Tuesday, September 6, 2011

Is IFC good enough?

It’s better than what it used to be.
Fewer rogue elements are left floating in space after conversions, storey-settings come through OK, so do most walls, slabs, doors and windows.
On the downside though, even a relatively small file still converts into thousands of objects.
Topographical meshes, stairs, curtain-walls turn into objects, parametrical and with some intelligence retained, nonetheless with very limited ability to do modelling work with or around them.
Based on this one can deduct, that at least 3 groups of specimens that work as ‘tools’ in one program will only function as objects in another when imported as IFCs.

Should this be a worry?
Not necessarily. It is quite reasonable to have a complicated terrain-model created by one party inserted into another’s file as a relatively dumb object.
Similarly a stair and/or a curtain wall.

The problem I see is where elements that belong to multiple parties are at stake.
...most building elements. A bit like the column I wrote about previously, the column that was simultaneously everyone’s and no-one’s.

This issue is not insurmountable; there are numerous ways to move forward.
Step one however is to accept that building elements are a bit like children, belong to many people, all wanting to put their stamps on them...
... and they are people in their own right and owned by no-one!


  1. Don't mistake the quality of IFCs with the poor excuse for an IFC
    interpreter Autodesk have ship with Revit. IFCs are far from perfect, but that's like judging the quality of Shakespeare's writing by watching a production put on by a bunch of hyperactive monkeys.

    The reality is Autodesk are not serious about IFC, because if they were comprehensively implemented and adopted it would foster true competition in the BIM marketplace. IFC exists instead to satisfy the checkbox feature comparison needs of certain governments and organisations. It is not meant to work or be used, it just keeps Revit "in play" during any purchase or IT planning process.

    Given the immaturity of the BIM market this situation isn't surprising, and initiatives like IFC will not stand a true chance of succeeding until BIM has emerged from its current primordial ooze. Unfortunately IFCs were too far ahead of their time, and it is highly likely that by the time BIM achieves the same maturity as CAD it will be time to start over with a new "interoperability" initiative.

  2. Davids comment above is "spot on"!
    But not limited Autodesk Software. I have been doing some research lately, targeted at determining what 2D annotation can be expected to be useful in various SW. Made a file from ArchiCAD, and put it into every SW I could get my hands on. And also into ArchiCAD (13,14 and 15)
    None where perfect, but if you added up the results, it showed that all info exept, fill definition, was in the ifc file! So the problem is not much the ifc standard, but implementation, and then mostly the import part of the implementation. It is expecialy interesting (in a positive way) to see how the Nemetschek ifc Viewer handled text. Try it youself!