search

buildingSMART / Sample-Test-Files

Sample files of various formats and schema versions.
Other
251 stars 115 forks source link

IfcSectionedSolidHorizontal and Directrix #100

Closed BernhardWehrle closed 3 years ago

BernhardWehrle commented 3 years ago

Hi all

To which entity must I set the relation "Directrix" in the entity "IfcSectionedSolidHorizontal" in Ifc4x3RC2. I think i should not use IfcAlignmentCurve because it obsolate.

IfcSectionedSolidHorizontal

SergejMuhic commented 3 years ago

Hi Bernard,

do you want to test sweeping along an alignment or would it be enough to test along any curve? I would propose a more generic curve for now just to test the concept itself because sweeping has not yet been finalized in RC2.

pjanck commented 3 years ago

Hi Bernhard,

you are correct, IfcAlignmentCurve should not be used anymore.

Following the comment from Sergej, if you're using a simple line, it should be enough to check the concept of 3D swept geometry. The concept is not finalized, so there is no definitive answer yet to give. We still have open issues as to "what does sweeping along a IfcGradientCurve mean".

In my opinion, one would link IfcGradientCurve if you're sweeping along 3D, or IfcCompositeCurve if you're sweeping along 2D.

BernhardWehrle commented 3 years ago

Hi Stefan

for the unit test IfcCourse I will set the link to IFCCOMPOSITECURVE

33 = IFCCOMPOSITECURVE((#39), $);

52 = IFCSECTIONEDSOLIDHORIZONTAL(#33, (#54, #58), (#53, #57), $);

In the future will it be possible to set a link to the semantic part of the alignment?

EDIT: removing email signature.

TLiebich commented 3 years ago

In the future will it be possible to set a link to the semantic part of the alignment?

no, this is not planned (at least not directly), any geometric representation should be computable purely from its geometric definition. An open question: should we plan to have an INVERSE backpointer from the resulting curve to its hor/ver alignment definition in the semantic part.

jmirtsch commented 3 years ago

no, this is not planned (at least not directly), any geometric representation should be computable purely from its geometric definition. An open question: should we plan to have an INVERSE backpointer from the resulting curve to its hor/ver alignment definition in the semantic part.

I would support the inverse attribute for this, similar to IfcLinearPositioningElement attribute for axis. I'm not sure if it makes sense to change the base class for vertical and horizontal alignments (and cant) to to IfcLinearPositioningElement to reuse this attribute.

pjanck commented 3 years ago

Opinion: introduce a new Inverse, do not change the base class.

EDIT: moving the discussion about the possible inverse to the linked issue on IFC4.3.x-development repository.

pjanck commented 3 years ago

Closing, since the discussion moved to other repository.