A discussion of hybrid open source and proprietary licensing models.

[Rob Myers <robmyers AT mac.com>]

On 1 Sep 2004, at 20:11, Marshall Van Alstyne wrote:

>  There's an interesting post from slashdot on how open source may not 
> be the best development model for games. 

I worked in games for two years. We almost finished the project. :-)

> The author argues that the main reasons are (attempted summary):
>  (1) long development cycles with very short consumption cycles, that 
> limit useful feedback.

Successful games have long consumption cycles, and rely on new 
materials to extend their lifespan. Role-playing games (like the 
Dungeons and Dragons 3e computer games) rely on player-created modules 
for gameplay. Single-player games, like "The Sims", have expansion 
packs of new materials released to supplement those shipped with the 
game.

"Modding", producing new characters, maps, challenges and stories for 
existing games, building on the materials shipped with the game, is a 
major ingredient in building communities of consumers for games from 
the more successful companies. This is Open Source in all but license.

> (2)  reuse of characters & graphics is difficult
>
> Now imagine that the developers for our open source Doom 3 can take 
> the art from Doom 2 to use as a base for Doom 3. But this isn't very 
> useful. The artist can't load the art for an Imp monster circa 1993 
> into The GIMP, apply a filter, and suddenly have an amazing 3-d model 
> with bump mapping. In fact, the only area of game development where 
> reuse is a major advantage is the ability to use an existing game 
> engine. But most closed source developers already do this.

This is a hacker's argument regarding art...  :-) Those low-polygon 
models for the monster may have been reduced from high-poly models. 
Those low-resolution textures likewise will have been reduced from 
hi-res textures. So if the source materials were available, this would 
be useful And even if they were lo-poly/lo-res to start with they can 
be smoothed algorithmically as a starting point.

If nothing else, you can re-use the look without plagiarism if it's 
Open.

> It would be interesting to see if we could design licensing mechanisms 
> that would facilitate open content for markets with these 
> characteristics.

One poster to the Slashdot discussion mentioned that ID (authors of 
Quake) factor the licensing revenue from the game engines that they 
write into the budget for the games that they create the game engines 
for. The poster argued that they would be committing financial suicide 
by giving away the game engine for free (sic).
Which is a funny way of saying that if they wrote the game engine as 
Open Source their financial risks on each project would be reduced. :-)

One interesting thing for licensing with computer games is that there 
is a clean conceptual and programmatic  split between the core system 
(the "game engine" that loads and runs the resources and gameplay for 
the game, written in a compiled language such as C) and the support 
materials (art, models and the 'scripts' that drive the gameplay).

So one could, in theory, keep the engine closed whilst licensing the 
art and scripts liberally for community-driven modding. This would 
return value to the authors of the game engine by maintaining interest 
in their product, extending shelf-life and increasing sales.

Or the engine could be opened, reducing the risk of relying on 
licensing revenue, and the game materials kept closed as proprietary 
added value over the closed engine.

It is an interesting area for Open Content, in many ways there is 
already an Open culture there with the Modding scene that could be 
formalised.

- Rob.