I've been doing a lot of 3D modeling lately, using Google's SketchUp (formerly from @Last software). I love the software so much that I put down my credit card and purchased a full version after enjoying the liberal 8 hour preview (8 hours of actual runtime in the software, which lasted me for quite a while).

The thing I find most impressive about this software is the gesteral nature of it. The interface does a fabulous job of "guessing" what you meant. I remember using AutoCad with a large tablet input device in the 90's - they had separate commands for snap-to-endpoint, snap-to-midpoint, etc... in SketchUp when you get close to the middle - snap... close to the end - snap. It's really quite nice.

One of my coworkers (Dave Peck) said that a bunch of the work in this space (3D gesterual interfaces) started croping up in the late 90's. I'm still buffering the video, but he pointed me to Robert Zeleznik's SIGGRAPH '96 video, as well other papers from Brown University including An Interface for Sketching 3D Curves (ACM 1999) and Drawing Strokes Directly On 3D Models (SIGGRAPH 2002).

I find the space facinating, especially considering the rather frustrating way this type of interface appears in Word (autocorrect seems to frustrate me consistently).