[This] an entry submitted to the 2013 Global Game Jam.

Our < 48-hour team was primarily 4 people:

Myself (lead programmer/technical-artist/3D modeler)
My buddy from college/work (producer/designer/programmer/technical-artist)
and 2 cool guys we met on site (programmer/sound-engineer, 2D texture artist/3D modeler)

The 4 of us created:

A top-down perspective prototype that allowed a player to move with an X-Box controller
Shooting and destructability
AI for enemies
Spawning for enemies
Sounds for in-game actions (thanks www.bfxr.net)
3 textures with normal maps and light maps
4 particle effects and projectiles
3 non-primitive 3D models (not rigged)
several images for HUD UI, including 3 animated sprites, and code to make it all work together
map loader that could generate maps from a text file (and a test map)
2 player controls, including camera code to keep them both visible (still buggy)
a heart-rate +cool-down mechanic (for the game theme: heartbeat)
tweaks to improve performance and fix bugs
simple UI code for game and title screen

We had additional help from extra people for:

1 3D model
In-game music (removed from web build for download-size issues)
Title screen graphics
technical problems with Unity (particularly for XBox controller input)

We did not create (despite planning):

In-game items and pickups
Special abilities, including new attacks
A-star path finding
Random level generation
Customizable controller settings
Rigged models

Our process was fairly simple, but I think it went as smoothly as it did because my friend and I have some industry experience, including programming and management/planning. Our general goal was to "just make the game", and I think the actual execution went quite well. I would generalize our process as:

Early phase

Design: basic game concept+genre, itemized+prioritized gameplay features, and itemized+prioritiezed art assets. Oh, and maybe a story. Ideally 1 person leads the design, with creative input from others.
Start the game prototype ASAP (potentially right when gameplay genre is defined), and get something playable that can be appended to by other team members. This is a Lead Programmer Bottleneck/Critical-Path-type task.
While the prototype is being developed, a technical manager is organizing a shared drop-box/repository/source-control, centralizing contacts, and getting communication and asset sharing processes clearly established.
Non-lead engineers are doing research and writing code for important-but-not-prototype-required code.
Artists are drawing concept art and working with a designer to establish a theme for the game, and make the vision as visible as possible. An art asset list is created with a designer/producer.

Mid phase

Programmers program. Incrementally add features to the prototype till it looks like the game you expected. A lead engineer identifies tasks for other engineers while the lead does the "heavy lifting". Regular engineers write gameplay/UI/scripts/support-systems, the Lead engineer integrates them, and does the really hard systems otherwise.
Artists do work identified in the planned asset list, and the engineers integrate the art. "Technical Artists" do the art assets AND integrate, which improves the speed of the process and quality of the art.
Producers find roadblocks for the team members and clear them, generally staying out of the team's way otherwise, spending time absorbing (and shielding the team from) administrative responsibility

Late phase (crunch time)

Engineers fix bugs and cram final features and art in. During the early/mid-phase, technology is designed with a long-term perspective. During the late phase, the perspective is "just finish this, and we're done."
Artists keep doing art, but a bit faster and less polished, with less design freedom.
Producers keep doing the same as mid-phase, just under more stress. At the end, final administrative tasks are handled while the technical team crams in work behind the scenes.

Every actual game development experience I've had has been in that pattern. Even the game-jams follow the pattern, just more compressed.

Michael Vaganov

({{monthName[d.date.m]}} {{d.date.y}})