Paul Solt

I presented at BarCamp #5 Rochester http://barcamproc.org/ at RIT on my ArtworkEvolution project. Here are my presentation slides.

ArtworkEvolution-Barcamp

I evolved images during the presentation and these are some of the results:

Curved Sky

TV

Pillar

A lot of people have been talking about the iPad. Here are my opinions on the future of iPad, computing, and entertainment.

The iPad is set to revolutionize how we interact with multimedia content and computers. There are a number of reasons that make the hardware and software standout. First and foremost, it is affordable cutting-edge technology. The $499 price point means that it is not out of reach for average consumers who are interested in an updated “all-in-one” computing device. All 9.7 inches of the screen are multi-touch, which will allow software developers to create very interactive applications. Star Trek, Avatar, and other science fiction movie computer interfaces can finally be realized on a large multi-touch screen. The device is connected, which allows the consumer to use it anywhere. Lastly, the device will provide the ultimate responsive user experience.

Home Entertainment Revolution

Apple now has the ability to revolutionize the home entertainment market. They are provided a multi-media portal, which will change the way we use our TV’s, computers, and music players. Imagine controlling a TV from the couch without attaching any wires. A user might want to watch “Batman Begins” on their 42” HDTV. A few touches will open iTunes and start the movie. I mentioned the TV, how does that fit into the picture? The movie streams wirelessly in HD from the couch to the 42” TV. Gone are all the cables, remotes, and hassles. Don’t bother with power cable, since the device will play content for 10 hours straight. The entertainment cabinet can be cleaned out. Throw out the VCR, CD player, DVD player, Blue-Ray player, cable TV, satellite TV, and digital antennas because they are not needed. Apple will be the one stop remote control into all media content and it will be seamless to use and control.

Affordable Technology

A few years back, in 2007, Amazon set out to take over the digital books arena. They did a pretty good job at providing access to books, but that is about all they did. The Kindle DX costs $489 and is just a digital book reader. It has limited processing power and storage space. The main attraction is the e-ink technology that is supposed to be easier to read. Overall, the device is nice, but is very limited in the target audience and lacks multi-media capabilities like an iPhone.
Apple worked hard to set the price point of the iPad as close as possible to the Kindle, because they are directly competing with Kindle’s e-book market. For $10 more one can get a fully color display that can play videos, music, games, display e-books, and run applications. Apple did a wonderful job in selecting a set of features that could be combined for a relatively low price point. The device is slightly more expensive than other eBook readers and net books, but not overly expensive.

Large Multi-Touch Screen

For the longest time computers were something that required skill to use. However, this learning steep learning curve is almost no longer the case with the iPhone, iPod Touch, and iPad. The iPhone revolution brought capacitive multi-touch screens to the public. In English this means that a user just touches, not “presses,” the screen to perform actions. iPad is riding on that revolution wake and it is taking it step further by increasing the size of the screen. This technology is not foreign; it is mainstream and it is here to stay because it works. If a user knows how to use an iPhone or a laptop track pad then the transition is smooth. The touch screen is key, because it allows people to interact with a device just like they might interact with a microwave or a washing machine. A user physically touches, taps, and slides controls around that directly mirror the physical world. The iPad is a natural user interface and it is what most people what, but do not know how to ask for.

Software is Key

The main attraction with any working piece of hardware is software. People want to use a piece of hardware that is customizable. At any given point the device can assume different roles, because it was built to be extensible. In one instant it is an email program, movie player, music player, and then an entire college library. Apple has created a platform that provides many inputs and outputs that software developers can hook into to provide new and novel user experiences. The software development kit (SDK) has given developers direct access to technology that was locked down or too expensive to use. Developers can use a digital compass, accelerometer, multi-touch screen, microphone, and motion sensors to interact with a user in astonishing ways.

