Non-optional SET [0:?]

[iegorychev]

Hi All,

ENTITY IfcCurveBoundedPlane
SUBTYPE OF (IfcBoundedSurface);
BasisSurface : IfcPlane;
OuterBoundary : IfcCurve;
InnerBoundaries : SET [0:?] OF IfcCurve;
END_ENTITY;

Validation "Validate required attributes assigned" should return TRUE, if all non-optional attributes have values, or if object has non-optional attributes [only]; FALSE if any non-optional attribute has no value (10303-22).

I've just got a file with InnerBoundaries=$ which doesn't pass validation as it should be () at least in correct case. =)

Shouldn't such cases for explicit attributes be transformed into InnerBoundaries : OPTIONAL SET [1:?] OF IfcCurve; ?

[TLiebich]

Hi @iegorychev

to me it is a known anomaly - in IFC development we always used optional set[1:?], with the exception of those entities taken over from STEP (mainly part 42). As it is the case here. Changing it now would lead to incompatibilities () vs. $ in the part 21 file structure. So it should be an issue addressed with a next major version to unify modeling style and also in general to minimize the often redundent ways to express geometry..

[SergejMuhic]

Ah yes, this is definitely a pain for us implementers. I would support this. Empty lists still take memory so either there is something in it or we do not need it.

EDIT: We are also missing a lot of WRs from the definition (similar to IfcCompositeCurveOnSurface). I think this should be addressed for proper implementation of these definitions.

[SergejMuhic]

@TLiebich I am not sure if you are talking about IfcCurveBoundedPlane here:

to minimize the often redundent ways to express geometry

If you are, what is redundant about this definition?

[TLiebich]

@SergejMuhic - this statement was not specific about IfcCurveBoundedPlane. More a general observation.