Quote:

Original post by Omegavolt This would all be a part of the way a dynamic world would work, however. If these NPCs were all a part of a city, then most of them had jobs right? Now theres vacancies for these jobs. That could affect the player. Plus, I would think that other NPCs who had relationships with the dead NPCs would find out about the deaths, and want to take some sort of action. If this action is something the player could do, it could turn into a new quest, or line of quests really, for the player.

Still, I think you might be overlooking the amount of stuff that needs to happen in between the AI's "doing their thing" and it turning into a quest that is presented to the player in a way he can understand. Voiceover is right out; you can't record dialogue for dynamically generated plot. You'd probably need some sort of dialogue-writing AI, similar to the conversation-having AI's out there, which is a huge effort in and of itself. (And don't even think about localization...) There are other ways to do it without creating a dialogue synthesizer, like the way the Sims does it... having the characters communicate in abstract picture bubbles, and directly showing the player their emotions and desires in abstract thought bubble pictures. I thought this was really clever (and the fake language was a good way to avoid localization), but I don't think it would work to present a standard RPG-style "quest" to the player.

I guess you could do a fill-in-the-blanks style dialogue, but I've never seen it done well. Something like:

Guard: "Hello. It appears that <Bob> was <murdered>. I heard that <Sam> was <angry>. I heard that <Susie> was <jealous>. Find the <killer> and I will give you <100 gold>."

But if you limit yourself to fill-in-the-blank style dialogues like that with pre-decided ideas for what sort of quests you are going to present, like the guards asking you to find a murderer for money, then I think the game would be vastly more interesting if you just pre-scripted a series of partially random, repeatable "Find the murderer for money" quests. All the development trying to get the AI to decide to murder on its own is a waste of time since the only difference the player will see between your game and the other is that yours has much worse dialogue.

Quote:

Original post by Trapper Zoid This is also part of why I don't feel the simulation approach is the best way to achieve interactive stories: firstly because events will happen where the player cannot experience them, and secondly because they have a tendency to spiral out of control; if the backstabbed guy happens to be integral to a subplot, then the whole thing just got wrecked by no fault of the player. ...

I agree completely. I think the best way to get dynamic plot would be to not think of the controller AI as a storytelling AI, but instead as a Dungeon Master AI, for anyone familiar with p&p roleplaying. Thinking of it as storytelling or authoring doesn't mesh that well with what should actually be happening, whereas Dungeon Mastering is almost exactly what people are discussing: you need to dynamically adapt your characters and plot to provide an interesting story to a person or persons who have the freedom to do whatever they want in your world. The Dungeon Master AI in a dynamic plot simulation would have the same duties as a DM in real life would; it would try to read into the player's backstory and actions to determine what he likes and present plots that will hook him in, it will try and subtly railroad characters back into the borders of the story world if they start doing things that are taking them into unfinished territory, and it would gloss over parts of the ongoing action that are uninteresting and focus on the parts that are exciting and directly involve the player.

Quote:

Original post by makeshiftwings I agree completely. I think the best way to get dynamic plot would be to not think of the controller AI as a storytelling AI, but instead as a Dungeon Master AI, for anyone familiar with p&p roleplaying. Thinking of it as storytelling or authoring doesn't mesh that well with what should actually be happening, whereas Dungeon Mastering is almost exactly what people are discussing: you need to dynamically adapt your characters and plot to provide an interesting story to a person or persons who have the freedom to do whatever they want in your world. The Dungeon Master AI in a dynamic plot simulation would have the same duties as a DM in real life would; it would try to read into the player's backstory and actions to determine what he likes and present plots that will hook him in, it will try and subtly railroad characters back into the borders of the story world if they start doing things that are taking them into unfinished territory, and it would gloss over parts of the ongoing action that are uninteresting and focus on the parts that are exciting and directly involve the player.

That's it exactly! For some reason the Dungeon Master analogy completely slipped my mind when making that last post [grin].

I still consider this to be storytelling, but of a different kind from what might spring to mind when considering a story. It's less about high drama in the vein of Shakespeare and more like a parent making up a story for their child with the child as the hero. Or a good Dungeon Master trying to come up with scenarios that the other players want.

I think you need to talk with Will Wright about this game - his new game Spore combines different style games into one - and after watching the E3 videos that are all over the place, it has been done well.

This could relate to your problem/challenge in a couple ways: You would have to combine different genres into your game if you wanted o be able to:

A)fight - standard RPG - Obliviion is a good game for this genre - copy it
B)farm - don't know about this - I wouldn't want to farm
C)rule a kingdom (which has elements of economy, sim style interface for structure building and maintenance, army and personel management, etc..)
D)smith or small business owner - another sims style game
E)adventurer/explorer - this could be implemented with an oblivion style also

Then comes the scripting - definetly have to be dynamic although there is nothing out there that quite does what you need - a true dynamic dialouge generator based on events which are dynamically generated. Then managing the player economy is multiplied in complexity if businesses and kingdom taxes come into play - if you are unhappy do you gather a group of equally disatisfied peasants and rally in front of city hall until you are arrested or things are changed; how about a revolution to overthrow the kingdom's ruler's?

