Update deserialization notes for FMU states

[pmai]

Clarified deserialization requirements and limitations for FMU states, closes #2186.

[chrbertsch]

Looks good.

(@pmai: Shall we add a link to FMI-LS-REF with the possibility to reference platform specific serialization files? Or is it a bad idea to reference layered standards from the core standard?)

[pmai]

Looks good.

(@pmai: Shall we add a link to FMI-LS-REF with the possibility to reference platform specific serialization files? Or is it a bad idea to reference layered standards from the core standard?)

I would not do so just for this, and not as long as FMI-LS-REF is not yet released. I can see the benefit of pointing to layered standards for particular issues, but for FMI-LS-REF I would point to that for more generally interesting things (e.g. to embed multiple parameter sets, etc.), and as part of that, maybe mentioning stored state then also makes sense.

If a new LS for specifying stored state capabilities were to be created, then I think it would make sense to point to that, once released, from the current text here.

[KarlWernersson]

Looks good