Monthly Archives: May 2013

Papers Accepted to IEEE Visual Languages/Human-Centric Computing (VL/HCC)

Good news! We received notification today about two papers accepted to VL/HCC later this year. Here are the paper titles and abstracts. When the camera-ready preprints are ready, I’ll be sure to post those as well.

Helping End Users Help Themselves with Idea Gardening

J. Cao, I. Kwan, F. Bahmani, M. Burnett, J. Jordahl, A. Horvath, S. Fleming and S. Yang. End-User Programmers in Trouble: Can the Idea Garden Help Them to Help Themselves? to appear in the IEEE Conference on Visual Languages and Human-Centric Computing (VL/HCC), San Jose, USA, 2013

Abstract—End user programmers often get stuck because they do not know how to overcome their barriers. We have previously presented an approach called the Idea Garden, which makes minimalist, on-demand problem-solving support available to end user programmers in trouble. Its goal is to encourage end users to help themselves learn how to overcome programming difficulties as they encounter them. In this paper, we investigate whether the Idea Garden approach helps end-user programmers problem-solve their programs on their own. We ran a statistical experiment with 123 end-user programmers. The experiment’s results showed that, even when the Idea Garden was no longer available, participants with little knowledge of programming who previously used the Idea Garden were able to produce higher-quality programs than those who had not used the Idea Garden.

Keywords—Idea Garden; end-user programming; problem solving; barriers; mashups; quantitative empirical evaluation

User Interface Explanations in Intelligent Agents

T. Kulesza, S. Stumpf, M. Burnett, S. Yang, I. Kwan and W.-K. Wong. Too Much, Too Little, or Just Right? Ways Explanations Impact End Users’ Mental Models, to appear in the IEEE Conference on Visual Languages and Human-Centric Computing (VL/HCC), San Jose, USA, 2013

Abstract—Research is emerging on how end users can correct mistakes their intelligent agents make, but before users can correctly “debug” an intelligent agent, they need some degree of understanding of how it works. In this paper we consider ways intelligent agents should explain themselves to end users, especially focusing on how the soundness and completeness of the explanations impacts the fidelity of end users’ mental models. Our findings suggest that completeness is more important than soundness: increasing completeness via certain information types helped participants’ mental models and, surprisingly, their perception of the cost/benefit tradeoff of attending to the explanations. We also found that oversimplification, as per many commercial agents, can be a problem: when soundness was very low, participants experienced more mental demand and lost trust in the explanations, thereby reducing the likelihood that users will pay attention to such explanations at all.

Keywords—mental models; explanations; end-user debugging; recommender systems; intelligent agents

Advertisement

A Brief Ludum Dare 26 Update: Results

The final results are in for Ludum Dare 26, a game competition in which participants have 48 hours to program a game from scratch. As I previously wrote, I participated in Ludum Dare for the first time last month and learned an immense amount in an extremely short amount of time. Here are the ratings I received for Painter’s Cat. The ratings were out of 5, and there were 1610 games submitted to the Compo category (which I participated in).

Ratings

Coolness 60%
#47 Audio 3.85
#176 Innovation 3.64
#187 Humor 2.96
#219 Mood 3.33
#265 Theme 3.87
#457 Overall 3.26
#569 Graphics 3.00
#700 Fun 2.82

(For those who are wondering, Coolness is a measure of how involved you are in the community, especially with playing, rating, and leaving comments on other people’s games. The more you play and comment and rate other games, the higher your coolness is).

First of all… I’m in the Top 100 for Audio at #47? Holy crap. That’s amazing. I’m pretty blown away by this – all of that time doing One Hour Compos has really paid off! I couldn’t find where my overall ranking was, but it looks like I managed to place in the top 40%, which isn’t bad for a first game.

Ludum Dare was an amazing positive experience for me overall. I’m pretty excited for next time. There’s still a lot for me to learn and this is one way to help me not only learn about software development and game development, but to also have fun doing it.

Our paper, “The Role of Domain Knowledge and Cross-Functional Communication in Socio-Technical Coordination”, to be presented this coming week at ICSE2013!

