Swarm Development: Postmortem

Started by Baron, Thu 21/04/2011 04:58:20

Previous topic - Next topic

Baron

Millions of people get up every morning, feed themselves, and get themselves off to school or work.  People can self-regulate.  Now if you put a lot of constraints on the process, for example everyone had to get up at 6:13 am and make waffles from scratch just the way you like them, then ride a bicycle using no hands to get to school or work; well, then you'd start to run into problems.  The swarm would by necessity have to work by instinct (or the wisdom of each individual's accumulated AGS experience, in our case), which precludes micro-management.  The end result would certainly in no way resemble any pre-conceived game, just as an unmoderated Wikipedia would in no way resemble the authoritative encyclopedia it pretends to be.  But both would result in something interesting and, arguably, something that could not be created through even the most meticulous planning.  Now whether that interesting something would make for a good game is debatable....


Wyz

So when are you going to start this? :D
Life is like an adventure without the pixel hunts.

Baron

#22
Quote from: Wyz on Tue 26/04/2011 13:16:19
So when are you going to start this? :D

   Alright, fine, let's go through an exercise and see how it turns out.  The best case scenario is that we'll change game making forever, and the worse case scenario is that Radiant will end up saying "I told you so!"  In all likelihood we'll end up somewhere in the middle: maybe he'll just say "I told you....", I don't know.  

   So to get the ball rolling I'm going to create a character.  He's a vampire, for no better reason than people want more vampire games to play.  No, he's a ROBOT vampire!  That would be cool.  He's some sort of army experiment that brings him back to his (human) senses, only now he is doomed to obsess over (as well as hunt down and destroy) his lost love, a sexy vixen who seduced him into becoming a vampire in the first place.  They can never be together now, but there's this platonic love thing (at least from his end).  To make things a little disorienting, maybe he's just woken up after the cyborg operation (thus accounting for weird perception difficulties as he moves between rooms with different styles).  Here I leave the torch, since no doubt I will sink well over an hour stitching together contributions, but from here in I want all contributors to exercise enough discipline to only spend ONE HOUR on the project.

   What I am looking for now:

   1) Sprite designer for the main character.  Resolution will be 320x200 to minimize effort, and characters should take up about 25% of this space (so roughly 50 pixels, give or take).  Design means draw forward and side shot (back too, if possible given time constraints).  Something easy to animate would be a good idea, but we'll work with what we get.

   2) Writers, approximately 3.  I need a very BASIC plot and very quick puzzle ideas.  Thinking up some special "power" for the robot vampire would be helpful -otherwise it would be a waste to have such a cool main character.  Maybe he can fly, or see infrared or slow down time or something.  Plot can lead to a suspenseful ending -doesn't have to wind up.  Plot should involve no more than two settings (ie rooms).  For now I'll combine the most compatible ideas into this project resulting in about a 5 room game.  Feel free to write in other compelling characters (with descriptions) and potential dialog snippets (we'll try to get other writers to stitch them in during the wind-up phase!).

 3) Art director (1): All you have to do is come up with a short list of stylistic guidelines for background artists, maybe with a sketch or two (indoor & outdoor).  Remember all backgrounds will be completed in an hour, so limit "style" to drawing and maybe palette: if feasible we'll get other artists to worry about lighting and shading.  Be specific about detail required, so there's not too much inconsistency between artists.  Remember "rough" style is probably better, since we might need quick paint-overs for compatibility reasons.

   Also, if someone wants to co-chair the "co-ordination committee" (ie help me out trying to organize things) that'd be swell.  I'm going to copy this post into the RAT thread and see who I can get to contribute an hour to the project.  Any and all help welcome!

Radiant


Baron

Quote from: Radiant on Wed 27/04/2011 17:39:50
I told you so!  ;D

Well, I was goaded into it, wasn't I?

