- created example for ModelVariables in modelDescription.xml

[PenndorfT]

This is an example how I interpret the required configuration if an internal XCP slave. The XML schema requires a different kind of start value configuration for simple types (Boolean, UInt16) and strings. According to https://github.com/modelica/fmi-standard/blob/main/schema/fmi3Variable.xsd#L331 it is not enforced that exact one start value is given (minOccurs="0" maxOccurs="unbounded"). This should be described in the layered standard as an obligation of the importer.
I will take this a todo for me in addition to referring to the example in the document.

[PenndorfT]

When reading table 21, section 2.4.7 of the FMI standard I think the Start tag can also be left out. In this case the FMU implementation sets 127.0.0.1 as default. Comparing with the given example for the array of string parameters there it seems there must be a check for a scalar variable that can have a start value.

[PenndorfT]

I have tried the following example with fmpy 0.3.15

    <String name="org.fmi-standard.fmi-ls-xcp.ListenIpAddress"
            valueReference="2"
            causality="structuralParameter"
            variability="fixed"
            description="IP address where the XCP slave listens for XCP protocol commands.">
        <Dimension start="6"/>
        <Start value="127.0.0.1"/>
        <Start value="127.0.0.2"/>
    </String>

It's accepted by the generic fmpy importer and should be rejected by the fmi-ls-xcp implementation of the importer.

[pmai]

Structural Parameters must always have a start value, so the Start element for the address is mandatory. The number of start values provided is determined by the dimensions of the variable, which for a scalar means exactly one value. This is defined by the base standard, so the XCP LS might at most want to point out that the required variables are to be scalar, i.e. not arrays.

[PTaeuberDS]

I agree with Pierre. @PenndorfT Is that fine for you and can you change this in the document?

[PenndorfT]

Obsolete after discussions in #19 and #20