So, our paper at the International Conference on Software Engineering, titled The Role of Domain Knowledge and Cross-Functional Communication in Socio-Technical Coordination, will be presented this coming week in San Francisco. Daniela is going to be presenting this paper on Thursday, May 23 at 1:30 PM in Grand Ballroom B!

The preprint of this paper appears on my blog. The main story is that we examine how diverse roles in two teams in Brazil working on requirements and their related artifacts coordinated along task dependencies using a case study method, and report on how knowledge and work dependencies affect their work.

There are a number of other great papers that are appearing in the same session, including two co-authored by Prem Devanbu. It’s a good session to be at, in my opinion.

Hope to see you there!

TextMate and Running a Master File

Many people find my blog searching about how to make their multi-file LaTeX projects build in TextMate.

It’s worth nothing that the TM_MASTER_FILE works for almost any programming project, not just LaTeX. So, if you’re working on a Python or a Ruby project and have a main file that you want to run, set the TM_PROJECT_MASTER to point to that and save yourself some effort. Here’s a reminder on how to do that.

  1. Open your project file (or create one if you don’t have one).
  2. De-select any selected files in the TextMate drawer by clicking on empty space.
  3. Click on the “i” in the bottom-right corner of the drawer.
  4. Add a variable named TM_PROJECT_MASTER.
  5. Set the value to the name of your main file (in my Ludum Dare game, it was painterscat.py

Now your “run” ⌘r shortcut will always run your main file with minimal hassle!

Inheritance in Javascript: Getting it Done from a Newcomer’s Perspective

The Idea Garden is made up of a number of suggestions that each follow a certain structure. The person who wants to use the Idea Garden in a new environment has to write templates and a bit of boilerplate code. As part of the Idea Garden meets Gidget project, I’ve been doing some work porting the Idea Garden architecture from a Firefox plugin, CoScripter.

The purpose of this post is to discuss how I used a particular style of Javascript prototype inheritance that I discovered on Stack Overflow. In some respects, it’s trying to communicate to new developers from the perspective of a new Javascript developer and explain some gotchas on the way. This isn’t meant to be a really in-depth guide to Javascript prototypes or the Javascript object model – there’s some good reading on the Internet already that discusses the intricacies of the language.

The Prototype-based Programming Model

Javascript’s object-oriented model is prototype-based rather than class-based: what this means, practically, is that unlike most classical object-oriented languages (Java, Smalltalk, C++, Python, Ruby, etc.), there is no distinction between a class definition and an object. In other words, as soon as you “declare” a “class” in Javascript, it exists in the world immediately!

This has led to some oddities in Javascript, such as there being at least two ways to declare a new class, with each of these ways looking a little different from each other. In some respects, trying to think in terms of normal object-oriented programming in Javascript only gets you mired in some unnatural programming practices.

Because Javascript isn’t classicly object-oriented, trying to comprehend its inheritance model is a bit tricky. For this reason, along with reasons outlined in the Gang of Four book), and others, I’d actually recommend composition over inheritance.

But, in this particular case, I was asked to implement an architecture to use inheritance because we have a number of Idea Garden suggestions, and Idea Garden suggestions behave similarly and borrow a number of common functions, but might have some small specialized differences in their content.

Which Inheritance Style, Exactly?

There is no shortage of Javascript inheritance styles. Because there’s really no accepted “way” of doing it (or of NOT doing it), anyone who wants to do it has to really work at studying what they want, exactly.

In the end, I emulated a style following this Stack Overflow post by Sean McMillan because it (1) used the Javascript Module pattern, which was pervasive throughout this code and (2) used Object.create, which basically does the inheritance for me (Object.create is ECMAScript 5 and from what I understand, basically replaces new). However, the answer, while providing a pretty good code example, didn’t explain all of the concepts behind it that I needed to know.

