3:36pm PT: After waiting in what can only be described as an epic line (by far the longest we've seen at GDC so far, and it's Friday afternoon!), we've found a respectable spot at the Portal postmortem, one of the most anticipated sessions of the entire week. Erik Wolpaw and Kim Swift are on stage, seemingly unaware of the sea of humans piling up outside the door.

3:53pm PT:
Ludwig posted a photo of the Portal postmortem line. Best part: you still can't see most of the line which wrapped around both ends of the frame. We're just happy to be in here. We're prickly with anticipation. It's like seeing the Beatles on Ed Sullivan ... just really nerdy ...

4:01pm PT: And we're off. They're here to talk about "integrating narrative and design" – first, a spoiler alert! They'll be talking about the game so, if you haven't played it, shame on you. Oh, and you may want to leave if that sort of thing bothers you. Why should you care about Portal? 1) "We had a small team" 2) Both a critical and commercial success for Valve 3) -- That small team size "imposed constraints" on their design choices. Example: Glados being a disembodied voice may not have happened if they had more money for animators, etc. We're thinking Jaws' and the point-of-view shark.

Wolpaw is talking about games with a "high story delta," unlike Portal. He mentions "Clive Barker's Undying" as an example that he actually likes. He addresses the gap between cinematics and gameplay - "a good game being undermined by the story the gameplay is telling and the story the story is telling. He says, "'Story' story must never intrude on 'Gameplay' story."

4:08pm PT: They're emphasizing playtesting, encouraging developers to watch playtests personally -- "exposes what isn't working." If a playtester can't recall the story, then it isn't working. Using the example of the player's opening "cage" they used playtesting to refine art direction as well.

Wolpaw says writing a funny game is a difficult task. "About three years ago I worked on a game called Psychonauts" -- the crows begins clapping. "More people are clapping than bought it." Wolpaw's slides are hysterical -- this one has a topless character with large "Don't" and "Despair" badges in the important parts.

4:14pm PT: They originally planned on having you meet "The Rat Man" -- the character responsible for all the wall scribblings found throughout the game. Because they only had two artists, they instead just realized him through the hidden rooms.

They're explaining the inception of the WCC: the plan was for a "Box Marathon" level -- long level with a box that needed to be carried through the entire level. Playtesters hated it, so they revised the functionality of the box so they wouldn't hate it so much, but they were still forgetting about it. "Still needed something else ..."

Voila! "Erik to the rescue ... when all else fails great dialogue is an excellent hint." Wolpaw says he read interrogation manuals that had psychological information on detainees to get the idea for humanizing the cube. It worked! People stopped forgetting their boxes.

4:19pm PT: Players needed to be trained to use the incinerator to defeat the boss, so they decided to add an incinerator to the WCC level and force the player to euthanize the cube, instead of GLaDOS saying it was going to be done. "You get to incinerate GLaDOS the same way she made you incinerate your best friend."

First boss battle idea: "James Bond lasers." Swift says, "The lasers were extremely boring to dodge." They were also difficult to aim, so they rested on a rocket launcher.

Second attempt: "Portal Kombat." Hardcore FPS players wanted more action and players never appreciated the story, while others didn't understand the change from the more cerebral format they were used to.
4:26pm PT: Final boss battle idea was a chase scene where the players had to chase GLaDOS. The pacing was bad, and players were confused. It was too complex. The more complex they made the battles, it slowed players down and ruined the pacing. They were screwed -- had to get an ending so they could ship on time with Episode 2 and TF2. Wolpaw makes a joke about how humble their game was compared to those other ones.

Testers really liked the Fire Pit puzzle, even though it was one of the simplest puzzles in the game: time pressure, visual impact, high drama, all wrapped around an easy puzzle.
They realized they didn't need a complex puzzle at the end, they just needed to add those elements to the battle. They added a timer which counted down to a release of neurotoxin (they insist this was because the green particle effect was cheap). They also used the geometry of the room to help players identify the

4:34pm PT: The boss battle was so important because they really wanted players to feel "genuinely happy and leave the game with a smile on their face." The answer: a song, in this case "Stay Alive" by Jonathan Coulton. "The cost to happiness ratio of putting a song over some scrolling text was really high."

