Flash on the Beach 2010 – This guy’s experience.

I’m writing this on the train home from Flash on the Beach. I should probably be coding a game like I did on the way to the conference, but I’m finding that after sensory overload from 3 days of sessions, a few late nights and the terror of doing two(!) sessions on stage, my brain is unable to do code right now.

My main session seemed to go very well, although I’m gutted for all the people who couldn’t get in because it was so packed. The demos from the session are already available at DullDudeGames.com/game-designer and I’m going to add a transcript of what I said and my slides very shortly. If you were an attendee there should be a video at some point, although I didn’t see any cameras so maybe it’s just a screen cast or something. It was my first full session at a big conference, and definitely an item ticked on my bucket list. I’m mega grateful to conference organiser John Davey for putting me up there, and Pokemon Trainer Seb-Lee Delisle for being my hype-man for the last 2 years. Without Seb, probably nobody would know who I am. He’s also a bloody ruddy nice bloke.

As happens at every talk I give (will they never learn!), some people asked me afterwards if I would share the source code from my talk. The demos are built on my work-in-progress game engine, so I can’t share the source to the games without also including the game-engine, which for business reasons I am not able to do, although a big chunk of it is already open-sourced in my Gamepad library.

The funny thing about geeks is, even when I do a session about design – purely the creative decision making side of games – geeks still just want to look under the hood! The main feedback was that people wanted to know the maths for the car handling physics, so I’m going to turn that into a tutorial and post that here on my blog. Also, if you’re more into the wizard stuff, I have the full source for a different Zelda-style wizard game that I made a while ago, that I am going to give away here too.

Slightly weirder was my last minute addition to the “jam throwdown” on the terrifying main stage. I was only asked to do it the night before, and had planned to show my complete greatest hits from the last 11 years of my work. Not having a chance to rehearse, I didn’t know that this was actually impossible, so what the attendees got was about the first 5 years, which was loads of terrible, daft lo-fi experimental Flash stuff and a few games. It seemed like people got the joke, but I did have a small bout of paranoia that everyone hated me straight after. Do not underestimate the social paranoia appearing at conferences can give you! I have had some nice feedback about the slot on twitter since though, so I feel ok about it now. People were probably just being nice, but thank you, I appreciate it!

(photo All rights reserved by oyvindnordhagen)

Even though I’m just a geek talking about geeky stuff, it’s really hard not to get a bit of a rush from making a whole auditorium of people laugh or go “oooh”. I wonder whether this is why people like Hoss Gifford and Brendan Dawes pack their sessions with jokes? I don’t think I got into speaking as a back door into stand-up comedy, but to be honest I have no idea why I did want to get into it. Certainly a large part of me finds the whole experience completely harrowing. What is “speaking” or being a “speaker”? Why do it? What is it for? I have genuinely no idea.

That’s enough about me though, what about all the talented speakers? Well it was a slightly strange one for me. For one thing, since last time I have become even more focussed on game development, and there wasn’t much representation for gamedev, so there weren’t many sessions that apply directly to what I do day-to-day. From the queue outside my talk and Jon Howard’s (excellent) session, there’s obviously a lot of pent-up demand for it among attendees. Maybe next year John could try to book some Flash gaming luminaries like Dan Cook or Adam Atatomic? I’ll definitely suggest it to him.

I want to word this next bit very carefully so that I don’t offend any of the amazing speakers who I saw and spoke to during the conference, and who are all some of the most talented and inspirational people you will meet in this industry. Here goes then: basically what I’ve noticed is that not every speaker can inspire you as much the second and third time you see as they do the first time you see them. Even though everyone updates their material regularly with new work, nobody has an unlimited supply of mind-blowingness. Ideas that are revolutionary when you first see them get synthesised into the way you see the world. The same idea can never inspire you again in the same way it did the first time. This is no reflection on the talent of the speaker, it’s just the nature of ideas.

And this is why I loved the elevator pitch so much. Twenty speakers from completely different disciplines blasting you with as much information as they can get into a 3 minute slot. I know from experience that each 3 minute speech has many hours, if not days and even weeks of preparation go into it. This shows through massively and you get exposed to more new ideas in this session than any other. Someone tweeted that speakers collect elevator pitchers like Pokemon. I kind of played Pokemon Trainer a bit this year, encouraging Andreas, Jasper and Tom to do the pitch. This reflects very well on how I choose my friends, because my Pokemon were easily 3 of the best in the whole session.

