Proposal to change description for CardinalPoint of IfcMaterialProfileSetUsage

This is a super useful attribute for structural modeling. The description for the `CardinalPoint` attribute of `IfcMaterialProfileSetUsage` is currently:

> Index reference to a significant point in the section profile. Describes how the section is aligned relative to the (longitudinal) axis of the member it is associated with. This parametric specification of profile alignment can be provided redundantly to the explicit alignment defined by **ForProfileSet.MaterialProfiles[*].Profile**.

However, `ForProfileSet.MaterialProfiles[*].Profile` is an `IfcProfileDef` and I think therefore it is referring to the wrong thing, because:

1. In the examples shown in the [Beam standard case](https://standards.buildingsmart.org/IFC/RELEASE/IFC4/ADD2_TC1/HTML/annex/annex-e/beam-standard-case.ifc) example and [beam varying cardinal points](https://standards.buildingsmart.org/IFC/RELEASE/IFC4/ADD2_TC1/HTML/annex/annex-e/beam-varying-cardinal-points.ifc) example (and perhaps others too, I haven't checked), the profile alignment is explicitly defined in the IfcSweptAreaSolid.Position, not the profile def.
2. Depending on the profile def, the profile def may have no position, one position, or multiple positions. It seems a bit tricky to derive.

I propose to change to:

> Index reference to a significant point in the section profile. Describes how the section is aligned relative to the (longitudinal) axis of the member it is associated with. This parametric specification of profile alignment can be provided redundantly to the explicit alignment defined by **the shape representation's extrusion position**.

Ping @berlotti @TLiebich @aothms @jmirtsch @Jesusbill @krande