IfcDirectrixDerivedReferenceSweptAreaSolid semantic definition

Dear all. I find that the definition of IfcDirectrixDerivedReferenceSweptAreaSolid could benefit from some additional clarification.
http://ifc43-docs.standards.buildingsmart.org/IFC/RELEASE/IFC4x3/HTML/lexical/IfcDirectrixDerivedReferenceSweptAreaSolid.htm

@SergejMuhic @peterrdf

I don't really know where the current description comes from. Doesn't seem to be in https://github.com/bSI-InfraRoom/IFC-Specification/tree/develop/IFC4x3/Sections/Resource%20definition%20data%20schemas/Schemas/IfcGeometricModelResource/Entities/IfcDirectrixDerivedReferenceSweptAreaSolid

I have some points below, but I hope everyone can have another look at this. This is at the core of the new additions in 4.3 it needs to be as unambiguous and clear as possible.

---

> In most cases the case

In most cases?

> the IfcDirectrixDerivedReferenceSweptAreaSolid has exactly the same behavior as IfcFixedReferenceSweptAreaSolid

(Does it really warrant another entity type then when it in many cases behaves exactly the same? The risk of this kind of gratuitous subtyping is that it is easily overlooked due to liskov substitution.) <-- we're not changing this now, just for future consideration.

> except when the Directrix not only defines a tangent direction but a tangent plane for each point on the curve

How would an implementer assert this?

> Like for example

That reads a bit informal. Is this an example, clarification or normative statement?

> in the case of a IfcSegmentReferenceCurve

Isn't this then the only case?

> the change in y directioin of the tangent plane is added to the fixed reference

Is this alteration of FixedReference constant for the entire sweep? Probably not? Then I think it should be mentioned.

Is the direction supposed to be normalized or can it introduce scaling? Applies to supertype as well I guess.

> The change in y direction at the start of the directrix is defined to be 0 independent from StartParam value, this means the change can be non-zero at the start of the resulting Swept Area Solid.

... in case of a non-zero StartParam value?

The magnitude of the change?