The Return of the Prototype

As I mentioned above, Javascript is a protoype-based language, and as such doesn’t have class definitions. The module pattern probably comes the closest to emulating a “class definition”, but is still not exactly a class definition. If I create a module:

IdeaGarden = (function() {
	var exports = {},
		environmentName = "Gidget";

	exports.hostInfoToIdeagarden = function() {
		// Do stuff here
	};

	return exports;
}());

it exists as soon as the code runs, and I can modify the object’s state, use its functions, and so forth, without doing an Object.create or new call.

An important part of inheritance is to separate the definition from the instantiation. In general, you want to inherit from a certain state of class, rather than whatever you happened to have used and extended most recently.

To do this, we define a prototype in our module that we will inherit from when we do an Object.create.

Vehicle = {};

Vehicle.AbstractVehicle = (function() {
	var exports = {};
	exports.prototype = {};
	exports.prototype.init = function() {
		this.identifier = "Abstract";
		this.passengers = 0;
		this.speed = 0;
		this.running = false;
	};

	exports.prototype.getName = function() {
		return this.identifier;
	};

	exports.prototype.accelerate = function() {
		if (this.running) {
			this.speed = this.speed + 10 - this.passengers;
		}
	};

	exports.create = function(identifier) {
		var ret = Object.create(exports.prototype);
		ret.init();
		return ret;
	};

	return exports;

}());

This code, which pretty much conceptually matches to Sean’s code in his post, illustrates a few important points.

First, we have a callable “exports.create”, which creates a prototype of the base class, runs its “init” function, and returns the prototype for us to use. The fact that it returns a prototype is important! Unlike in the normal module pattern, where you would declare additional private members somewhere in the body of Vehicle, in the inherited class from the prototype, only members declared using this in the init function will be visible to instantiated classes and their descendants.

The difference between “what belongs to the module” and “what belongs to the new created class” isn’t exactly difficult, but it tripped me up as a beginner because of the close association between the module class having its own private members and the members appearing in the init function.

Because the create function returns a prototype, everything needs to exist in the scope of the prototype, rather than in the module!

The corrilary to this is that public members need to exist in the init function and that public methods need to be part of the prototype, rather than the module. Here’s an example method:

exports.prototype.accelerate = function() {
		if (this.running) {
			this.speed = this.speed + 10 - this.passengers;
		}
	};

Inheriting from the Base class

Once you realise the importance of returning the prototype, the rest of it pretty much falls into place. Since you’re using Object.create to create a new version of the prototype, you can create multiple versions of the base class and modify them freely without worrying about changing the definitions(that’s something that you would have to worry about otherwise). Here’s some code to inherit from the base class.

Vehicle.Car = (function() {
	var exports = {};
	exports.prototype = Object.create(Vehicle.AbstractVehicle.prototype);
	exports.prototype.init = function(identifier) {
		Vehicle.AbstractVehicle.prototype.init.apply(this, [identifier]);
		this.identifier = identifier;
		this.passengers = 4;
	};

	exports.prototype.cruiseControl = function(newSpeed) {
		while (this.speed < newSpeed) {
			this.speed += 1;
		}
	};

	exports.prototype.startVehicle = function() {
		console.log("Starting the I4");
		this.running = true;
	};

	exports.create = function(identifier) {
		var ret = Object.create(exports.prototype);
		ret.init(identifier);
		return ret;
	};

	return exports;
}());

Let’s go through some of this code as well.

First, you’ll notice that exports.prototype in this version does an Object.create(Vehicle.AbstractVehicle.prototype). This basically matches the prototype to the base class. In the exports.prototype.init function, we call the init from the base class as well – it’s essentially equivalent to our “super” call from other languages. Note that we pass in our arguments (identifier in this case) in an array – that’s just how apply works. Our variables from the base class will be initialized at that time, along with any additional variables we want here.

As mentioned above, because we return the prototype, every method that we want to extend the base class with needs to be preceeded with prototype. That’s why we have exports.prototype.cruiseControl and exports.prototype.startVehicle. Note that we can use any of the variables and methods we declared in the base class within these methods as long as we precede them with this.

