A comment from a non-developer... I understand Eric's argument, which basically favours lightweight, specific formats that, as they are limited, allow to bypass their limitation by better support of e.g. im- and export to other, similar limited formated - just because they are simple to implement. On the other hand, I think Sean identified the solution to the complexity of a broader format by providing a consistent library support. In fact, my experience is somehow contradicting the idea of simple formats at least in 3d. I have a lot of problems with mesh-files like obj, 3ds, as the ease of implementing also led to a lot of semi-implementations that fullfill some task, but are not complete or even try to model additional functionality by misusing format specs. Other then expected I then just experienced how cool it is to download some parts in step-format from a vendor website, open it in freecad and have it there just as a native cad object. Also please do not forget that the decision on success of a cad format are not only about geometry. There are lots of specialized applications going beyond that, and lots of approaches to achieve something that step provides - a unified language to describe such additional entities. I am from the building side and see IFC, the various proprietary formats, all just to allow adding intelligence into the cad format that could be supported by step in a compatible way. Still - I have no idea whether anyone ever worked on the 2d-part of step in open source... Cheers, Lars.