A lot of triple-A games these days are trying to incorporate personal choice, with excellent examples like Mass Effect and Deus Ex...and not-so-good ones like Fable. But meanwhile in Type-Moon visual novels, you get games where one single choice puts you on a totally different plot-rail. Even if you ignore the abundance of hyperviolent 'bad end' paths, you still have totally separate plotlines with different villains and themes. Where Mass Effect was always headed toward an ultimate showdown with the Reapers, Fate/Stay Night can make your greatest enemy Gilgamesh, Archer, or the Holy Grail itself all depending on what you do.

D'you think more traditional gaming can take after visual novels in this regard? Where rather than amazing writing only slightly changed by your choices like Bioware gives us, or being so impersonal that you can kill everybody or nobody (Way of the Samurai), you have grand, artistic, emotional paths waiting at the end of each choice, all with full production values and character development. Is that asking too much? And more importantly, would you be interested? =P

Most RPGs have linear story lines. I recall several articles I read mentioning alternative paths are designed for replayability (or replay value). And if you look at games like World of Warcraft, its character race design is very typical. It tells the player that even after you get bored with your character, you can still create a new one as another race to get a new game.

I also recall reading from somewhere that games should not be like movies. The view argues that a game should be more of providing an interactive decision making environment to the player, and less of just telling the story. In that sense, alternative paths also serve the purpose of having more varieties, which makes the player feel like he's in control, rather than having to do something he's told to.

Nonetheless, I figure nobody hated Call of Duty because of its linear story setup. So if it's done right, linear story can also be great.

This was one of the 7 goals that I'Ve set myself when I began designing my current project 7 years ago (yes, that's two 7s there, one more and I go for the jackpot).

Because of the production constraints it entails (Khaiy pointed that out very nicely) I had to turn it into: "Few, but meaningful decisions".

Let's break this down:
"Few": The more decisions the player can take, the less significant the difference it can make in the game both factually and by contract.
With a given timeline and budget where each "exception" consumes a large portion of your budget, you need to carefully plan ahead what specific the decisions you want the player to make.
By contrast, fewer choices will have the player stop and think. Imagine a rpg with 99 levels but you only get 1 skill point every 10 levels.

"Meaningful": It's not just giving them choices but making sure they care, and that there is an endgame somewhere.
Caring can be achieved by making drastic decisions on think they know (the village they started in, the characters they control)
The endgame is more tricky. You, as a designer, must be able to evaluate what will make the most fun content. It can only trigger at the very end of the game based on what you did along the way, but you must make sure that the connection between the player decisions and the outcome is made clear: You get this final boss instead of this other final boss because of what you did at that specific event.

"Decisions": This is often overlooked, but a decisions must be something tangible to the player. Don't make a room with a door on each side which appear to be the same and have a trigger in place that considers the first door opened to be a choice. It really isn't.
You must hand out information, even if it is not always accurate. You must hint at the possible outcome or at least at the ramifications.
You want the player to know that siding with the evil empire will have at least some chances of inflicting terror upon the neutral population through invasion if you plan to have this decision turn into the destruction of an entire city.

If you can achieve this, people will often say that they had liberty of choice even if there were only rare occasions to express themselves. They felt they had the data they needed to think it through and express themselves through a decision. Compare this with open-world games that are not sandbox, where you can technically make decisions all the time and you'll soon realize that it is impossible for them to capture the same level of choice.

Its really like politics: A player given a chance to customize every single tiny bit of detail of the endgame scene won't be surprised when he gets there because he's been doing all of the micro choices along the way. If they did vote once for a political party, they'll be surprised by everything that happens afterwards they didn't necessarily think of ;)

I'm less pessimistic. Mass Effect players were angry that their choices in ME3 were pretty much worthless. They (and I) wanted those choices to have an actual effect, as it was in ME2 when many main characters were capable of permanent death. So the demand is definitely present.

well, ME3 was the last game, so permadeath is just pointless... if you think of it, the end of the game was the same as the end of ME1, you chose the ending in a dialog.

they only have failed on the cutscene, that was the same for every ending. they failed in delivering the story, not in the decision making. as a matter of fact, at the ending, there are some options you can choose only if you used only renegade/paragorn options on the entire game (with that character).

the choices of the player must feel different, even if they are almost the same. ending of ME1 was an example. even if there's no clue that the council survival would matter, i just couldn't let them die. the endings feel different.

sorry, my english is bad, in this part I was Talking about the Mass Effect 1 ending, where your choice is to let the council die or order the aliance to save the council, I agree with you on the final of Mass Effect 2. that wass a BIG decision let to the player, their choices since the begining of the game matter, and much.

ME2 is the best example here of a game where the player choices really matter. but in ME 1, if you let the council die, a new one is there in ME2. if wrex die, their brother do exact the same actions he would have done (the dialog is a little different, but the actions are the same...) the only hard choice in ME1 is who dies on virmine. this is the only choice that really change something on the future games.