Finally, we have our create function, which is NOT in the prototype space because we want to be able to call it from our code. This is pretty much boiler-plate code from the base class but we need it.

Now, we can use this just as we might expect (with a little help from JQuery assigning an identifier to outputDOM):

var workCar = Vehicle.Car.create("Work");
workCar.startVehicle(); // Outputs "Starting the I4 of Work" to the console.
outputDOM.append(workCar.getName() + ". Speed: " + workCar.speed + "
"); // Outputs "Work. Speed: 0". workCar.accelerate(); outputDOM.append(workCar.getName() + ". Speed: " + workCar.speed + "
"); // Outputs "Work. Speed: 6"

If we create another car using Vehicle.Car.create it’ll hold its own set of variables.

In Closing

I find Javascript is a little bit of an adventure because it’s a language that appears to have evolved to do things that no one envisioned it to do fifteen years ago. It used to be written directly in HTML as part of “onclick” attributes, and now it’s manipulating the DOM, applying CSS elements, and animating on-screen. Everything’s evolved heavily in the back-end as well. Douglas Crockford attempted to apply a classical inheritance model to Javascript but then gave up, eventually saying that “it was a mistake”. However, a lot of this information is still on the web, and it’s difficult to identify what the state-of-the-practice is for Javascript these days.

I do want to, however, keep up to date and be sure that if I’m writing Javascript now, it’ll be recognizable to people who are reading my code two, three, five, ten years from now. I like trying to keep code reasonably simple and clean – especially in a situation like the one I’m in where others are going to be working off of the code that I started.

If anyone notices that I’ve made a mistake in my understanding, or would like to recommend other ways to accomplish what I’ve outlined here, I’d be glad to hear about it and update this post accordingly – I’m not a Javascript expery by any means, but by writing and self-reflecting on what I’ve done, I would like to keep improving my skills and my understanding. So, if you have any suggestions or improvements please feel free to drop me a line or leave a comment.

The Challenges and Rewards of Non-Competitive “Compos”

As I mentioned in my previous posts, I participated in Ludum Dare, a game development “compo”, where you build a video game by yourself, from scratch, in 48 hours (a variation gives 72 hours and a team, but I did the 48-hour version).

A “compo” is a “composition competition”, but I’ve yet to participate in a compo where the competition aspect is what actually gets people energized. If anything, the compo is more community-oriented than it is competition-oriented. I participate weekly at ThaSauce.net One Hour Compo, which is a music compo in which you create a song from scratch in one hour. The competition aspect supposedly is because people vote on their favourites at the end, but in the end I don’t think the votes are what anyone’s really fighting for.

In any case, I feel that compo has probably been one of the top ways for me to improve my music making skills and that doing the Ludum Dare compo was an excellent way for me to simply program for the sheer joy of it.

But, I think one of the greatest benefits of doing these compos is simply being able to succeed, and to feel happy and proud doing it. It’s a real self-esteem booster and it also helps you beef up your skills and the ability for you to work under huge time pressure.

Below I’m going to present the post that I wrote for Ludum Dare 26. Most people there tend to write technical post-mortems, but I thought that the emotional barrier was actually a bigger barrier to cross than the technical ones!

An earlier version of this post below first appeared on my Ludum Dare 26 blog.

I think I’m over my Ludum Dare Adrenaline Rush now.

In my first initial post, I said that I was going to fake Ludum Dare. To my surprise, a few commenters actually wrote in and encouraged me to just enter anyway. I did. And I came out with a game that had some interesting ideas, along with a number of problems.

I’m actually happy I participated and would like to thank the handful of people who encouraged me (some strangers, some friends) to enter anyway. It was a great experience and I want to be around next time, if I have time for it.

But, what led me to go down a road of, “I don’t think I can do it?”

I’ve never designed a game before

Well, technically I did. In my first ever C programming class when I was 15, the final project was a video game. Mine compiled but didn’t work – we had Macs at school, and when I realised that my game wasn’t getting close to finished, I brought my game home, wrote almost all of my code on my PC, and hand-checked it to see if it would work, in theory, when I brought into school the next day.

