Probably obvious to some, but thought I'd share, for those in a similar learning curve.
While it may seem a good idea at first glance, it also introduce limitations like stairs walls not always bound between slabs, and in "stair like" buildings it will quickly become a nightmare.
This is great! I'm sure there are many uses of this that people will discover as they continue to use Blender.
You know me, now I'm wondering how we can format the IFC file, such that these parametrics are recreated upon import. :)
Personally I've always been very wary of constraints since when they break they're so hard to locate and fix (and they break a lot in Revit if you're not very careful). I prefer to focus on good modelling (of course) and easy visualization of existing simple constraints and geometry. So BIMOne Color Splasher for Revit for example can color objects based on their properties, for example their offset from their level or which level they are hosted on. This type of functionality is worth gold for the quality of a model and I look forward to learning how it is/will be implemented in Blender & BBIM (I'm gonna keep pushing that name :-) ).
I generally agree with you, @duncan . Constraints also can be the cause of a lot of unintended mistakes in large projects. Coming from both sides - the manual mesh world and the parametric solid world, I know my preference :) That said, I have seen some clever constraints - and hopefully someone more engaged and creative than me can discover them!
By the way, there already is a feature to colour by attribute, property, or class in the BlenderBIM Add-on. In the search panel, there is a paintbrush next to each search filter.
I found drivers in blender a bit unsafe, eg when you copy objects, as the constraint values "data path" are string based, they does not update to copy but still rely on source objects.
Not even talking about linked copy simply not allowing different values in drivers.
So unless we find a pretty clever way to implement such things i would try to avoid them.