Jesper Juul's Game Diagram could be a good way to start a better definition of nongames. If you know his http://www.jesperjuul.net/text/gameplayerworld/, you probably are familiar with the six features proposed by him to define games: (1) Rules, (2) Variable and quantifiable outcome, (3) Valorization of outcomes, (4) Player effort, (5) Player attached to outcome and (6) Negotiable consequences.
According to Jesper, in order to qualify as a 'game', an activity must cover all six features. If not, it's either a borderline case (and some of them we still call games, like gambling nad games of pure chance) or not a game at all (which is not equal to nongames, in our case). So how would nongames qualify?
Nongames, yes, have fixed rules. In fact, as a general rule, computer games have fixed rules (even if this rule allows the system to change them, combine them diferently and/or make new ones).
My only problem with that feature is that all computer-based systems have rules. Photoshop is (arguably) not a game, but it has rules (for instance, I can't use more than 256 colours if the image is in 'indexed mode'). Because of that elasticity, I am sure that nongames have rules, I am just not sure if they have it in the sense that we expect from a game.
Elektroplancton, for instance, uses conventions from musical software. The important thing to notice is that those conventions are now translated into a videogame environment.