Connected

The iPhone provided the all-in-one experience because it can double as a music player, movie player, email program, Internet browser, and eBook reader. It was small, but it had the ability to execute each of those tasks. It has those abilities because of the Wi-Fi and 3G data connections. These connections make it possible to see content beyond the walls of a single hard drive. It provides a much richer experience to user. The iPad takes these same tasks and now makes it better by providing a bigger experience. Users can use these connections in a larger form factor and can be more productive. For most users a simple Wi-Fi connection will be all they need from the couch in the living room. Some users might be active and on the go, so they will need a 3G wireless connection. Apple has realized this connection issues and separated both technologies to reach different consumers needs. Users can get the Wi-Fi by itself, or combine Wi-Fi and 3G if they need to always have a connection the the internet.

User Experience

Users want fast responsive devices, not sluggish devices. A lot of users complain that they cannot run multiple applications (multi-tasking) on the iPhone, but what they do not realize is what they have to give up for multiple applications. Running anything in parallel on a mobile device means that it is dividing computing resources and power among applications that are invisible in the background. These resource hogs will slow a device down and drain a battery.

Traditional multi-tasking is not what users want. Apple supports multi-tasking, but only to first party applications. In restricting access, Apple has complete control of the user experience. Third-party multi-tasking is not supported for a few reasons.

• Window’s Task Manager is a power user feature that is unnecessarily complicated. On a Windows Mobile 6.x device, task manager is a terrible experience. For example, pressing the ‘X’ on an application is not guaranteed to close the application. The button may only minimize the application, in which case it is still using computing resources and draining the battery. The ability to manage open applications is a power user feature on a mobile device and should be hidden from a typical user.
• What is the difference between running an application in the background and running an application one at a time if the transition from one application to the next is fast and seamless? Does the experience have to differ solely from a technicality? iPhone applications can save state from the last thing they were doing when they are closed. For example, if a user is composing an email about a trip on an iPad. They need weather information and decide to check the weather with the following steps.
  1. Press the home button.
  2. Touch the weather application.
  3. Press the home button.
  4. Touch the email application.
  5. Resume composing the email with the updated weather knowledge.
• Running applications in the background allows companies to directly compete with Apples multimedia business. iTunes is one of the few applications like Mail and Messages that run in the background. A user can play music through iTunes while using different applications. If a user could use Pandora for music in the background, then they would have a smaller reason to stay on the iTunes platform. I do not see Apple changing this policy, since it is not in their best interest.

Conclusion

The iPad will simplify the experience to download new movies, games, music, books, and utility applications. There is no doubt in my mind that Apple will continue to innovate on this new iPad platform to further simplify and connect the multimedia experience in every persons home. The iPad is magic and just works.

For my Genetic Algorithms class I decided to work based on the work of Karl Sims. I have built a system that can evolve images based on “user natural selection” over a period of time.

The user plays god and can select those images that will live and die to produce new images. Here’s a fun sample of images:

Curtains

Waves

Loops

Logo

Squirley

Hungry Curves

A Future Road

Wallpaper

Pipes

Peace

Planet Rings

Here’s my second presentation on iPhone Development at RIT for the Computer Science Community (CSC).If you enjoyed it let me know. I decided to look into graphics and OpenGL for the presentation.

Slides: iPhone Development II – Paul Solt

Demo: Triangle Demo: OpenGL ES on iPhone

OpenGL ES Triangle Demo

The first demo is a basic OpenGL ES iPhone project using vertex/color arrays and an orthogonal view. It’s based on the tutorials from Jeff LaMarche This is the starting point for any iPhone OpenGL ES project. It’ll give you a window that you can draw in and manipulate OpenGL state. Use Jeff’s xcode project template to make this process painless and easy to get started. It’s easier than using GLUT!

Demo: Cocos2D for iPhone: Cocos2D iPhone Graphics Demo

Cocos2D Demo for iPhone

A 2D graphics package to aid graphical applications on the iPhone. It provides some really neat animation support along with project templates to make setup a breeze. Cocos2D comes with two separate physics packages that you can incorporate into your game. The demo shows an animation trigger when the user pressed on the screen. Animations can be built from simple actions and combined to create a complex animation. It’s perfect for a platformer.