Upon re-reading my above post I notice I've not explained very well how I envisage the swarm organizing itself, so I'll try to rectify that here:

   To start the project we need a basic idea, like the one I suggested above.  From there we need plot ideas, a sprite designer and an art director (all described in my last post) to each put in one hour creating what essentially amounts to a basic game design document.  Once those tasks are complete, more swarm members can contribute their hour of effort as informed by the work of the tier before them: background artists use plot, character scale and loose design guidelines to draw rooms, for example, and animators draw actions necessary to the plot.  Based on their work more can move forward, etc., until the bulk of the project is done.  Then the pyramid begins to shrink as editors bring it all together.  At no point does one mind control how the game will develop: it will be an organic venture that will depend on the development instincts of each swarm member at each stage -only the rough guidance of the plot outline and art style will be provided.  So for example an animator might have one hour to make the main character sprite reach out to grab something.  He/she would be provided with the main character sprite from the previous stage.  It would then be entirely up to the animator to decide what the reaching might look like.  Or another example: the plot calls for a secondary character.  A sprite designer would be given the plot outline for the character and the main character basic sprite for scaling/style guidance, and the rest would be left up to him/her.  Again: program a simple puzzle.  The programmer would be given the game files (by that point) and the plot outline, and any relevant variable information.  Then he simply scripts the result and communicates variables for the next programmer to use.

Maybe this works better visually:

I concede that someone -probably me- is going to have to spend several hours breaking work down into manageable chunks and polishing the game in the end (especially if there are bugs it doesn't make any sense for new people to try to figure the whole script out).  But I contend that the vast majority of the work for this project can be done in one hour chunks by autonomous agents with minimal direction.  That is the experimental part, though.

Overall, I envisage the experiment running to a MAGS game length, eating up something like 60 man hours (the vast bulk of which would be one-hour contributions).  There won't be anything radical in terms of gameplay that would require a lot of focus of one mind.  Just a simple point & click.

Anyway, I notice everyone got awfully quiet when this became less a theoretical discussion and more of an actual experiment (well, except for Radiant.... :) ).  Thoughts?  Suggestions?  Volunteers?

Babar

When you originally spoke of it, I thought you meant simultaneous man-hours, which would have been more troublesome (and probably more hilarious) to organise. The way you've got it now, it seems that one person would have to do one bit in the chain while the others in that chain wait for it to become their turn.
The ultimate Professional Amateur

Now, with his very own game: Alien Time Zone

TomatoesInTheHead

QuoteThen the pyramid begins to shrink as editors bring it all together.
A pyramid that grows and shrinks is more a diamond than a pyramid :=

I think room scripting could be started in parallel to all the graphical stuff for rooms: as soon as you know what exits and objects etc. there are, you can script it with some dummy sprites. When the room is finished, the script can be polished in terms of coordiantes, walkable areas etc.

From what I get, the process is currently pretty much centered around the plot-writers. A dialog-writer cannot include something about rooms, objects and NPCs that don't exist, a designer cannot make objects that are not used for a puzzle, a room-scripting person cannot include puzzles that need inventory items where he can't be sure the player has them, and so on... so the plot and puzzles must be very detailed from the start.
But let's see how this first experiment turns out, then we can try to take it one step further and see how we can make the process more parallel and remove as much dependencies as possible.

Oh, and you should give that project a fancy name so it can make its way to the list of crowdsourcing projects eventually. :D

Wyz

Well, I'm in for it (obvious from my "So when are you going to start this?" remark). ;) Just to be sure: each of us will only be doing one hour of work right?

That isn't a easy pick but I'm up for writing, music or programming. From my experience with Hourgames I should point out that putting the game together takes an equal amount of time as producing the resources; I guess anyone that is familiar with how AGS works but does not have an obvious talent should enter for the sprite importing, walkables/hotspots importing and interaction writing tasks. Maybe for each room a person that includes finished stuff as it comes, and also codes the interactions. When the room is finished it can be saved as .crm and send to someone who wraps the total game.
The same could be done for characters, they can be saved as an AGS character file. And GUI's could also be stored as a file. That only leaves puzzles, these can not be stored, unless the are contained as a separate room each. Well code by contract could be used: the puzzle designer tell the program what functionality he needs and the programmers write the functions and send in the code as script files.

example: a puzzle that needs certain items (ingredients) added in the right order to from a potion. The puzzle editor calls for a function:

Code: ags

// Order of potion: iFeather, iFrogLeg, iCrowsEye, iCowSpit

int AddIngredient(Inventory *item);
/* Adds an ingredient to the potion:
Returns:
0 - right ingredient, right order
1 - right ingredient but in the wrong order
2 - wrong ingredient
3 - last ingredient, potion is done
(this could actually also be done with an enum but oh well you'll get my drift)
*/

void ResetPotion(); // clears out the mortar


That actually can be used as header of a script file. So there are my suggestions. :)
Life is like an adventure without the pixel hunts.

