By Design - Half-Life 2: Orange Box

You know what's inside The Orange Box. Now, meet the people behind it.


alf-Life 2: The Orange Box is a heck of a packing job. Combining one of the most essential and engrossing single-player campaigns of all time, stylish and unorthodox online content, and a totally unique space-bending puzzle, it asks one price for three elements lesser studios might've been happy to let stand on their own. It's quite an achievement, but it wouldn't have been possible without the members of three distinctly talented teams. So we thought we'd introduce them and see if they can tell us what went into The Orange Box, aside from so many great hours of gameplay.

Half-Life 2 - Programmer/Designer

David Speyrer: Programmer/Designer
David Speyrer: Programmer/Designer

David Speyrer

Age: 37
Hobbies: Rock climbing, home improvement
Likes: France, Team ICO
Dislikes: Artificial sweeteners
Favorite musician: Nick Cave
Favorite story: On Stranger Tides by Tim Powers


How did the episodic nature of the series affect your approach to design?

David Speyrer:

The smaller scope of the episodes relative to Half-Life 2 allowed us to choose a few key design elements to focus on and execute those elements really well. In Episode One the focus was on Alyx as a companion; in Episode Two we continued that focus and added the car, the Hunter, and new outdoor environments.

We're also able to sketch out the complete arc of an episode earlier in development than we can on a larger project like Half-Life 2, which means we're able to do more broad revision on the product as a whole. Reading people's feedback on the game, I think that people are picking up on that as they play Episode Two; there's a cohesion and polish to the game that we might not have achieved had it been a 20-hour game, because the cost of a top-to-bottom revision in a game of that size can be so high.


The Half-Life series is known for using scripted sequences in place of cutscenes to tell its stories. How did this play into the design?


One big factor with our interactive scripted sequences is that we can't guarantee that the player will look at them. We believe it's important to maintain this player freedom for a whole bunch of reasons, most notably to maintain the player's sense of immersion. As a result, we have to employ a bunch of tricks to get players to look where we want them to look. We also have to design the level in a way that encourages players to stick around and listen to the story.

In Episode Two you see two complementary approaches to storytelling. In some cases we tell the story "as you go," so while you're fighting or adventuring alongside another character they say things that flesh out the story or develop their character. Many of these story moments are important but optional, because some players will definitely not be in the right place at the right time to hear them. In other cases we lead the player into a place where they can witness a larger, more monolithic story event, and these typically are more vital to the main plot line of the episode.


Tell us about how you approached designing Half-Life 2 and the episodes.


We design in small teams we call "cabals" that consist of people from all disciplines of production such as programmers, level designers, modelers, and animators. Our core mandate is "he or she who designs it, builds it," which helps us maintain a high level of investment in our designs.

We start with a statement of the design goals for a small section of the game and try to keep the pure design phase very short, preferring an imperfect prototype to a perfect design. As soon as we have that imperfect prototype, we play-test it to see if it achieves its goals. If the experience is fun even in prototype form we know we're onto something. If our early play tests yield ideas for ways to improve the prototype, we try a few of those and then test again as soon as possible.

So the heart of our design process is very iterative and everyone from the cabal watches every play test to apply their expertise to the game experience.


Half-Life 2 was one of the first modern games to really establish the importance of in-game physics. Why are physics important to the Half-Life 2 series, and what do they mean for the way games will play in the future?


We went after physics in Half-Life 2 because it was a technology opportunity that was fundamentally interactive in nature--it wasn't like a new graphics feature that lets the same old game look better. It also dovetailed well with the immersion goals of the game because it met people's natural expectations of how a real world should behave. Physics was such fertile ground for building gameplay that as we realized its potential it became more and more the defining element of the game.

The gravity gun was a big facilitator of this and once it came together a bunch of other pieces fell into place: the physics puzzles, the saw blades, throwing explosive barrels, all the way down to deeper mechanics like grabbing live grenades with the gravity gun and throwing them back. In the future when we start simulating things like fluid dynamics in real time we'll see more interesting mechanics emerge. I also think that over time the line between the parts of the world that are physically simulated and the parts that are not will blur and our gameworlds will be more and more interactive and intuitive.

Half-Life 2 - Lead Writer

Marc Laidlaw: Lead Writer
Marc Laidlaw: Lead Writer

Marc Laidlaw

Age: 47
Hobbies: Playing games
Likes: Miyamoto
Dislikes: Star Wars
Favorite band: Camper Van Beethoven
Favorite story: Down There by Damon Knight


In your opinion, what distinguishes storytelling in a video game from any other kind of writing?

