Are there any guidelines or best practices for the design of multi-monitor systems?

By multi-monitor I mean systems which arrive out of the box with several physical monitors, each with predefined functionality. These are usually industrial systems running their own software.

This type of systems raises interesting questions, such as – do we duplicate important stuff on all the screens? Do we gray out all screens to display a modal dialog on one of them? Do we create a hierarchy of screen “importance” or are they equal? Etc.

There is some research on this from the ergonomics perspective but I’m looking for more of a UX-oriented work.