Wolpaw says, "Embrace your constraints as fuel for creativity." Oh, and Swift says, "playtest, playtest, playtest. Onto Q&A time!

4:45pm ET: Q) Why does GLaDOS have an incinerator in her room? A) "The answer is actually kind of fun." They made the room so that if she lost her mind, they could incinerate her. "It was bad of Aperture Science. THey're sort of half-geniuses and half-morons."

Q) How did they get new playtesters to test their game? A) Evidently it's not very hard to get kids to play games nowadays. They'd also use friends, family, Valve staffers working on other projects, and others.

Q) How will they make the next Portal (if there is one), how will they keep a lot of their design philosophies intact? A) Wolpaw jokes that he's not sure.

Advice for getting into game design? Make games. Make lots of games. Fail and try again. Just keep trying.

Q) Was it intentional to make players feel like they've broken the game after The Fire Pit? A) Yeah, they planned it but it ended up working better than they thought. They're taking some credit for it, but not all.

Q) Why did they decide to throw in comedy for Portal because Narbacular Drop really wasn't that funny. A) Wolpaw: "If no one was going to object, I was going to make it funny."

Q) How did the element of cake come into fruition? A) It started as an inside joke. "It really didn't have any more meaning than that."

Q) Why did they decide to make the protagonist and GLaDOS female? A) Gabe suggested it and they said "alright, sure."

Q) Obviously Portal was a very short game. Do they see a movement towards shorter-form games? A) Wolpaw says he's not sure that people want longer games. No one finishes games unless they're 14. "You just stop playing at some point and never finish it." They wanted everyone to be able to finish their game.

Q) How much final dialogue got cut and could they release it? A) Virtually everything was done as text-to-speech and was cut at that level; some of it Wolpaw just isn't proud of; the rest, he may want to reuse.

Q) Were there any considerations for making Portal a co-op or multiplayer game? A) They did consider it, but just didn't have the time to consider the gameplay. Turns out Portals in deathmatch aren't fun (hrm, really? Color us surprised).

Q) If your development process was a style of painting, what would it be? A) Cubist?

Q) Chell has good DNA, Gordon Freeman has good DNA ... A) Are they going to hook up? "You guys have
access to Garry's Mod, don't you?"

Wolpaw on game writing: "I almost think more of what I'm doing is like writing a film score." It's not central to the role of a game, but it's

Q) At what point in development did you make the decision to drop Portal into the Half-Life universe? A) Maybe middle of development or so; earlier on for Wolpaw. At first it was because they were using Half-Life art, as the game adopted its own style, they worried less about that.

Q) What do we have to learn from Portal because the best games teach us something? A) "Don't trust anyone." "The cake is a lie."

Q) How did they determine how to sell the game? A) Around the middle of Portal's development, they started talking about The Orange Box -- from the Portal team's standpoint, they were honored to be packaged with HL2 and TF2.

Q) How did they get people to experience the entire story without experiencing the entire game? A) Luckily, the game has always been pretty short, so for the most part they watched testers complete the game.

Q) Does less in more in terms of story apply to other genres? A)

Q) How do you resolve the differences in tone between the Half-Life universe and the Aperture universe? A) Inside the Aperture Science funhouse, it's really GLaDOS that affects that tone.

Q) Is there any chance of an artbook? A) They're talking about it, more of an Orange Box sort of thing. Really, there's not a lot of nice concept art, especially when compared to HL2 or TF2.

Q) Why are the Portals ovals? A) Ovals look better than rectangles.

Q) How much backstory is there that isn't in the game? A) There's a lot, it turns out. Some of it is buried on the aperturescience website.

Q) InflueWolpaw labels "Destination Void" by Frank Herbert about a rogue AI as an influence.

Q) Do they think having girls on the team was partially responsible for the uniqueness of the game? A) Swift says she's not sure ... Wolpaw says, "Having girls on the team makes the room smell better."

And that's it. They've got to catch a plane, we've got to stretch our legs.

This article was originally published on Joystiq.

GDC08: Joystiq interviews SFIV's Yoshi Ono