Demo: Raytracing on the iPhone

Raytracing with Photon Mapping

I ported my computer graphics II ray tracer from OpenGL/GLUT to OpenGL ES on the iPhone. The performance is very slow on the iPhone 3G (133-230 seconds), but that’s to be expected with the given hardware. In the simulator it can render a scene in about 15 seconds. This demo shows how the performance of the actual device is vastly different from the iPhone simulator. I’m sure I can rework portions of the ray tracer to be more efficient, but that wasn’t the goal of the quick port.

Resources:

• OpenGL: “The Red Book” OpenGL(R) Programming Guide: The Official Guide to Learning OpenGL(R), Version 2.1 (6th Edition)
• OpenGL ES Tutorials: http://iphonedevelopment.blogspot.com/2009/05/opengl-es-from-ground-up-table-of.html
• Cocos2D: http://code.google.com/p/cocos2d-iphone/

Today I gave a presentation on iPhone Development at RIT for the Computer Science Community (CSC).If you enjoyed it let me know. I’m looking into starting an informal iPhone Dev workshop for more topics.

Here are the slides and Xcode projects:

Slides:  iPhone Development – Paul Solt

1. Demo: Hello World Pusher:  Foo2

2. Demo: Touch Input:  Stalker

3. Demo: Robot Remote Control:  See my previous post

*The Touch Input demo was based on a demo given during the Stanford iPhone courses available on iTunes here.

Resources:

• Cocoa Programming for Mac OS X by Aaron Hillegass (Third Edition)
• Stanford iPhone Course (cs193p.stanford.edu)
  • Search “iPhone Application Programming” in iTunes
• Beginning iPhone 3 Development: Exploring the iPhone SDK by Jeff LaMarche
  • Blog:  iphonedevelopment.blogspot.com
• iPhone Dev Center
  • developer.apple.com/iphone

Here’s the iPhone Player/Stage Remote Control project! There’s a .pdf that describes how to setup Xcode in the .zip file.

Controlling a Robot over Wi-Fi

A Virtual Robot in a Virtual World

The goal of this project was to use the Player/Stage robotics code on the iPhone to communicate and control robots. I discuss how to setup the Xcode development environment. There are two example Xcode projects. The first one is an Objective-C project that wraps around the C++ Player/Stage code. The second project is a very primitive C++ program running on the iPhone without any UI. Both of these Xcode projects are fully documented and will serve as a starting point to iPhone Player/Stage development.

iPhone Player/Stage Remote Control Project: iPhonePlayerStage

Feel free to ask questions and let me know how you use the code.

Get ready for an iPhone robot remote control. I’m working on creating a couple samples and documentation.

Check out the video from Imagine RIT: http://www.youtube.com/watch?v=MOxTV41Lqac

Here’s the link to my follow-up post: iPhone Player/Stage Sample Projects

I created a C style driver program that used OpenGL/GLUT for my computer animation course projects. It worked fine for the first project. However, there were multiple projects and they all started to use the same boiler plate code. In order to reuse code, I decided to refactor and make an extensible class to setup GLUT. My goal was to make it easy to extend the core behavior of the GLUT/OpenGL application.

As I refactored the code I decided to make a base class to perform all the GLUT setup. There’s a first time for everything and I didn’t realize the scope of this change until I was committed to it. I will summarize the problems and how to solve them. The code and explanations should provide the basic understanding of what is happening, but it will not compile as it is provided.

Problem 1: My initial attempt was to pass member functions to the GLUT callback functions. However, you can’t pass member functions from a class directly to C callback functions. The GLUT callback functions expect a function signature exactly as function(BLAH) and I was giving it something that was function(this, BLAH). Where the “this” portion was the object passed under the hood.

class AnimationFramework {
public:
	void display();
	void run();
	void keyboard(unsigned char key, int x, int y);
	void keyboardUp(unsigned char key, int x, int y);
	void specialKeyboard(int key, int x, int y);
	void specialKeyboard(int key, int x, int y);
	void startFramework(int argc, char *argv[]);
};

… in the setup function for GLUT

void AnimationFramework::startFramework(int argc, char *argv[]) {
	// Initialize GLUT
	glutInit(&argc, argv);
	glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
	glutInitWindowPosition(300, 100);
	glutInitWindowSize(600, 480);
	glutCreateWindow("Animation Framework"); 
 
