific tool to be able to link in an easier way Capella models artefacts and Doors requirements
Q: Does the System Engineering correspond to System Analysis and Logical Architecture levels of ARCADIA method ? We cannot easily recognize the ARCADIA levels in this modelling.
System Analysis level is used for “SSS engineering data level” and Physical Architecture for “S-SSDD engineering data level”.
Q: What version of Capella Were you using for this project? Did you transfer to later versions?
The latest version available in our Thales engineering environment (4.5.2)
Q: How do you manage interfaces refinement iterations ? Do you go back systematically on the system model ? or not ?
What do you mean behind “interfaces refinement iterations” ?
If the interface reference is in the Capella model, it is always generated from the model if it needs to be updated.
Capella Model contains the “true” data
Q: Do the Sub-system capabilities are “named/worded” consistently regarding the System’s ones? e.g. if you use a computer (a sub-system) to realize part of the system functions, its capabilities may be “generic computer” ones. If so, do you try to word them with a generic name or keep to the system original name (on purpose)?
System and subsystems share the same capabilities / functional chains list. The goal is to give to sub-systems the view of how they are involved in global system capabilities. However, each subsystem can have its own engineering data (including generic capabilities) and link their own data to the one received from the system
Q:How do you capture non-functional constraints in Capella? Function? Constraint? Only doors’ requirement?
We use documents and requirements that are not formally connected to the Capella Model for non-functional constraints
Q: Did you couple strongly capella with Jira ? Using which kind of mechanisms?
Capella and JIRA are not strongly coupled. Both are linked for functional chains and system component. Fortunately, the system components list and functional chain list are very stable and maintain the consistency manually is not time-consuming
Q: Can you speak more about the engineering data model? Is it just a picture? Is it just lists of items? Do you “run” it (the model) in some way?
The engineering datamodel is a way to communicate and help us to structure our engineering process and environment. We don’t “run” the model.
Q: Can we generate AIT test cases from capella model?
Tests cases are not generated from Capella, they’re written manually. However, functional tests cases are written from functional chains : each functional exchanges at the edge of the system is translated in a test step (an input FE is a stimulation step and a output FE is a check step). Once functional analysis has been matured in a collaborative way, writing the functional test sheet is easy.