New take on Conversation in Video Games

Started by
32 comments, last by GalvonicBond 8 years, 7 months ago

This takes you away from a dialogue system per say but I've been playing with the idea of an icon based language. Many RTS use icons for most everything anyways so much of the character interaction with NPCs could use emotes to express feelings and icons to indicate things, actions, places, people, etc.

I thought a basic short term, long term (based on repetition of short term) and recall memory (recalled by other characters) could allow AI to apply emotes to events and actions of other characters or the player character. The way to weed out minor events is influence. A character's influence is based on their connection to other characters earned through socialization (with individuals), success in combat and interaction with scripted events. Connecting to characters with higher influence works like a shortcut to earn fast influence with those already connected to the higher influence character. Scripted events can help connect the player character to many NPCs all at once as they watch or hear about the event. The key to make this viable is creating a system for characters to forget. The recall memory would allow a character to be reminded of some events returning it to short term but with all the old data of the amount of times it was communicated about. Keeping short term memory reduced to a certain size with influence of events and long term is a limited collection of repeated short term events could keep every character's memory to a manageable size. Hope this makes sense.

Advertisement
[edit] As with the "Grrenlight" thread, my apologies that I took so long to respond to this. ^^; [/edit]

... it still does not change the fact the dialogue is the tool to provide info/action.

Thinking more about your arguments, I somewhat agree, but it also seems to me like a somewhat reductive perspective for the writer to take: thinking about a dialogue in terms of what it does for the plot seems somewhat counter to letting the characters act freely. But then this may be a difference in approach between you and I, much like the difference between those who "write by the seat of their pants" and those who carefully plan out every element before the first word is written.

I think that the argument about picking up a book is misleading, however. By the same argument, one can open a book, note that it can be deconstructed into tropes, and thus conclude that it's a good idea to write a book by constructing it out of tropes. However, while drawing out the tropes of a book may serve well for analysis, and while one can learn from tropes, I think that constructing a book out of tropes will tend to produce rather limited characters and plots. "Troping" is a tool of analysis, not of writing. But note that this is in counter the the argument itself, not the conclusion, which I addressed above.

... real life chatter (which in 90% cases has no purpose biggrin.png) ...

There's another point in which I disagree with you. tongue.png

However, more relevantly, in a game it may serve the purpose of fleshing out the world, and giving an additional layer of verisimilitude to the NPCs. Characters that do nothing but point to "that dragon over there that killed me ma!" are quite clearly just signboards with faces; having them shout at the baker about what he said last week might make them feel more human.

Another thought is that not all games need to have strong narrative direction; if games can work with a quest-sandbox, why not a narrative one? Just as some players love to read every scrap of writing presented, no matter how irrelevant to the plot at hand, some players may enjoy taking time to just sit and chat with a character.

The thing is, games are far, far more primitive we are willing to admit (both as designers and players).

I believe that this is in part a lack in the writing found in this medium, not purely a limitation of the medium itself. We could be making games that are far more varied, but we have a tendency to fall back on "kill things and take their stuff". (Not that there aren't games that go further than that; I just feel that they could be more common.)

I do think that more complex interactions can be difficult to model--but then this very topic is about an attempt to model one of those things. Should we not be attempting to support it in order to broaden the medium?

MWAHAHAHAHAHAHA!!!

My Twitter Account: @EbornIan

designing a system like this would not only take a HUGE amount of time. but consider how difficult it would be to code this in, and make the conversations unique and memorable

Your system reminds me a lot of Fallout 1 & 2, OP. Disposition, speech skill, intelligence stat, and topic all matter in the Classic Fallout dialogue.

This topic is closed to new replies.

Advertisement