Open jasonweiss opened 8 years ago
I suggested that for the AOESW-IRIS ICD because AOESW has the C-based RPG. So it was impossible to see what was being used by the RPG vs AOSQ in the Word document that was written.
The AOSQ is a component in CSW (a sequencer) so is modeled in ICD-DB. I believe RPG will also need an Assembly/HCD to integrate it with CSW so an RPG component makes sense to me. All I was really saying was breakup AOESW into its CSW components, which is what is done in ICD-DB models. The AOESW-IRIS ICD is still between level 2 subsystems, not level 3 components within the same subsystem.
So I don't think that your example of AOESW to IRIS is an accurate comparison to what you are asking for.
I can see why you might want to do this since some components within IRIS are being produced by different teams, which I imagine is your motivation, no?
Ok, kim, I see. Nevermind on the AOESW-IRIS ICD then. I suppose if we migrate to the ICD-DB for that interface, the subsystem to component functionality will give us what we need already.
Yes, your last sentence is correct. Although level-3 ICDs are only an internal requirement, it'd be nice to be able to use the ICD-DB for this purpose.
In the current icd-db 0.9 it is possible to specify --subsystem
and --component
. I think that a valuable addition that would facilitate "Level 3 Subsystem" ICD generation is to simply allow --component
to take array arguments, as well as some kind of label. For example, in IRIS I may want to know the interface between the CSRO and the Imager. From the point of view of TMT as a whole, the OIWFS and Imager are just collections of assemblies within IRIS. So, to specify CSRO, something like:
--subsystem IRIS --component CSRO:oiwfs-adc-assembly,oiwfs-detector-assembly,oiwfs-poa-assembly,rotator-assembly
where the CSRO
is the label given to that collection of assemblies (and which would appear in a generated PDF).
For IRIS, we'd like to construct ICDs between level-3 subsystems, which are subsystems of IRIS, such as Imager, IFS, CSRO, etc. Each of these subsystems may contain more than one component, so the ability to produce ICDs between components isn't sufficient for us. The implementation of this also help us follow Kim's suggestion on the AOESW-IRIS ICD by breaking the AOESW up into subsystems (AOSQ, RPG, PSFR) and presenting the IRIS interface to each of those.