The Evolution Of Games

Thanks to Reality Panic for pointing to a new IGDA-hosted article comparing evolutionary biology to the organic process of videogame creation. The author compares the Cambrian era, an "early period of developmental simplicity for organisms", to the '70s and early '80s for games, with both containing "...a number of... oddities with few or no modern descendants". He goes on to liken the possible wiping out of the dinosaurs with "the impact of a giant meteor" to "...the arrival of the Sony PlayStation... [marking] a mass extinction of 2D games", and concludes by suggesting that, like the evolution of fauna and flora, "...periodic outbreaks of originality, and the corresponding extinction of certain game genres, are useful to drive the form forward, but the conservative intervals between these events are what serve to sustain."

Comments

2D games

[sofakingl]

2D games are not extinct; games such as Guilty Gear X and SNK vs Capcom continue to be made, and have a solid fan following.

Re:2D games

[TRACK-YOUR-POSITION]

Though there are few examples of 2d games being made on the home consoles (some of which are absolutely incredible) even more significant are the vast sums of money being made by 2d games in the portable (game boy, cell phone, pda) markets. 2d isn't just still alive, it's the next big thing.

Re:Gaming is at a Nexus.

[PainKilleR-CE]

I'm not going over the whole thing, but I will give you an idea of where my mind is at. I see polygons going away for the most part. Future, I know, but I'm picturing 3D models being sculpted with curves that are expressed mathmaticaly.

That's really just a matter of getting the tools in the hands of model designers. As it is, the game will treat the model as a series of mathematic expressions, and curves are just another level of math to deal with on the software and hardware fronts (and every curve is usually just broken up into approximations in polygons anyway, even in math).

Instead of coloring 10,000 2D tiles why not have 1 object expressed as a formula that renders the entire object?

This is the way it's already handled by most 3D engines. If you have a 3000 polygon character model, the model was designed as a whole by the artist, and is stored as a whole, and one texture (skin) is applied to the whole model. The formula that renders it, on the other hand, has to break it up into polygons (or it's broken up during compilation and stored as a series of polygons) because the graphics cards handle triangles and polygons (because they're easier to work with).

This would allow it to scale tremendously, and if the output does require polygons (variable by the capabilities of the hardware) a driver or engine update could increase the end result rendered polygons.

This sounds similar to the MRM technology from Intel licensed by Valve and others (the only released game I know uses it is Dark Reign 2). The idea is that you design a model with as many polygons as you want, and then the mesh (the exterior of the model if you will) is reduced in the number of polygons by a mathematical model according to how many polygons are on screen. It only works in one direction (and only ever could), which is to reduce the number of polygons. Increasing the number of polygons requires a human eye. However, if there were unlimited budgets and infinitely patient designers out there, developers could have models that are well beyond the polygon count that any current system could produce, even if it was the only thing on the screen. Of course, this works (once it's implemented) without any changes to the engine or drivers, because it's based on what the machine can handle under that engine. In theory, the perfect model could look like just another Half-Life model today, and almost exactly the same as what the designer put together in 2 years, on the same version of the game with new hardware.

As long as old games are created with something vectorish they should be able to scale upwards quite well. Maybe there will be new features in the engine the old games wont use, but as long as the engine contains improved pieces that are compatible with the objects of the older engine it should work better.

Vectors, unfortunately, don't improve the quality of 3D objects when they're scaled. You could use vectors to resize models and (hopefully, depending on the angles used) preserve the model's quality at different sizes, but it wouldn't change the number of polygons.

For example take a MIDI file. I've played MIDI files on cheap hardware that sounds better than an ATARI 2600, but it still beeps, boops, and sounds electronic. I've played the same MIDI files through really good high end sound cards and pro-audio equipment and it sounds like a real instrument. Maybe thats a bad example, but the input source was the same.

This is the same example that you would find with polygon-based models and any technology that scales them well for the hardware. That MIDI file is only going to sound as good as the source, regardless of how much better your hardware gets. If it was designed on an Atari, the individual instruments may sound better, but it will still be limited by the Atari's capabilities (ie you won't get massive 128-voice sound from the MIDI file if the source had a maximum of 8 simultaneous voices). Or, just like you can play most ster