Baron

Quote from: Babar on Thu 28/04/2011 06:10:34
When you originally spoke of it, I thought you meant simultaneous man-hours, which would have been more troublesome (and probably more hilarious) to organise. The way you've got it now, it seems that one person would have to do one bit in the chain while the others in that chain wait for it to become their turn.

For experimental purposes we've had to compromise the simultaneous aspect of it, it's true.  But there will be many chains -are already many chains, I should say -in progress.  Trust me: from your point of view on the ground the fog of war will be thick indeed, so it might as well be simultaneous.  The payoff (the finished game), however, will take several weeks to assemble, so alas there won't be the "instant" gratification this time around.  Maybe if we swarmed a pre-existing project and remade it we could do it all simultaneously, but that would cut down on the opportunity for creative discovery which I'm really looking forward to.  If you have any ideas how we can practically speed things up, I'd love to hear them.

Quote from: TomatosInTheHead on Thu 28/04/2011 08:53:38

I think room scripting could be started in parallel to all the graphical stuff for rooms: as soon as you know what exits and objects etc. there are, you can script it with some dummy sprites. When the room is finished, the script can be polished in terms of coordiantes, walkable areas etc.

Yes, I think we can have most of the rooms going at the same time, once their usage has been nailed down.  You are correct that much of the project is dependent on plot writing.  It might be fun to try the writing at the end -a challenge to make a bunch of seemingly unrelated actions make sense! - but I think a lot of contributors would get frustrated drawing/scripting/dialog writing into the void.  A few sentences of guidelines can be very helpful, and I'm having a hard time thinking of how we'd get there without at least some writing first.

Quote from: Wyz on Thu 28/04/2011 16:05:20
Well, I'm in for it (obvious from my "So when are you going to start this?" remark). ;) Just to be sure: each of us will only be doing one hour of work right?

That isn't a easy pick but I'm up for writing, music or programming. From my experience with Hourgames I should point out that putting the game together takes an equal amount of time as producing the resources; I guess anyone that is familiar with how AGS works but does not have an obvious talent should enter for the sprite importing, walkables/hotspots importing and interaction writing tasks. Maybe for each room a person that includes finished stuff as it comes, and also codes the interactions. When the room is finished it can be saved as .crm and send to someone who wraps the total game.
The same could be done for characters, they can be saved as an AGS character file. And GUI's could also be stored as a file. That only leaves puzzles, these can not be stored, unless the are contained as a separate room each. Well code by contract could be used: the puzzle designer tell the program what functionality he needs and the programmers write the functions and send in the code as script files.

Indeed a great deal can be done separately without the cumbersome necessity of one large file being transfered back and forth (a nightmare for version control or a huge bottleneck).  I will try to have rooms and characters developed as separate files so that the maximum number of people can be working simultaneously.  I wasn't aware that separate GUI files could be created, but we might as well experiment with that as well (although a conventional GUI is probably a better idea for such a short project with so many different minds working on it).  Using separate script files is also a good idea: indeed, experienced coders might be asked just to write the code for each puzzle "in theory", and then it can be adapted and implemented by one mind very quickly.  Definitely something to ponder.

Thanks to everyone who has volunteered so far, and to all of you who have spared time to put some thought into how a project like this might work.  The more interest the better!  We're still looking for an art director -just one hour writing up some very vague stylistic guidelines and maybe doing an example sketch or two.  It's a remarkably small investment of time to put your stylistic stamp on game!

Snarky

#29
I can't do the art, but since Baron asked for my input despite my extensive experience in unsuccessful self-organizing game projects, I'll pitch in a game design idea:

The main guy -- let's call him... Merrick -- is a vampire, so he has to stay away from direct sunlight. But his robot enhancement includes a small flying probe or drone or what have you that he can send out and control remotely. It can only do very simple tasks (give things a gentle push, perhaps carry one item). And it runs on solar power!

So you can have a general game mechanic where he has to stay in shadows and dark areas, and must send the probe out to do stuff for him in the light, maybe flip switches, push things so they fall over, pick up or drop items, etc. Maybe draw the curtains or even itself cast shadows for you to stay in. And it will only move in light areas.

