Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[lordnincompoop]

Are you going to have them do anything else? Friend/foe recognition? Stuff like that?

[rarborman]

Okay, so I've fixed a few more issues relating to the unlocking process,
 and I've started adding in switches.

I've added support for static items (permanent fixtures) and I've linked a lever up with a door.
The character is identifying the lever as a possible mechanism to open the door,
 but the process of testing the lever and door is still floating around in my head.

I think I'll have them try pulling a lever, and if there are no immediate results they will recheck nearby rooms for changes.

I'm not really sure how I should have them handle this.

Just have them go about like normal, because do you check every room in your house for changes when your lightswitch doesnt work?

If you can find a way to have them link changes to the lever by nondirect means then your AI is more realistic.

[Outcast Orange]

That sounds like a good idea.

I've noticed a lot of the reason it takes so long to accomplish anything
 with this project is there are too many classes that are too similar.

I'm combining a of the classes, and then working on a new system that will create
hypothesis/assumptions about things, then go about testing them.

[Armok]

*credit?*

[SolarShado]

Is sound implemented in any way? Locks generally make a sound when they're (un/)locked...

[Outcast Orange]

Armok gets credit for the idea of combining all of the classes,
 and he's been giving me a lot of suggestions over the hypothesis process.

If anyone else wants credit for something they once said or suggested, please just complain here.

Anyways, the sound thing is a great idea!

I've been wondering which sounds would be audible between rooms,
 and what sort of responses an AI would have to hearing them.

I'll confess that I haven't really worked on this at all this weekend,
 the new Minecraft update has been distracting me pretty heavily.
 

[Armok]

Sounds should probably wait until AI reacts to seeing other AI do things.

[SolarShado]

Sounds should probably wait until AI reacts to seeing other AI do things.

Actually... you might could use a similar system for both...

[Armok]

Yes, that is the reason. No point in cluttering the project with sound until it's gotten to the point where it might fill a purpose.

[Korbac]

Fantastic work, Outcast Orange. 

Also, for the sound thing - an AI would be more likely to shelve their current goal and investigate a noise the louder it is. An AI would be more likely to investigate a noise if they have no goal / no other option. An AI would be extremely likely to investigate a noise if it was concurrent to an unknown lever / switch they just used.

NINJAEDITED: Fair point, Armok.

Again, great work Sir Tangerine. (Can I call you that? I think it sounds pretty awesome.)

[Outcast Orange]

Sir Clementine please.
Not only does it sound wonderfully absurd, but the fruit tastes better too.

Thanks for the suggestions.

I feel I've been over-thinking a lot of these things,
 and my common sense has abandoned me.

[Korbac]

Sir Clementine, common sense is easily forgotten whence you are in the coding zone! 

[SolarShado]

Sir Clementine, common sense is easily forgotten whence you are in the coding zone! 

Amen

[Outcast Orange]

Combining the classes has taken a lot longer than I thought it would,
 especially since I've been really busy lately.
I've finally done it, and all of the obvious bugs or related issues have been rooted out,
 though I'm sure there are a lot more waiting dormant.

I'm still not sure how exactly to proceed,
 but here are the overall goals:

- AI stores a list of possible actions that have simple effects and prerequisites

I already have a system like this, but the structure is very rigid and wonky,
 which makes further expansion feel akin to guiding a spaghetti noodle into a sink drain.
I think gutting it and taking a cleaner route shouldn't be very hard at this point.

- AI Knowledge and Testing Hypotheses

I'm really not sure how to approach this one.
They hold tons and tons of information, tied in with abstract clones of everything they encounter.
There are certain aspects of each item unknown, and the AI should yearn to fill in these unknowns.

Hypotheses will have to be created by finding some process that would yield distinct results depending on the circumstances.
This part should be a bit more complicated, but still manageable.


I'll go ahead and start restructuring actions, since it shouldn't take more than a few hours.

[MrWiggles]

Update?