Im still looking through the link you posted but I still don’t see how a computer based story/quest generator could compete with a handcrafted game world. The best you can do is a random joining of pre-made elements because Computer don’t have imagination

is there any game on that site i can download, i read the link but any programmer can tell you theres quite a gap between a vauge description of how a programs going to work and actually making it

The problem is that being RANDOM is COHERENT. I read through it, and its a nice idea. Take a set of parameters, mix and match, and out comes a nice story arch. However, these story archs can get very reptitive very quickly. Even if you are being asked to do different things, like Save old guy x, or Go scout place Y, random quests just dont have that hand-crafted satisfaction to them.

A great example would be when I was designing the story to the game I was working on, I wanted some good names, so I went to a random name generator. These generators seemed to be great. They had defined syllabelles that would match to make hundreds of names. However, after going through probably hundreds of names, I found only five that I found any use out of. In reality, I only kept one as a city name.

The point is to go back to the orginal topic, An rpg where you can actually role play in. Even when random generation is used, it still feels generic.

Quote:Original post by Vanquish
Wrong. I'll find that thread about the quest generator. It can and is being done. Sorry to burst your bubble.

It is possible, just hard to do right. The term 'random' is not really correct since such a system is based on static templates (of situations and behaviors). To create quests of any complexity, the system will require adjustment logic and even logic to decide if the particular quest pattern is relevant/redundant ro the current state of the game world. Each template could have subcomponents and nested levels of templates can be used to fill in details, with adjustment and fitting locally (and some flavor scrambling etc..).

Ive been thinking of such a system for my long term project. One important abiliity is to have the 'quests' not interfere with each other (but some clean themselves up after solved or they timeout and dissapear, allowing reuse of
world space). Another aspect is the generate on the fly (similar to 'lazy evaluation') where depending on what parts of the world a player touches, the local quests get triggered otherwise they dont take up world data resources waiting for a player who may never show up... Trigger seeds on the world contain the directives for relevant quests (some triggers can insert/spread secondary quest triggers or chain them).

Example:
Player wanders thru world, at a location a trigger object activates and places a clue prop near the player. If the player sees/picks up the clue, IT activates and procedes to set up the quest -- placing NPCs (with appropriate behavior templates), props, scenery features(structures, tunnels etc..). Placement is done nearby in a spot where no other quest is active (and possibly on terrain selected to match the theme). More clues might be placed around the quest
location to guide the player to it. Eventually the player will get close enough and wake up the NPCs and interact. The quest can then trigger a chain of quest phases (setting them up similarly to the fisrt phase).
The Quest object has its own script that monitors the situation and can be able to clean up after too much time goes by or test world conditions to terminate itself.


Notice that the 'trigger' seeds must be preplaced (or a area of terrain will have a 'random encounter' list of appropriate quests (these could be nested sets too) and matched to the world regions/theme areas etc...


Alot of filter/adjustment logic may be needed to intelligently fit quests into the existing world situation and theme (hardly 'random').

An over-controller could distribute new seeds -- following ques in the main storyline (I suppose they could have yet more script logic to adjust to the area they were placed onto, ex- Enemy Invasion is translated into raids by local enemies (possibly orcs) and appropriate events and sub triggers are
placed/distributed. And when the root object is terminated all the corresponding elements can be terminated/cleaned up (possibly leaving appropriate debris to show what had transpired etc..)

Templates would have parameters to allow flavor/mode selection from the caller and also the dynamic subcomponents which can be changeable (calling down to other templates).

Script Functions are needed to check existance/data of other quest objects and world situation (the main storyline) and to place (and program) NPCs, place props, deform scenery (like clear an area to allow placement of specific quest (mini-plot) props etc...). Other functions manipulate/place/unplace trigger objects.


A wide range of quest complexity could be supported from, 'Whats under that weird rock' to a grand quest to 'Find/Kill the Dragon, Get the Princess, Win a Kingdom'. Many small loacalized quests/encounters could fill in all that empty space in a world (and the seed mechanism minimizes overhead).


A simplified form of this scheme could be used to improve the fixed
encounters that many games seem to rely on (especially the MMORPGs which cant have enough online GMs but duplicate servers where the work put into scripts can pay off).

The scheme could be expanded to control entire subplots in the world/storyline (the method is the same, just the scale is larger).


Ive heard that some of the newer MMORPGs (like StarWars) is starting to use 'holodeck' like areas (reuseable) where they can place new game situations
(usually precanned chunks) and later remove them. This is basicly the same idea, except at a clunkier (more manual) and coarser granularity.

Quote:Original post by Kaze
Im still looking through the link you posted but I still don’t see how a computer based story/quest generator could compete with a handcrafted game world. The best you can do is a random joining of pre-made elements because Computer don’t have imagination

It cant, but because hand crafted stuff is so expensive (at least until the game mechanism starts letting players create it) that the worlds are quite limited and the player has this feeling of being led around by the nose and frustrated at all those doors that never open and precanned situations which allow little creative action.

By doing hierarchical templates, you could vary what is generated quite alot
and with built in logic you could make 'on-the-fly' creations fit/adapt better
than the monolithic static scrambled building blocks that you are thinking of.

I don't think there's any argument that the best human plot generators are better than the best automatic plot generators. On the other hand, there are plenty of examples out there where the manually generated plot could be improved upon by just about anything. Yes, mix-and-match with modular plot components can only produce what you put in in the first place, but then again a completely hand-crafted plot can only produce what you put in in the first place, and gets a lot more predictable when you replay it...

There's a quite successful line of books here in the UK, called Mills & Boon Romances, which follow a very rigid template, and could probably be churned out by a computer program almost as easily, but are eagerly devoured by grannies (of all ages) who want something slightly titillating without being shocking, and don't want to have to make the effort of following a complex plot.

Or there's numerous TV soaps where story-lines are sufficiently modular that if one soap does something the others will all have the same storyline within a few months (allegedly - I only watch one soap myself, so can't really compare)