This would be such an awesome game if we have around 10 more years of technology under our belts - One idea is an event repository the game could call on (using XML with webservices or something or the sort) so all new types of dialouge and coded event modules can be shared among all the computers that are 'connected'. That way when one person generates script for a random event, it could be shared and brought into the game via the web somehow - obvious challenges arise there too. And AI, NPC's, etc..

That is my 2 cents on the issue.

interesting thread-jacking...

Can a webservice be used to concentrate player-created content and distribute whatever has been proof-checked by a dedicated beta service, to generate a pseudo-livefeed, and therefore a pseudo-interactive game?

As a single player RPG.. honestly no fun what be the bs apprentice and make spears until the npc's heros save the world, the game would just not that have enough content generation to keep a person entertained, there would be a finite life

However if you could take secondlife(and other type games) player generation and a rule based online rpg (aka ultima, wow, everquest) you would have a great game ^.^

thats just my view, and one id definatly want to see started and play, i want to be a passive role, I dont want to go and kill monsters

( I think it would be hard, but not impossible )

Ive seen a couple of places say they where making such a game.............
there either not around, or not updating, or just not showing any progress even with there updates

(About spore, i watched the demo, the actual gameplay wonderful, but the user to user intereaction looks like it was non exisitent, you could not affect another player and there world directly.. no fun)

[Edited by - Dinner on July 6, 2006 6:13:58 AM]

I don't necessarily mean the game will be like spore - I just used it as an example to show that there can be more than one type of game within a game: sim city, the sims, a fighting style game, etc...

Somehow this game would have to accomplish the same thing, but in a much more dynamic way. If there was to be a truly 'do whatever' type of game, you would need different interfaces and game elements to satisfy all the different 'jobs'. And if programmers could easily add different mods to incorporate a different type of job, then it could theoretically have infinite types of jobs (farmer, ruler, figher, wizard, smithy, hunter, inventor, etc...)

Store the codebase and modules in a repository and update all the 'connected' games with new content and let the users play and programmers keep adding content.

Just an idea...

Quote:

Original post by ryn0virus If there was to be a truly 'do whatever' type of game, you would need different interfaces...

Not necessarily. In theory, I could 'do whatever' in real life, with only one 'interface.' If I want to kill an orc, I <pick up> a <blade> and <swing> it at the <orc>. If I want to make a blade, I <put> a <piece of iron> in a <fire>, then <pick up> a <hammer> and <swing> it at the <hot piece of iron> repeatedly to acheive the desired shape. If I want to grow crops, I <pick up> a <shovel>, <dig a hole>, then <put> a <seed> in the <hole>, and repeat.

You could define macros for these actions so it is only tedious the first time.

Quote:

Original post by josh_w314 Quote: Original post by ryn0virus If there was to be a truly 'do whatever' type of game, you would need different interfaces... Not necessarily. In theory, I could 'do whatever' in real life, with only one 'interface.' If I want to kill an orc, I <pick up> a <blade> and <swing> it at the <orc>. If I want to make a blade, I <put> a <piece of iron> in a <fire>, then <pick up> a <hammer> and <swing> it at the <hot piece of iron> repeatedly to acheive the desired shape. If I want to grow crops, I <pick up> a <shovel>, <dig a hole>, then <put> a <seed> in the <hole>, and repeat. You could define macros for these actions so it is only tedious the first time.

But I think you'd need different interfaces if you wanted each sub-game to be fun. For example, if you did what you suggested, and made the farming part a real-time FPS view that mimics real life, it would be the most intensely boring game ever made. Walking slowly through the field for eight hours real time clicking weeds just so I can eventually, after six months real time, pull in a meagre harvest and get a few gold pieces that I could have found after five minutes of adventuring? A truly "fun" farm sim would, in my opinion, need to have more of an RTS/SimCity UI, with a sped-up timeframe to match, and control over assets and employees from simple UI windows. On the flipside, a game about one-on-one melee combat would not do well in this UI.

I think that providing the interfaces would be a minor problem compared with making every possible occupation fun to play.

Most people do not consider manual labour to be fun in real life and there is very little you could do to make it so in a game. Sim games are only enjoyable because the player has some type of managerial role.

Playing the role of a storekeeper might be fun for players that enjoy social interaction if customers came by on a regular basis, but otherwise it would consist of standing around in one place day after day.