	// Function callbacks
	glutDisplayFunc(display); 		// ERROR
	glutKeyboardFunc(keyboard); 		// ERROR
	glutKeyboardUpFunc(keyboardUp); 	// ERROR
	glutSpecialFunc(specialKeyboard); 	// ERROR
	glutSpecialUpFunc(specialKeyboardUp); // ERROR
 
	init();			// Initialize
	glutIdleFunc(run); 	// The program run loop  // ERROR
	glutMainLoop();		// Start the main GLUT thread
}

Solution 1: Create static methods, and pass these static methods to the call back functions. All the logic for the AnimationFramework will go into these static methods. I’ve fixed the compiler errors, but it feels like a step backwards from what I set out to do.

class AnimationFramework {
public:
	static void displayWrapper();
	static void runWrapper();
	static void keyboardWrapper(unsigned char key, int x, int y);
	static void keyboardUpWrapper(unsigned char key, int x, int y);
	static void specialKeyboardWrapper(int key, int x, int y);
	static void specialKeyboardUpWrapper(int key, int x, int y);
	void startFramework(int argc, char *argv[]);
};

… in a setup function for GLUT

void AnimationFramework::startFramework(int argc, char *argv[]) {
	// Initialize GLUT
	glutInit(&argc, argv);
	glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
	glutInitWindowPosition(300, 100);
	glutInitWindowSize(600, 480);
	glutCreateWindow("Animation Framework"); 
 
	// Function callbacks
	glutDisplayFunc(displayWrapper);		// NO ERRORS!
	glutKeyboardFunc(keyboardWrapper);
	glutKeyboardUpFunc(keyboardUpWrapper);
	glutSpecialFunc(specialKeyboardWrapper);
	glutSpecialUpFunc(specialKeyboardUpWrapper);
 
	init();					// Initialize
	glutIdleFunc(runWrapper); 	// The program run loop
	glutMainLoop();				// Start the main GLUT thread
}

I’ve made a bit of progress, but when you think about it, static methods aren’t much different from the plain C functions.

Problem 2: I want to have the functionality in the instance methods, yet I’m stuck with static methods. I’m not encapsulating the behavior inside instance methods, and I have no option for inheritance. I’ve still failed to hit the initial goal of creating an easy to use object oriented GLUT wrapper that’s extensible.

Solution 2: I need to make virtual instance methods in the class and call them from the static callback functions. But I can’t make the function calls directly, I need a static instance of the class to make the instance function calls. (Re-read, as it’s a little complex) All I need to do is pass an instance of the class or subclass and I’ll be able to extend the functionality. It’s a little tricky/ugly, but it’s the best method I’ve found for encapsulating C style GLUT into a C++ application.

class AnimationFramework {
protected:
	static AnimationFramework *instance;
public:
	static void displayWrapper();
	static void runWrapper();
	static void keyboardWrapper(unsigned char key, int x, int y);
	static void keyboardUpWrapper(unsigned char key, int x, int y);
	static void specialKeyboardWrapper(int key, int x, int y);
	static void specialKeyboardUpWrapper(int key, int x, int y);
	void startFramework(int argc, char *argv[]);
 
	void run();
	virtual void display(float dTime);
	virtual void keyboard(unsigned char key, int x, int y);
	virtual void keyboardUp(unsigned char key, int x, int y);
	virtual void specialKeyboard(int key, int x, int y);
	virtual void specialKeyboardUp(int key, int x, int y);
};

The static methods are implemented here:

// Static functions which are passed to Glut function callbacks
 
void AnimationFramework::displayWrapper() {
	instance->displayFramework(); // calls display(float) with time delta
}
void AnimationFramework::runWrapper() {
	instance->run();
}
void AnimationFramework::keyboardWrapper(unsigned char key, int x, int y) {
	instance->keyboard(key,x,y);
}
void AnimationFramework::keyboardUpWrapper(unsigned char key, int x, int y) {
	instance->keyboardUp(key,x,y);
}
void AnimationFramework::specialKeyboardWrapper(int key, int x, int y) {
	instance->specialKeyboard(key,x,y);
}
void AnimationFramework::specialKeyboardUpWrapper(int key, int x, int y) {
	instance->specialKeyboardUp(key,x,y);
}

The startFramework() method is the same as provided in the Solution 1.