Andreas presented his DDConsole aka DoomsdayConsole which is a runtime tracing/hacking/debugging/tweaking console for Flash, inspired by the ones you find in games like Quake. It’s completely open source and completely ace. Through some wicked 8-bit graphics and animation, Tom presented his awesome sound effects generator AS3sfxr with such style that he also managed to subconsciously teach you how to do the creative side of sound design.

Jasper showed us all why Unity3D is the crown prince of 3D game engines by building his slide-deck as an Incpetion/Matrix inspired fly-through of a city falling into place. Nobody loves Unity as much as this guy. He’s thinking about it most the time. He’s probably thinking about it right now. What he didn’t tell you during his slot was that he taught himself to model in Maya just to make his presentation. And you’d never have noticed it, because his models looked awesome. I could tell that Jasper enjoyed the experience because I spoke to him after and he had a crazy look in his eye like someone who had just discovered crack. Anyways, if you went to the conference, please vote for these guys on your feedback form, they’re ace.

The stand-out session of the conference had to be Seb’s “What the flux?” There’s been a lot of mud thrown back and forth this year between Flash developers, HTML5/web standards advocates, and of course Apple, and Seb has spent a lot of time trying out these various competing technologies. He presented a some home truths that I don’t think everyone was ready to hear, but presented it a way that was optimistic, pragmatic and tried to build bridges. He had video interviews with some smart people from both camps, which he cut in with some political slogans, a game show skit and some jokes. I don’t actually agree with him that JavaScript and HTML5 are anywhere near ready to move into the immersive/experiential/flashy uses of Flash, but on everything else he was about right. He did point out that Flash is still the mofo daddy for web games.

Grant Skinner showed some cool physical device experiments like using Android phones as controllers for 8 player Asteroids and even as accelerator pedals for Scalextric. Also there was some marital aid hacking which raised a few eyebrows. In Grant’s session I began experimenting with starting rounds of applause, my first attempt failing and making it seem as if I was slow clapping, which was quite an amusing little moment that he dealt with like the pro he is. After that the audience really got into it, and through my further efforts I managed to increase the numbers of rounds of applause in all sessions I went to by at least a factor of 2 for the rest of the day. This is great for the atmosphere of the conference. If everyone is showing appreciation of a good speaker it actually increases the speaker’s confidence and makes for a better session. Towards the end of the conference I noticed an annoying trend that people were only clapping videos and not any other type of work that was shown. This is really silly and completely unfair on speakers who do different kinds of work, e.g. INTERACTIVE STUFF.

Other highlights were surrealist animator Cyriak Harris, design legend Stefan Sagmeister and designer/director Nando Costa who made this year’s amazing title sequence, which is frankly the sexiest form in which my name has ever been written. Watch it and see why. Nando showed some wicked film work he’d done with Modest Mouse (who I love) as well as other portfolio stuff. I chatted with Nando later about life in Portland and he’s a thoroughly nice bloke. I am now 1 degree of Kevin Bacon away from Modest Mouse which is a nice thought, although completely pointless.

I am completely exhausted now, so I’m going to get back to normality for the rest of the year, and take my Zero to Game Designer in 60 Minutes talk on the road in the new year. Thanks so much to everyone I met for being so supportive!

Gamepad ported to Processing / JAVA

Gamepad, my opensource keyboard input utility, has received a great reaction so far, with many retweets, over 1000 views, and some great partnerships with other libraries in the pipeline. And now it's not just for Flash, because Michael Heuer has ported the whole project to JAVA for use with processing. So, if that's how you get down, go check out Gamepad for Processing on Github. Big thanks to Michael for taking the time to create this.

Introducing Gamepad.

Gamepad is a free, open source project by me, Iain Lobb, with the aim of greatly simplifying keyboard input for Flash games. The idea was born out of 2 realisations – first, that since key.isDown was removed from ActionScript it has been more work for game developers to handle keyboard input, and second – that developers, me included, were not working with keyboard input at a sufficient level of abstraction. Trust me, if you make Flash games, you need this in your life. Update: I've had some feedback that it's hard to see the download button on Github, so please CLICK HERE TO DOWNLOAD if you prefer.

What does it do?

Gamepad simulates an analog joystick input using the keyboard. Many times when we access key presses, what we are really doing is pretending that WASD, the arrow keys or some other combination are actually a D-pad or joystick with an X and Y axis, and 1 or 2 fire buttons. Gamepad handles the event capture, maths and other details of this for you, so you only have to think about how you want your game to respond to this input. A detailed explanation follows, but why not just download the source code and play around?

