The efficient creation and manipulation of complex 3D scenes using 2D devices is an elaborate task that requires skilled people and many hours of training. To simplify the task of aligning objects, various snapping and constraint mechanisms have proven useful for computer-aided design (CAD) and related modeling applications. The authors of this paper propose the usage of constraints to limit the degree of freedom, according to physics (gravity) or human conventions, in order to simplify intuitive positioning and aligning of objects in 3D by means of 2D input devices. Special attention is paid to grouped objects that can be handled in an easy and intuitive way.

In general, the proposed techniques are not completely new; many aspects and variants of them can be found in planning and editing tools available for interior furnishing planning.

The authors focus on the creation of scenes using objects from a library, and introduce different types of object constraints. They have implemented an editing environment called multi-user intuitive virtual environment (MIVE) to test and compare the proposed editing methods. All examples described are related to interior furnishing planning, which is perfectly suited to this kind of constraint-based editing.

This paper focuses on the comparison of unconstraint-, partially constraint-, and fully constraint-based editing. For this comparison, the results (time, accuracy, and so on) of 15 computer science students with different skills have been gathered and evaluated. Since the study is not conclusive in all aspects, the evaluation of the results indicates that constraint-based editing has clear advantages over completely unconstraint systems.

While it is true that constraint-based editing makes 3D editing simpler, the problem is specifying object constraints, which is a complex task that requires application specific knowledge about the object and cannot be completely automated.

The authors provide a comprehensive overview of previous work in this field of research. They describe their proposed constraint-based techniques in detail, and discuss their benefits with respect to the evaluation of results from user testing.