EDIT:
I left out details on how the instance was set. In my program, I subclassed AnimationFramework and created classes for each program I needed overriding the appropriate methods. As an example, KeyFrameFramework was a subclass in my project.

I have a function in my AnimationFramework.h

static void setInstance(AnimationFramework * framework);

AnimationFramework.cpp

AnimationFramework *AnimationFramework::instance = 0;
void AnimationFramework::setInstance(AnimationFramework *framework) {
	instance = framework;
}

The main.cpp looks something like this:

int main(int argc, char *argv[]) {
 
	AnimationFramework *f = new KeyFrameFramework(); // Subclass of AnimationFramework
	f->setInstance(f);
	f->setTitle("Key Framing:");
	f->setLookAt(0.0, 2.0, 10.0, 0.0, 2.0, 0.0, 0.0, 1.0, 0.0);
	f->startFramework(argc, argv);
	return 0;
}

I set the instance variable of the AnimationFramework class to an AnimationFramework subclass called KeyFrameFramework. Doing so allows me to use polymorphism and call the appropriate functionality that is specific to each animation project. Note: Don’t set the instance within a constructor, since the object is not fully initialized until the constructor is finished. You need to set the instance after your subclass object has been created.

Let me know if you have any questions. Below are the references I used.

References:

• http://www.parashift.com/c++-faq-lite/pointers-to-members.html#faq-33.5
• http://www.parashift.com/c++-faq-lite/ctors.html
• http://www.gamedev.net/community/forums/topic.asp?topic_id=314567

I wanted to set default values for my application using a Settings.bundle and I ran into an interesting issue with iPhone SDK 2.2. If you don’t run the Settings application before your application runs for the first time, then the default settings are not set. It turns out that you’ll need to manually set them on the first time the application is run.

Setting defaults to the default value…

I’m not sure why there isn’t a function that can set the values for me, but after digging I found someone who ran into the same dilema. I don’t want to have multiple locations with default values (code and a Settings.bundle), so I found a programmatic way to set all the default values if they haven’t been set.

- (void)registerDefaultsFromSettingsBundle {
    NSString *settingsBundle = [[NSBundle mainBundle] pathForResource:@"Settings" ofType:@"bundle"];
    if(!settingsBundle) {
        NSLog(@"Could not find Settings.bundle");
        return;
    }
 
    NSDictionary *settings = [NSDictionary dictionaryWithContentsOfFile:[settingsBundle stringByAppendingPathComponent:@"Root.plist"]];
    NSArray *preferences = [settings objectForKey:@"PreferenceSpecifiers"];
 
    NSMutableDictionary *defaultsToRegister = [[NSMutableDictionary alloc] initWithCapacity:[preferences count]];
    for(NSDictionary *prefSpecification in preferences) {
        NSString *key = [prefSpecification objectForKey:@"Key"];
        if(key) {
            [defaultsToRegister setObject:[prefSpecification objectForKey:@"DefaultValue"] forKey:key];
        }
    }
    [[NSUserDefaults standardUserDefaults] registerDefaults:defaultsToRegister];
    [defaultsToRegister release];
}

All you have to do is call the above function if one of your Standard User Defaults returns null. Make the call once your application finishes loading like so:

- (void)applicationDidFinishLaunching:(UIApplication *)application {
	// Get the application user default values
	NSUserDefaults *user = [NSUserDefaults standardUserDefaults];
	NSString *server = [user stringForKey:@"server_address"];
	if(!server) {
		// If the default value doesn't exist then we need to manually set them.
		[self registerDefaultsFromSettingsBundle];
		server = [[NSUserDefaults standardUserDefaults] stringForKey:@"server_address"];
	}
    [window addSubview:viewController.view];
    [window makeKeyAndVisible];
}

References:

• http://stackoverflow.com/questions/510216/can-you-make-the-settings-in-settings-bundle-default-even-if-you-dont-open-the-s
  
• Display code in your WordPress blog: http://wordpress.org/extend/plugins/wp-syntax/