I was also thinking that when you're in probe mode, the probe always follows (i.e. drifts in the direction of) your cursor (but constrained to light areas), so that the control scheme is a little different. (I'll commit to code that up if/when we get around to implementing the UI.)

Light/shadow regions and using the probe to get at things you can't reach yourself could be the basis for many different puzzles and create some variety in the gameplay. Also, the probe is small and can get through places where you can't. For example:

In one room, you're blocked by light falling across the middle. There's a switch to close the curtains on the far side, but it's in shadow. You first have to use the probe to nudge a mirror so that it casts light there, then press the switch so you can cross.

In another room you can't go anywhere at all until the probe picks up an umbrella (or something to that effect) and shields you. Because it's on the other side of a fence or something, it can't hand it to you, so once you're past it has to drop it.

I was thinking of the probe as a simple floating orb, maybe with a hook. Or it could be his detachable cyborg hand. As long as it doesn't take a lot of animation.

Edit: It doesn't have to be actual sunlight, either. If the game starts in some kind of research lab, they probably have UV floodlights to contain vampires with. Presumably there's been some sort of breakdown that allows our guy to move around (secret research labs, so prone to catastrophic accidents!), a lot of things will be broken or half-broken. Maybe the facilities are overgrown with strange plants and overrun with even stranger creatures. (If they experiment on vampires, they probably have all kinds of other fantasy beasts as well.)

So another simple puzzle could be that there's a floodlight in the ceiling that keeps you pinned down, but the light is flickering and you can see it's almost broken. Just send the probe to knock into it (every time the light flickers, it pauses), and the light goes off.

Oh, and if we need more vampire-based gameplay: What if, although he doesn't normally have a lot of special vampire abilities, when Merrick drinks the blood of some creature, he briefly takes on the powers of that creature? So he could get super strength, ability to climb sheer walls, fly, all depending on what kind of blood he could get his hands on. By drinking human blood, he might learn stuff that person knows (e.g. certain access codes, for example). And if the game takes place in a research lab overrun by fantastic creatures, there could be a lot of strange powers knocking about.

Let's say one room is a medical lab with blood samples. You can't get to them yourself, only pick up one at a time with the probe. You'll have to figure out which one(s) you need, maybe use two or three in the right sequence. (Let's say they spoil quickly, so you can't hoard them for the rest of the game.)

cat

I already did my 1-hour part yesterday and sent it to Baron, but I guess I could also post it here so other people could self-organisely improve it or start the animation. And also to motivate other people to join the swarm.

I did the main character with some kind of bionic hand and an infrared eye sensor to detect people in the dark.



Ali

I asked Baron if I could be Art Director. As soon as I have a free hour, I'll return to this thread with some stylistic guidelines for alla dang y'all.

EDIT:

Here is my attempt at a style guide for backdrops that can be completed in 1 hour. I'm sure it can and will be improved upon in the course of the experiment. Baron is within his rights to reject it, and punch me in the face if he chooses.


STYLE GUIDE:

Sample Backdrop:


It's a little crazy because I crammed a lot of sample textures in to test how they could work.

Layout guidelines:

  • 320x200 with a 20px letterbox at the top and bottom.

  • Flattened perspective, objects drawn like 2D stage scenery (eg tree, rocks, desk).
  • Walkable area limited to bottom third (except in the case of stairs, multi-level locations etc.).

Colour guidelines:

  • Colour palette limited to greens, yellow, purple and orange (see below).
  • Shaft of lights which the characters can walk behind are solid patches of yellow at 50% transparency (see opaque shaft below).
  • Double-pixels used for major outlines and shading (e.g. the walls).
  • Single pixels used for texturing (e.g. floorboards, stonework).

Texture guidelines:

  • Foreground objects drawn in silhouette.
  • Hatching (vertical or horizontal lines) give shape and shade sparingly.
  • Non-continuous patches of texture Use hatching (vertical or horizontal lines) to give shape and shade sparingly.
  • Add shadows/texture around objects e.g. (around tree/rocks).

Okay... this has taken a bit more than an hour to write up so I'm going to stop!

Baron, feel free to revise these outlines or alter them entirely!