A simple example

First we create a gamepad. It needs a reference to the stage so it can capture keyboard events, and it needs to know whether it is simulating a circular movement space, like a thumb-stick, or a square one, like a flight-stick. This argument is called “isCircle” I’ll get into this distinction later.

var gamepad:Gamepad = new Gamepad(stage, false);

Then in your update / enterFrame function, you simply adjust to the position of your character based on the “x” and “y” values of your gamepad. The “x” and “y” values are always between -1 and 1, so x = -1 would be mean the virtual D-pad is pushed all the way to the left, and x = 1 would mean it would be pushed all the way to the right. I have chosen to use y = -1 as up and y = 1 as down, so that it matches Flash’s screen coordinate system.

character.x += gamepad.x * 5;
character.y += gamepad.y * 5;

And to access the fire buttons, we simply look at the “isDown” property of the “fire1” button.

if (gamepad.fire1.isDown) fire();

That’s it! Your character will now happily walk around the screen using the arrow keys and fire when you press the CTRL key. As Gamepad also has easing by default, the character will also accelerate and decelerate smoothly!

The “isCircle” property

A common mistake that developers make in top-down perspective games, is to allow the player to move too fast diagonally. They say “If the up key is pressed, move up 5 pixels, and if the left key is pressed, move left 5 pixels, so if both are pressed move up 5 pixels and left 5 pixels”. This is wrong! Pythagoras tells us that the speed of their character would now be the square root of five squared plus five squared, which is the square root of fifty, which is about seven. So now their character moves 5 pixels per frame horizontally or vertically, but 7 pixels per frame diagonally. Disaster. Once you know this you can handle it yourself, but Gamepad makes it easy by giving you the isCircle option on creation. When you create your gamepad, simply pass in true for the second argument:

var gamepad:Gamepad = new Gamepad(stage, true);

Now the “nub” of the virtual joystick is limited to a circular area, meaning if you hold the “down” and “right” key together, it will report values of roughly x=0.7 and y= 0.7, instead of x=1 and x=1, and movement speed will be equal in all directions. Typically you would use this option is arena shooters and Zelda-style adventure games, and you may want to use it in scrolling shmups, but that’s a greyer area in terms of “realism”, as your vehicle is already supposed to be moving at a high velocity.

The “ease” property

By default, gamepad will give you a nice easing motion. One advantage it brings is that the player can “tap” the keys to achieve the effect of a half press on an analog input such as the XBOX 360 thumb-stick. Often, though, you won’t actually want to use this. You can easily turn it off by passing in 0, or pass in some other value for more/less responsiveness. Typically you’ll want easing activated for simple games, but in more complex simulations you handle acceleration elsewhere, so you may want it deactivated. To deactivate easing:
var gamepad:Gamepad = new Gamepad(stage, false, 0);

The “autoStep” property.

Gamepad needs to update at the same rate as your game, so that the easing, and the “downTicks” and “upTicks” properties (which I’ll cover later) always keep in sync with your game. If you are simply using Event.ENTER_FRAME for your update, you don’t need to do anything, as this is the default. However, if you have some other system, you should pass in false for the autoStep property and manually call the public “step()” method every time you update.

I advocate using a frame-based tick, and using the “fix your timestep” methodology if you need to stay in sync with real time. This way your game is deterministic, you will have far fewer inconsistencies with collision detection, and you can safely use basic Euler calculations for acceleration. However, I understand that some developers, and some game engines, such as Flixel, use a deltaTime based approach. If you are using a time-based approach, you should use the “fix your timestep” principle with gamepad, by calling the “step()” function an appropriate number of times each update, based on how much time has passed since the user started the game.

The GamepadInput class.

Each Gamepad instance has a set of GamepadInput objects that represent individual “buttons” on a virtual joypad. These are up, down, left, right, fire1 and fire2. However, these do not map one-to-one with keys on the keyboard – one GamepadInput can be linked to one, two or more keys, so that you can easily provide simultaneous alternate control schemes. The classic example would be having both WASD and the arrow keys control your character, so that players can use whichever scheme they prefer without having to set any menu options.

For setting up keys, GamepadInput has the “mapKey” function. It takes two arguments: “keyCode” – an integer representing the key, for which you should use the constants in the handy “com.cheezeworld.utils.KeyCode” class - and “replaceAll” which specifies whether to overwrite existing mappings. If you want to have multiple keys mapped to the same input, pass in false.