Marc Laidlaw:

In a written story, the story is the entire reason for the endeavor. It's everything. In a game, the story is just one facet of the experience. It could be an incredibly minor part of the game, or completely nonexistent. Even in Half-Life, where we have committed to putting a lot of our emphasis on storytelling, it's still only one aspect of a complex experience.

My original goal in joining Valve was to put the storytelling on equal footing with the technology--to come up with narrative techniques that were as cool, in their way, as the monster artificial intelligence or cleverly designed combat sequences. We experiment with narrative the same way we experiment with rendering technology, keeping in mind that these are games--not books, not movies. Games.


Tell us about how you approached writing Half-Life 2 and the episodes. What was the most difficult part? The most fun?


We approach everything by developing a shared vision. We decide what challenges we're going to bite off for ourselves. The hardest part is figuring out the blend of active gameplay and dramatic staging--to make sure that the game never bogs down in exposition, but that we take the time to let the story breathe and develop naturally. The most fun is taking the script into a recording studio with good actors, and seeing the written scenes come to life in the game after our animators have worked their magic.


In most games you're saving the girl. In Half-Life 2, she tends to be by your side. Why'd you take that approach?


In Half-Life 2, we chose to shift the emphasis from a solitary "you alone must save the world" approach, to an experience that placed more value on cooperation, on allies and friends. Gordon's goals are really Alyx's goals: You spend much of the game trying to rescue her father, knowing that Eli is a figurehead for the resistance. So Alyx is our way of embodying the player's goals, making them desirable and human and real--much like Alyx herself.


Why is romance so rare in games, and what role does it play in Half-Life 2?


It's rare for characters in games to have anything like a real interior life, or realistic hopes and fears. When it's all about saving the world, there's not much time to develop relationships. Characters tend not to have families, friends, or much of anything else the player can relate to. They have goals and tasks that need doing. Romance doesn't really thrive in goal-driven environments, unless perhaps it's a game like Chulips, where kissin' is the goal. We have tried to ground our characters in relationships that are familiar to most people: Alyx has a father, her hopes and fears are mostly right out there for you to see, and she tends to treat Gordon like an increasingly close friend. Still, so far we merely, ah, flirt with romance.

Have thoughts on The Orange Box or the people behind it? The world awaits your perspective!

Team Fortress 2 - Class Designer

Robin Walker: Class Designer
Robin Walker: Class Designer

Robin Walker

Age: 31
Hobbies: Playing games. Reading. Soccer.
Likes: Sci-fi / Post-apocalyptic settings. Squad-based tactical combat games.
Dislikes: All fantasy novels written in the last 20 years.
Quirks: Pathological liar.
Favorite musician: Bob Dylan.
Favorite story: Anything by Jack Vance.


How did you go about trying to balance the game among so many classes?

Robin Walker:

Balancing is an iterative process. We watch many, many play tests, and use automatic systems to gather all the underlying data. We spot balance problems, use the data to come up with solutions, implement the fittest ones, and then run another play test.

One of the most important pieces to class balance is the weaknesses of the classes. We've found that it's much easier to balance a class if it has a significant weakness to counter out its strength. For example, tuning the Scout's combat strength is easier when we know that an enemy team can cause him a lot of trouble through the use of sentry guns. The Soldier, on the other hand, is one of the hardest balancing acts, because he has no particular weakness.


Were any classes left on the cutting room floor? Can you tell us about them?


Internally, we tend to talk about the roles that a class fulfills. So while we didn't have a big list of classes that failed, we do have a variety of roles that did. Some classes, like the Pyro and the Scout, went through several roles before we found the one we were happiest with. For an example of a failed role, take a look at the TFC Scout. It's a role characterized by its desire to avoid combat. TF's gameplay revolves around combat, so that role is already off to a bad start when it's focused on avoiding the game's core.


What makes a good class?


A unique role, relative to the other classes, which could be a unique purpose, or a unique method of execution. Weapons and abilities that allow the class to achieve that role. A core weakness that makes him fear some other classes and situations. Weaknesses in his weapons and abilities that add interesting factors to the decisions he needs to make throughout the game.


Tell us about how you approached the design of Team Fortress 2. What was the most difficult part? The most fun?


