Just to cut back to a previous topic for a moment: 3D engines, while not essential, can definitely be employed to make a game less linear. Real-time 3D allows the designer to rely less on pre-rendered cut-scenes, and instead design cut-scenes that change dynamically based on the chosen path. This puts less of a strain on the story to conform to whatever is set in stone in the cut-scenes. For instance, Post Mortem had three different pre-rendered versions for all the major events in the game. Had those cut-scenes been rendered in real-time, it would have been easier to create more different scenarios with more different outcomes. Similarly, conveying the story through in-game text (signs, letters, notes), scripted sequences, dialog and NPC behaviors will make the script easier to adapt in more subtle ways.
Besides the game's overall narrative structure, interactive dialogs are another area where a game can be made to feel less linear. Innovation in this field is much desired. Instead of the usual completely pre-defined conversations, it's possible to add AI to a character that will respond to the choices of the player. Bateman's FreeSpeak engine has a so-called affinity system, which does exactly that. The affinity system uses a set of variables that can be affected by the choices of the player. For example, by saying all the wrong things a character's fear variable might increase. For every threatening sentence the player selects, the fear value would be increased by a certain amount. If the variable exceeds a certain number, say five, the character becomes uncooperative and the player will be forced to take another path through the narrative. If the fear variable stays below a certain number, for example two, the character might be more helpful and give the player an object that would have otherwise taken much longer to find. In other words, many things can subtly affect the variable, but by writing different scopes for the possible outcomes the conversation will not get trapped in a dead-end. Dialogs should be more part of the gameplay itself, and offer interesting choices that have noticeable effects. Usually I just exhaust all the dialog options on the screen without putting any thought into it, as none of my choices really matter.
Another point I should make though, is that not only should games actually be less linear, they should also feel less linear. Quality writing can to some extent mask the shortcomings of linearity. The Longest Journey is both an example of how to do that, and how not to do that. The first few chapters of the game seem to flow pretty naturally, as we play out parts of April Ryan's daily life. Eventually we're given the choice to have April spend the evening in one of two different ways. This choice does not affect the story, but it does give the illusion of being in control. In the middle part of the game, April Ryan has to recover four different artifacts. This is a rather trite story device, as found in many RPGs (e.g. collect the five magic stones, the three pendulums, the four swords, whatever). Once this quest is introduced, the dramatic tension that was built up to that point subsides. The real story of The Longest Journey is pretty much put on hold, in favor of this elaborate subquest. The player turns into an errand boy, going through various awkward and at times contrived obstacles to acquire these artifacts in four long chapters. The player already senses that what he's doing is just designed to further him on the on-track narrative. When April finally gets to use the four artifacts, the cutscene plays out exactly how the player expected. Not only is this payoff unsatisfying, the way the story was set up made players feel like they were constrained in what they could do.
Getting the map pieces in Monkey
Island II felt less frustrating.
Finally, puzzle design plays a major role as well. Bad puzzles can make a game feel more linear than it already is, e.g. by offering only one solution or by being too present as an artificial obstacle. In a way, puzzles are the doors that open to different parts of the story. A lot of adventure games don't seem to be overly concerned in providing a varied mix of different types of puzzles. Ironically, it is often Doom that is mocked for its lock-and-key puzzles, but graphic adventures are not much better off, except instead of key cards there are rubber duckies, plucks of cat hair, cogwheels or magical stones.
Developing good puzzles is arguably the hardest aspect of adventure game design. It's tempting for any designer to just cut down on the puzzles. Tim Schafer removed a lot of them from the final version of Full Throttle, presumably out of frustration with having to add puzzles just to increase the playing time. However, this resulted in many players completing Full Throttle in one sitting. On the other hand, games that have too many puzzles will often have players reaching for a walkthrough, which makes the gameplay self-defeating. I believe games should monitor the player's progress and provide more hints when it's apparent that he or she is stuck. This already exists in action games. For instance, Max Payne adjusted the game's difficulty according to the skill shown by the player. Secondly, the “take X and use on Y” style inventory puzzle is very hard to do right and should probably be kept to a minimum. They're an open invitation for using the try-everything-on-everything strategy. It seems the latest trend is to use only inventory puzzles, which makes adventure games all the more predictable. I felt that Runaway, despite its success, might have been guilty of this.
In Half-Life 2 players will be able to
use the physical properties of
objects to their advantage.
To summarize, I believe that through the use of non-linear narrative, and clever stories and puzzles that are designed to not feel constraining, adventure games can be made a lot more free flowing and fun to play.Continued on the next page...