Optional attributes

[michelangelo-acca]

It seems that attributes that were optional in the previous schema

now are not

In the provided Linear Placement example, a default value of 0 is used. But maybe sometimes "optionality" is necessary?

[tomvandig]

@michelangelo-acca Indeed I think we will have to add this, it is inevitable also for use cases like schema-evolution, and its much stronger to put it in the schema type system than to have special values represent optionality.

I will think about this after the projects.

[aothms]

Agreed. @tom if I can make requests I'm also interested in schema type references, so that we're all using the same point definition for example and not array vs x/y/z object and all using the same data reference and not just an object with a coincidental `"ref"` key in it. Sent from a mobile device, excuse my brevity. Kind regards, Thomas Op do 12 jun 2025 16:03 schreef Tom van Diggelen ***@***.***>:

…

*tomvandig* left a comment (buildingSMART/IFC5-development#45) <#45 (comment)> @michelangelo-acca <https://github.com/michelangelo-acca> Indeed I think we will have to add this, it is inevitable also for use cases like schema-evolution, and its much stronger to put it in the schema type system than to have special values represent optionality. — Reply to this email directly, view it on GitHub <#45 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAILWV7UZF7L6RX5IMZA6OT3DGCDZAVCNFSM6AAAAAB7FCC3H6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDSNRWHEYTMNBZGA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[berlotti]

This is also related to the "IDS module" that we envision to be part of the IFC 5 series.
Optional and mandatory could be an IDS style definition exchanged in Ifcx format.

[tomvandig]

@michelangelo-acca @aothms
See #50