Here is a simple technique for camera sequencer in Softimage.
I use one model with multiple cameras inside it.
One is the camera target (Camera_final)
Other cameras are sources for the sequencer.
I use a model for isolate the mixer of the sequencer.
The technic is to add a pose constraint and some expressions from each cameras source to the target camera and a movie clip is stored. this clip is deposed on the sequencer (the mixer).
This allows you to change length and position of all shots