The best way I can think of to keep things from getting boring is to add in a little competition. If you''re being briefed by a gov''t agent, there''s bound to be leaks in his department. So an opposing gov''t gets wind of what''s going on. Now maybe you have two people looking for the weponds report in Bob''s desk or something. If you get to it first, your competition tries to track you down and snatch it from you. Or the other way around. It might also be important to not kill your competition no matter how much you''d like to. Also if it turns out your competition is Agent X and Agent X repeatedly shows up, maybe eventually you''ll get a mission to kill Agent X. Or keep the same from happening to you as your profile goes up.

Add a few NPC with player like abilities. If they''re trying to track you down it''d be difficult to let them interogate other NPCs, but they could certainly follow your financial records or something. If it''s a space game you gotta know everything''s on computer, including what you eat for breakfast.

::snips discussion for future reference::

I think another important thing would be to make sure you vary the missions as much as possible. In a good random mission generator you might be able to have billions of combinations but there only maybe 1000 (or less) of them that seem actually diffrent to the player. The key is to get the 1000 diffrent ones to happen and space out the variation ones between them.

One idea I had is testing how much a mission varys from the last say 10 missions. The last 10 missions should be stored from game to game (even if they are playing on multiple saves). Then the computer can compare a randomly proposed mission with each of the 10 and every time i gets a rating 1-100 about how similar it is. 1 is exactly the same, 100 is completely different. Then you add up the values of the originaltiy to get the total originality vs the last 10 missions.

You can code it so that if the originality is below a certain level not to accept the mission, if it is very original to accept it and if it somewhere in between to try generating another and take the better of the two.

The process for deciding how different the proposed new mission is from the other ones can be checking each "waypoints" elements. For example if they are both attack missions it loses a lot of originality. If one mission is a attack and one is a protect it could gain originality, but not as much as comparing a attack to a collect. The amount of originality gained or lost based on the goals would be hard coded and depend alot on the game. If your game was fighting and roleplaying you would subtract originality for two fighting or RP missions in a comparison. This would not be the case in an all fighting game because there is no other alternative.

The originality comparison code would also compare involved items, towns, magic, monsters etc. for similarities (rescue princess vs. defend princess would score low originality.

What do you guys think of this method? Is there anything I am overlooking?