Print

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

[salmonjockey]

Well, I just started on one, but won't ever in my life have 72 consecutive hours to complete it. Does it count if I add them up?

Only if they're all in February.

I'd say it would be silly to add up individual hours. But adding up individual days, that sounds fair, at least to me.

[Bricks]

I wonder if I'd have the initiative to do this.  Assuming exams scheduling isn't insane I might just give it a shot.

Reading this reminded me of this article I recently stumbled upon on roguebasin:
http://roguebasin.roguelikedevelopment.org/index.php?title=The_Incredible_Power_of_Dijkstra_Maps

It's written by the creator of Brogue, and you can see the power of it in the simple but effective AI (damned monkeys!) and the auto-explore feature.  Assuming you want your actors to both follow directions and have their own set of needs, it could be a very effective alternative to waypoints.

[Biag]

I wonder if I'd have the initiative to do this.  Assuming exams scheduling isn't insane I might just give it a shot.

Reading this reminded me of this article I recently stumbled upon on roguebasin:
http://roguebasin.roguelikedevelopment.org/index.php?title=The_Incredible_Power_of_Dijkstra_Maps

It's written by the creator of Brogue, and you can see the power of it in the simple but effective AI (damned monkeys!) and the auto-explore feature.  Assuming you want your actors to both follow directions and have their own set of needs, it could be a very effective alternative to waypoints.

Oh wow, that's extremely helpful. And brilliant. Roguebasin has some fantastic articles.

[Virex]

That last part of the article makes no sense though. You can't combine desire maps because desires aren't additive. If you've got to go to the toilet and make yourself some food, you're going to prioritize one over the other. Instead, you'd have to have a priority queue which determines what the monster thinks is most important at this time and use the desire map for that action. If you then reevaluate the map every turn, the creature should follows it's original plan until it sees something to eat or feels threatened. Combining only makes sense when the desires aren't excluding each other, for example a monster may flee towards a wand it can use, but it makes no sense for a monster to flee towards a corner because it's threatened by something and wants to get the food in that corner at the same time.

[Biag]

@Virex: yeah, I don't think that was the most well-explained potion of the article. The way I understood it was that higher-desired things would have higher to MUCH higher numbers, and overpower the other things in destination selection. For example, if the monster has a Hunger value of 20 and a fearOfPlayer value of 25, then they'll probably swing by the food on their way out, but if their fearOfPlayer value is increased to 200, they'll probably just hightail it to the door.

[Bricks]

Virex, I disagree.  Say you have an entirely symmetrical situation where an NPC is trying to escape.  It will have to arbitrarily choose which direction to head.  Alternatively, place an item like a wand in a corner.  Now which way should it go?  Like Biag said, and as I believe the article described, you sum the product of the coefficient and the map for each desire.  Desires are additive in the sense of how much you want to visit a specific location.  Otherwise you might find a desert just as appealing as a beach.

I wonder how much this could improve companion AI.  It's part of the reason I downright refuse to touch the Bard class in ADOM (Batsworth!  Ratscallion!  Stop screwing around so we can go the next floor!), and I'm leery of Necromancer for similar reasons.

[MaximumZero]

Considering that this is what I want to do with my life, I'll be following with great interest. However, since I have no actual programming experience, I won't be contributing anything to the contest.

[Endek]

I'd definitely be up for this. Count me in

I've already implemented those dijkstra maps in my javascript RL prototype, and was amazed at how they seemed to imbue the NPCs with a clear sense of purpose. Found it very slow to update the map, though, but I guess that's the fault of me having an incredibly inefficient map structure in my prototype and iterating every tile.

[SolarShado]

I'd definitely be up for this. Count me in

I've already implemented those dijkstra maps in my javascript RL prototype, and was amazed at how they seemed to imbue the NPCs with a clear sense of purpose. Found it very slow to update the map, though, but I guess that's the fault of me having an incredibly inefficient map structure in my prototype and iterating every tile.

I'm not terribly familiar with JS, so I'm not sure it this's even possible, but you might want to try multi-threading your calculations. I assume you've got more than one dijkstra map, so you could assign 1 thread per map. Since each thread is accessing shared data in a read-only fashion and writing only to its own separate output, locking/etc. shouldn't be a problem.

But again, I'm not sure if/how you can do multi-threading in javascript.

[qwertyuiopas]

How flexible will the theme be? Could someone make a game where you are a single character who digs trenches to change the flow of various substances to cause both alchemical reactions and defensive trenches, and still fit the theme? If not, could they change it to "flows" of various critters and fit the theme? (I'm not planning to do either, but knowing how far people can stretch the theme is important, otherwise the end result will be twenty or so DF clones...)

[Biag]

How flexible will the theme be? Could someone make a game where you are a single character who digs trenches to change the flow of various substances to cause both alchemical reactions and defensive trenches, and still fit the theme? If not, could they change it to "flows" of various critters and fit the theme? (I'm not planning to do either, but knowing how far people can stretch the theme is important, otherwise the end result will be twenty or so DF clones...)

I almost addressed this in the OP, but decided that would present a different image if put with the rest of the guidelines. The player can have an avatar in the world, like a general or a digger or whatever, but that should simply be part of the control mechanism. Meaning, the gameplay should not focus on the player's avatar. Does that make sense?

[Thendash]

Posting to watch. I've only just begun to learn java so I don't feel like I'm good enough at programing yet to tackle writing a game under such a short time constraint but I'm interested to see what people come up with.

[Endek]

I'd definitely be up for this. Count me in

I've already implemented those dijkstra maps in my javascript RL prototype, and was amazed at how they seemed to imbue the NPCs with a clear sense of purpose. Found it very slow to update the map, though, but I guess that's the fault of me having an incredibly inefficient map structure in my prototype and iterating every tile.

I'm not terribly familiar with JS, so I'm not sure it this's even possible, but you might want to try multi-threading your calculations. I assume you've got more than one dijkstra map, so you could assign 1 thread per map. Since each thread is accessing shared data in a read-only fashion and writing only to its own separate output, locking/etc. shouldn't be a problem.

But again, I'm not sure if/how you can do multi-threading in javascript.

JS isn't capable of true multi-threading, unfortunately. It's far too high-level to give the developer any meaningful agency when it comes to memory management, process control, or other things of that nature.

It does, however, have event handling, so it is possible in theory to simulate multithreading by creating event callbacks which would perform particular calculations asynchronously, which would help stop intensive processes from causing the browser to become unresponsive.

My prototype is a very nastily written thing that is used more as a testbed for quick prototyping of concepts (my most recent one is an attempt at simulating atmospheric pressure to see if an idea of mine set in space would be feasible) before I decide that either the concept doesn't work well, or, if I like it, I then refactor it for inclusion into the more advanced engine that handles things like map subdivision far more elegantly.

[ILikePie]

Can I use bash and dialog for this? Using bash would mean you could only play the game on Unix machines, and adding dialog (for some cute menuconfig-like menus) would mean it'd work on Linux only, or maybe OS X as well, if it's comes with dialog.

[salmonjockey]

Can I use bash and dialog for this? Using bash would mean you could only play the game on Unix machines, and adding dialog (for some cute menuconfig-like menus) would mean it'd work on Linux only, or maybe OS X as well, if it's comes with dialog.

The standard in competitions like these is that you can target whatever platform you want, with the obvious consequence that your audience will match your choice.