The most difficult is definitely the class balancing. It's a long-term problem, and one that's never really finished, because we plan to keep adding new features to the game postlaunch. The most fun was in some of the smaller features that used our art direction to solve design problems in amusing ways. For example, both the Spy's paper masks and the gib callout arrows in the deathcam were a lot of fun to put in and iterate on.

Team Fortress 2 - Artist

Charles Brown: Artist
Charles Brown: Artist

Charles Brown

Official Job Title: Programmer/Designer
Age: 35
Hobbies: Cooking, reading, programming, video and board games
Likes: Co-op and multiplayer games
Dislikes: Griefers/Cheaters
Quirks: I own a Charlie Brown bowling ball
Favorite musician: Beatles
Favorite story: Demon Princes by Jack Vance


Team Fortress 2 has a unique visual style that clearly isn't trying to be photorealistic. Why take this approach?

Charles Brown:

Team Fortress 2's gameplay drove the art direction. The weapons in TF 2 are diverse and their effects on players and world are exaggerated. Humorous scenarios often arise while playing TF 2. So instead of combating the gameplay with a realistic look we decided to complement it with a stylized art direction.


Where did the inspiration for the game's look and feel come from?


We were inspired by the likes of J.C. Leyendecker, Dean Cornwell, and Normal Rockwell. Their early-to-mid 20th-century commercial illustration style complemented our 1960s industrial design; [it] made up the building blocks of our characters and world. In addition, we were also influenced by Hayao Miyazaki's art style, in particular his broad brush strokes and color palette choices.


Do you feel realism in games is overrated? What are the advantages, or disadvantages, of a more cartoonlike visual style?


The art style should solve design problems for the game. Thinking of the art in this way, as a tool rather than just a subjective choice, means certain game designs are better served by a realistic look while others benefit from stylization.


: Tell us about how you approached the art design of Team Fortress 2. What was the most difficult part? The most fun?


The most important element in any multiplayer game is the players themselves, and as such, the characters were our primary focus. The goal was to design aspirational characters whose abilities needed to be easily understood at a glance, while not being cliché.

Additional gameplay and presentation constraints made this a difficult but rewarding task. Some art problems were solved with technology. For example, the Phong and rim-lighting shader, when applied to the character models, helps them "pop" out of their environment. Other problems were solved in design. When a character comes into view a player must quickly be able to identify their team, class, and intention. The led to the design of a character read hierarchy: Team identification happens through our use of warm and cool colors, class readability via designed distinctive silhouette shapes, and intention through weapons at the ready. The combination of design choices and technology helped us create characters and a world that were both functional and fun.

Have thoughts on The Orange Box or the people behind it? Enlighten us with your wisdom!

Portal - Level Designer

Kim Swift: Level Designer
Kim Swift: Level Designer

Kim Swift

Age: 24
Hobbies: Playing video games, obviously, but beyond that any random hobby I can pick up; currently I'm sewing odd stuffed animals
Favorite band: Bad Religion
Favorite story: Hitchhiker's Guide to the Galaxy by Douglas Adams
Likes: Team Fortress 2 and puppies
Dislikes: Trying to figure out my dislikes
Quirks: Doesn't working in the game industry require you to be at least mildly quirky?


We understand that Portal's roots date back to a computer science project created by a group of students. How did it become what it is and make its way into The Orange Box?

Kim Swift:

Two years ago, everyone on the Portal team (with the exception of our writer Erik Wolpaw) were students at DigiPen Institute of Technology. One of the course requirements is that each year you have to create a game from scratch for our game class. For our senior-year project, we made a game called Narbacular Drop, which is the predecessor of Portal.

Every year, DigiPen holds a job fair for graduating students where they gather all sorts of developers from across the country to come in and take a look at the seniors' projects and resumes. A couple people from Valve swung by and took a look at Narbacular Drop and invited us in to demo the game for Gabe Newell. After 15 minutes or so, Gabe stops us and asks, "What are you guys doing after you graduate?" and offers the entire team a job on the spot to make a game based on Narbacular Drop using the Source engine.

After suffering from massive heart attacks, we all jumped at the opportunity and started working on Portal. Two years later and we have the utmost honor to be packaged in The Orange Box with two of the best games of the year.


The game gradually ramps up in difficulty as players become more familiar with the gameplay. Tell us about how you managed the progression of difficulty without making the game frustrating.