With some work I made it compile, but it didn’t really work as I wanted it to.

In some respect though, it doesn’t take much to be a game. I’ve seen a lot of things that people recognize as games. Top-down shooters, side-scrollers, role-playing games, adventure games, text-based games, board games, and so forth. But I’ve also seen a lot of creative work as well. A game where the main idea was to walk in a city. A game where  you woke up, experienced a main character’s morning, and looked at all of the objects in his or her room. A game where you planted seeds in a garden and watched them grow. They’re barely games by the standard definition – but they are all welcome in Ludum Dare.

Even if you don’t have a strong idea of what you want to do, build a game archetype anyway and then see what comes out. In some respect, one of the thrills of doing a game in 48 hours is that the first few ideas you get, you have to stick with because there’s no time to really make it better. So you get all of these raw, unrefined concepts that are the pure essence of creativity. And it’s great to see so many of these concepts work.

My Programming Chops

I don’t program a lot. I have a computer science degree so I know how to program, but my work is primarily focused on research activities that don’t require any development. I find a lot of programming boring – especially mathematics riddles. “Compute the least-common denominator of two numbers?” Snooze. “Write an algorithm that will identify is a string is a palindrome?” Ugh.

But then I start building a game and suddenly, everything is fun, even when I groan at thinking about the geometry and trigonometry. It’s because those things suddenly aren’t just mathematics. They’re situated in my game as a core concept now. I don’t need to understand them for the sake of knowing them – I’m understanding them because I know that they’re useful to me, now.

I learned a lot on the fly, and suddenly I realized that programming isn’t just about what I know – but about how fast I can learn what I need to know. I didn’t know PyGame existed until the morning I decided to do Ludum Dare. I didn’t know how to blit a sprite to a screen before reading about it on my lunch break. I had never thought about sprite rectangles, mouse movements and controls, or drawing tiled backgrounds until the hour the competition started – so I learned those things with a lot of help from the Internet. I can’t say that I know all of those things well, even now… but I’ve done them before now and I can only improve from here.

So even if you’re not a hot-shot programmer, it’s not just about how well you program – it’s about how well you can get what you need done by learning what you need to learn.

Ludum Dare and self-esteem

I think a lot of people who post on this site have a lot of confidence – you have to to enter something that is billed (albiet weakly) as a competition. The games that get all of the press are the ones that have the shiniest graphics, the best lighting engines, the cutest artwork, the most thrilling sound. In the end, history remembers the winners, and all of the winners kind of blob up together into this gigantic “super-winner” amoeba where it feels like one guy participated in 30 Ludum Dares and came up with a hundred amazing games along with a procedural level generator and a memory-management allocation system in the span of a week. The secret though is that this mythical superhuman game programmer doesn’t exist. That programmer is really hundreds of individuals

I think one big lesson to learn from this is that very few of us are superhuman, and more importantly, the majority of people who participate in a Ludum Dare are normal people. They’re not all rock-star programmers, hotshot artists, or amazing musicians. They’re regular people and normal people. Just like you, just like me.

The second point to remember is that Ludum Dare isn’t really a competition. Sure, it’s about getting votes and comments and people get into the top list at the end, but in the end there’s no reward and there’s very few bragging rights. This really isn’t a competition.

I think a big outcome of these two points is that Ludum Dare is a showcase. It’s not just “how good your game is”. It’s the fact that you’ve managed to produce a game at all. No where else would I have been able to build almost any program in 48 hours and then, in the span of under a week (so far!) convince 50 people to play and download my game and leave constructive comments.

In that sense, everyone’s a winner. And even though I said above that almost everyone in Ludum Dare is a “normal person”, the fact that we’ve programmed something from absolutely nothing to a working, deliverable product in 48 hours (or in a team in 72 hours) makes every one of us extraordinary.

I hadn’t been as excited about something before as I had right after Ludum Dare – and I think that’s because that, as soon as I had finished, I realized that I had done something extraordinary. A few thousand of us, together, had each accomplished something to be proud of.