Thursday 9:00am–10:00am: “Classic Game Postmortem – Out of This World/Another World, by Eric Chahi

This was a look back at the making of Another World.

Overview of the creative process. Freedom under constraint: an improvisation process.

Initial context. July 1989, the Amiga era. He’s 22 years old, but had written a lot of games; taken 3 weeks to 6 months, none really successful. Dragon’s Lair had just been released: how to do this?

2D game based on polygons. Cover lots of the screen with not so much data. Also wanted to use rotoscoping techniques, inspired by some new hardware. But: is it possible to draw polygons at the right speed? Experimented, yes. So graphics unit: polygons and only polygons, that’s a key constraint. He was obsessed with this: how would a potato look in polygons?

At times, he used “pixigons”, polygons the size of pixels. So the first level took too long, doing pixel art in an obtuse fashion. He ended up using bitmaps for the backgrounds, to increase detail, but only ten bitmaps in the whole game.

Create a 2D game with a cinematic feeling?

He didn’t have suitable tools available, so he had to create an engine and tools from scratch. He wrote a polygon editor, and an interpreter for a homegrown scripting language for the game logic. (Including animation!) Shows the main character in the editor, including a demonstration of groups of polygons. (E.g. the head.)

Fusion between code and graphics: you can modify the graphics in the polygon editor and see the interpreted code change, or vice-versa.  It was a very simple language, with only 256 possible variable names; he ended up writing down the meanings of variable names on paper! The engine had 64 tracks, each of which could run its own code. (Like a multi-track audio mixer, I guess.)

December 1989: tools and engine were usable, albeit flawed. One flaw: no common data shared between levels! So he copied data back and forth, which was a nightmare when tuning: each change in Lester’s behavior had to be done five times.

Did the initial animation. Polygons worked well; rotoscoping, not so much.

Tried hard to listen to his inner self. Including surprising himself: what would he expect as a player at a given moment, how can he subvert that expectation? E.g. when player is transported initially to the other world, initially thought that the player would appear on the surface, but much more surprising if he appeared under water.

The basic gameplay mechanics were inspired by Karateka.

16 colors in the palette. Which meant that he needed to have multiple uses for colors: e.g. used the main character’s flesh color for sunset highlights. Using black hair would have been useful to that end; but then the character would have looked too much like Eric himself, which would have been disturbing! Red hair loved that latter problem, while still being a reusable color.

He used parallel action in the background; but then the background action comes to the foreground (the lion appears), with a cutscene ad punctuation, then varies up the pacing in the subsequent chase, backtracking, rescue by a hooded figure.

That ends in a cliffhanger; the next scene starts by removing the hood (in a different location), so he could learn what’s under it! And then friendship gets introduced, including the only closeup of Buddy’s face.

Also, this level introduces both weapons and shields. That combination turned out to allow a lot of interesting gameplay combinations. (Especially whe you throw in the slow shield-destroying plasma shot.) Do you use a fast shot, take the time to set up a shield, take still more time to destroy a shield?

Also, going from the top to the bottom of the screen: he’d initially planned to use an elevator, but a teleporter felt better.

At the end of the jail level, it was clear that you’d have to meet Buddy again later.

December 1990: 17 months in, only 1/3 of the game is done. So he needed to work faster; he focused on puzzles that used the same building blocks he’d already established. The frame of the story ended up being a series of meetings and separations between Lester and Buddy. (That wasn’t the plan at the start: improvisation.)

He used separation of foreground and background more: that gave a cut scene effect without interrupting gameplay. And played more with pacing.

Publishing. He met with somebody at Virgin Games: they tried to convince Eric to make it easier by converting it into a point and click adventure. He was tempted, but decided not to; went with Delphine Software, who were supportive of the existing software.

June 1991: time is running out (it was scheduled to be released in November).  How should he end the game? He went through the game so far, and drew a map of key past events; he decided to play with the pace, leading to the area scene.

The cover illustrations of his previous games had been particularly unsatisfying. He was already passionate about illustration, and Delphine Software was willing to let him illustrate the cover himself. But he didn’t have time!

Also, they weren’t doing enough play testing. They found a few issues, but it was still unbalanced. Later, Interplay ported it: they polished some aspects, but also wanted to make the game longer, more difficult.

Interplay wanted to change the music, which he really objected to. They had a long war via fax, which eventually led to Eric sending an infinite fax, consisting of a looping strip of paper in the middle of the night, saying “keep the original intro music!” Interplay didn’t change their mind, but Delphine pointed out that Interplay had no legal right to change the music.

Console porting led to a bit of censorship, about an alien butt-crack. He considered sending an infinite fax, but decided to remove three pixels instead.

Eric was exhausted at the end, which is probably why Lester is almost dying at the end of the game.

He’s surprised at how long the game has lived: in fact, its being ported to iPad now!

Post Revisions:

This post has not been revised since publication.