Play testing, play testing, play testing. Every week during development, we would bring in someone new to play the game from start to finish. We wouldn't just have them fill out a questionnaire at the end of the game, but we'd watch them play through. There is so much value watching someone play the game. It's incredibly obvious when someone is stuck, or is having fun. We let our play tests dictate how to teach our players and when to ramp up the difficulty.


The game has a very satirical sense of humor. Where did that come from?


We all seem to have a pretty quirky sense of humor on the Portal team, and part of that humor has seemed to leak into the core of the game. The other part is our awesome writer, Erik Wolpaw, who wrote the monologues for the game's ever-present disembodied voice, GLaDOS.

When we first started working on the game, we didn't really have a story or any other characters besides the player. Our feedback was that the game was pretty lonely, so we talked to Erik and he agreed to help us fill in the gaps. It was really great to see that our players found the dialogue to be rewarding, and many people would completely stop what they were doing to laugh at GLaDOS's comments.


Tell us about how you approached Portal's design. What was the most difficult part? The most fun?


Designing Portal was a very collaborative and iterative process. For instance, to create a map, the entire team would get together in a conference room with a whiteboard and sketch out which gameplay ideas we'd want to use. One of us would quickly get the map up and running in a few days, and then we'd get someone who had never seen it before to play through. Based on our observations from the play test, we'd readdress the map, make any alterations that everyone agrees on, and run another play tester through. Rinse and repeat!

One of the most fun things about designing Portal was watching people play through a level in a completely new way that we hadn't thought of. Seeing this in many cases gave us the inspiration for a new puzzle or mechanic.

Portal - Programmer

Jeep Barnett: Programmer
Jeep Barnett: Programmer

Jeep Barnett

Age: 25
Hobbies: Playing games, making music, cooking
Likes: Satire, mental challenges
Dislikes: Whiners
Quirks: Shave my own head once a year, no phone (cell or home)
Favorite musician: Squarepusher
Favorite story: Socks for Supper


Where did the idea for Portal, or its original version, Narbacular Drop, come from?

Jeep Barnett:

Our team at DigiPen was required to create a game as a class project each year. We spent several months brainstorming ideas for our senior game, and the seamless portal graphical effect was an element in a few of those concepts. Narbacular Drop was a combination of those ideas stripped to the core portal mechanics. It ended up being a proof of concept for the portal gameplay ideas we developed in Portal.


What were the unique challenges of programming the game?


Any new element added to the game was expected to interact with portals and that often meant plumbing a much larger system of the Source engine. Everything from rendering, to audio, to AI was modified to accommodate portals.

The biggest undertaking was modifying the physics engine. We had to allow the player to dynamically cut holes in the static environment and process those collision changes in real time. Objects on opposite sides of a portal needed to realistically collide with each other. This code was rewritten several times for improved precision and performance.

Then there's the really weird stuff. In a standard Euclidian space you can ignore problems such as, "What if an object collides with itself?" In Portal, the shortest distance between two points isn't always a straight line. If the player's feet pass through a floor portal and touch a wall on the other side, they need to be able to stand on that wall as if it were a floor. For those challenges we first form an expectation for how a portal would behave in the real world, and then we assimilated that behavior into the game.


In Portal, it's possible to enter a room while leaving it. How were you able to wrap your mind around your own game?


I find it pretty amazing how well the human brain can adapt to reason within new rule sets. For example, a chess player sees beyond pieces on a grid to perceive the paths that each piece can travel. Likewise, anyone should be able to naturally navigate a room using portals after a little hands-on experience with Portal.

Before we had Portal or even Narbacular Drop for that hands-on exposure, we created a simple 2D portal demo. Playing with that allowed us to come to terms with the more advanced portal concepts, such as how momentum carries through portals.

So you've met the people behind Half-Life 2's incredible gameplay and engrossing story, the ones who made Team Fortress 2 look so cool and feel so different, and a couple of the minds behind Portal's weird science and unique wit. As you can see, Half-Life 2: Orange Box doesn't just represent a stunning collection of unusual games, it also represents an interesting assortment of professional gamers, people just like you. Thanks for reading, and by the way, if something is iterative, it involves repetition. To use that in a sentence, here's hoping Valve's success with The Orange Box is as iterative as its production.

Have thoughts on The Orange Box or the people behind it? Want to hear from more developers? Let us know!

GameSpot may get a commission from retail offers.

The products discussed here were independently chosen by our editors. GameSpot may get a share of the revenue if you buy anything featured on our site.

Got a news tip or want to contact us directly? Email

Join the conversation
There are 238 comments about this story