Baron

Awesome, awesome, awesome!  All these contributions (including the writing I've received so far) are top notch: how do you guys do this stuff so quickly?  At any rate, we're quickly getting close to being able to start some actual production.  If anyone feels that they can spare an hour to improve the main character sprite or the background style please feel free.  Also I'm wondering if there is a talented writer/editor who would like to spend an hour attempting to merge the main plot details of our writers into something resembling a coherent script (note: does not have to be actually coherent, just resembling coherent).

In other news TomatoesInTheHead has volunteered (or was volunteered, I can't remember....) to be our title guy.  He has come up with a working title: Project Crowd-venture!  So that's what we'll call it until he has seen enough of the actual plot to start coming up with actual titles.  I expect he'll come up with an hour's worth and we'll just have all the participants vote on them to determine the actual title. 

   I would also like to say that I'm very happy with how many people have volunteered in such a short period of time.  About 15% of our estimated required labour force volunteered in just the last 2 days, without so much as a screenshot or a confirmed plot!  I'm sure once we can cobble together a couple of screenshots to start a GiP thread and show actual progress we'll have volunteers stepping over each other to participate.

Jim Reed

Heh, this looks like it's shaping up nicely, Baron. Good luck with the swarm!

Charity

I don't know if there is room for random, non-time substantive brainstorms in this project, but:

If we're going with the vampire and probe idea, perhaps there could be some rooms that use artificial, non-UV lighting that allow both the probe and the vampire to move freely, though perhaps both experience sub-par functionality in these rooms.  There could even be a puzzle where the vampire has to figure out that the light is safe before entering it.

Armageddon



A continuation of the last background, I tried not to copy too much.

EDIT: I made a prefab .psd for fast indoor background creation, add more if you'd like, try to keep it organized though.

http://dl.dropbox.com/u/5479044/Background_Prefabs.psd

Also put me down as a background artist, this could be fun. I mean if you like my background that is.

Baron

PROJECT UPDATE:
Alright, we've got all the plots.  I might as well share them for everyone who's interested (I've disguised the identities of the authors so that there is no bias):

        Plot #1
        Plot #2
        Plot #3

   Currently there is an editor charged with the task of distilling a simple walkthrough for our project based on the ideas listed from the texts above (in one hour, no less!).  I'm not sure what he'll come up with, but hopefully it will give us enough of an idea of what the game will look like that we can start designing locations and additional characters.  Granted, the plots are all quite distinct from each other, but careful reading does show some similarities that I think we should try to keep: different species blood/juices give the vampire different powers and his robotic functionality should somehow be linked to evading death-by-light, for example.  Anyway, I know we're all waiting with baited breath to see what will emerge.

   In other news I spent all of five minutes playing with Cat's sprite.  I made the cloak contrast a bit more with his garments, and expanded the robotic arm and eye implants so that they'll read easier at small resolutions.  Any one have any suggestions?




Edit: By the way we have 17 volunteers on the books so far!

Armageddon

I liked the intro and idea of Plot #1 (GRAPEFRUITS!) although I didn't like the puzzles or location or the orange juice part of it. The twin wives would be cool characters if they sorta talked at the same time or finished off each others sentences or whatever.

I don't like the idea of Plot #2 having a probe fly around sort of makes the Adventure genre more complex than it should be, in my opinion.

I liked the name of Plot #3 I didn't really like the whole military facility thing though.


I was thinking this would be more of an old castle/hospital place by the sea, or like the original Castle Dracula where tons of gruesome experiments were performed except it's more modern with robots and things. I also think Merrick shouldn't be holding his hands together as it looks unnatural and a walking animation for that would seem strange.

Hudders

If someone can animate this, (from plot #1):

Spoiler
Contains:   Chair with straps (solution: use vampire teeth on)
[close]

that would be awesome.

Igor Hardy

I really like the art so far.

Quote from: Hudders on Tue 03/05/2011 13:59:52
If someone can animate this, (from plot #1):

Spoiler
Contains:   Chair with straps (solution: use vampire teeth on)
[close]

that would be awesome.

I could do an hour-worth of animations. Is there a list somewhere of what is needed? And do I need a badge before I'm allowed to animate? (I didn't read through the entire thread)

SMF spam blocked by CleanTalk