In many cases, you will never need to call the “mapKey” function, as there are presets for the most popular configurations in the Gamepad class. These are the functions “useArrows”, “useWASD”, “useIJKL”, and “useZQSD” (which is for the French “AZERTY” keyboard layout, where WASD doesn’t work). All of these methods take a “replaceExisting” argument which specifies whether you want duplicate mappings, as discussed earlier.

Unfortunately, the keys developers use for fire buttons don’t seem to be as standardised, but I have done my best, by providing the methods: “useChevrons”, “useGH”, “useZX”, “useXY” (for German QWERTZ keyboard) and “useControlSpace”. These are all taken from popular Flash games, for example Nitrome’s “Double Edged” which handles 2 players by giving one player WASD for movement and GH for attacking, and the other player arrow keys for movement and chevrons, “<” and” >” for attacking. Gampad’s default: Arrow keys, CONTROL and SPACEBAR should be fairly safe for all players, but there are many, many issues around international keyboards, so it may be advisable to allow the player to set their own keys.

Once you have your inputs set up, you can get information about their state: “isDown” simply tells you if a key is being held down, “isPressed” tells you if a key was pressed this frame/tick/update, and should be used instead of listening for KEY_DOWN events, “isReleased” tells you if the key was released this frame/tick/update, and should be used instead of the KEY_UP event, and “upTicks” and “downTicks” tell you how long the key has been held or released for. Basically, “isDown” is your go-to, but the others are there when you need more info. Generally, you should use the x and y properties of Gamepad for movement, but sometimes you may want to access the D-pad as “buttons” instead, for example using “up” for jump.

The GamepadMulitInput class.

There are also a further set of “buttons” represented by GamePadMultiInput objects. These are special as they aggregate the inputs of multiple other “buttons”. These are “upLeft”, “downLeft”, “upRight” and “downRight”, which let you treat these combined directions as if they were individual inputs on an 8-way controller, and “anyDirection”, which lets you know whether the player is pressing in any direction on the D-pad.

The angle, rotation and magnitude properties.

You may need these additional properties from time to time: “angle” gives you the direction in which the stick is pointed as radians, “rotation” is the same value but expressed in degrees and “magnitude” is the scalar distance of the “nub” from the origin, ignoring the angle. For example, if you set the rotation of a character MovieClip to negative the rotation property of your gamepad, your character will face in the right direction when they move!

The GamepadView class

If you want to visually see what your gamepad is doing, simply create an instance of the handy GamepadView class, and initialise it with a reference to your gamepad and optionally a colour.

var gamepadView:GamepadView = new GamepadView();
gamepadView.init(gamepad, 0xFF6600);
addChild(gamepadView);

GamePadTester and PlatformGamePadTester

In with the source code, you will find two visual “tester” classes. It’s not quite test-driven development, but these are the classes I use to ensure all the functionality of Gamepad is working – they’re also great documentation for Gamepad’s APIs, or could even be the starting point for your own game! If you’re using the Flash IDE simply open GamepadTester.fla or PlatformGamePadTester.fla and publish. If you’re using the Flex compiler you’ll need to create a new project and set the “always compile” / document class to com.iainlobb.gamepadtesters.GamePadTester.as or com.iainlobb.gamepadtesters.PlatformGamePadTester.as.

GamePadTester shows the basics of doing car-style movement and top-down character movement. It also demonstrates duplicate controls, with both WASD and IJKL controlling the character. PlatformGamePadTester shows the basics of a two player platform game, including variable height jumps (although in the end it transpired that these are mostly handled outside of the gamepad class). It also shows how you can create two instances of the same character class with different control schemes, without a single “if” statement or use of polymorphism. Composition FTW!

Final thoughts

Well done, you have made it through all the Gamepad documentation! Please start using it and submit feedback to my blog, or on github. I’ve had versions of this class kicking around for almost 2 years, but I had no idea how much work it would be to actually pull it all together, test and document it to a state where I was happy to release it as an open source project. I have insane new levels of respect for anyone else out there running an open source library. The license is MIT, which basically means you can do whatever you like with it, as long as you don’t blame me when it goes wrong. I’d appreciate it if you didn’t change the package names, and removing the copyright notice is forbidden. There’s a sweet gamepad logo that you can add to your game if you like, but it’s by no means compulsorily, and if you want to hit up the donate button on github, I’m not going to stop you. Enjoy!

Oh yeah, follow me on twitter: http://twitter.com/iainlobb