One of my pet peeves with the IGF and the GDC Awards and 90% of all game reviews is the way they kind of dissect the game like it's a toaster or a car or something.  The way it looks is graded separately from the way it plays, the way it sounds is graded separately from how it performs, etc.  For the last few years this has irked me in a really...simple, superficial way.  Movies aren't graded on their graphics and sound separately, right?  You never see this kind of review for a movie:

"Camera Work: 90.  Music and Sound: 40.  Story: 50.  Overall Score: 60."

Movies are reviewed as a whole, en totale.  A movie is judged as much by how appropriate all the choices were as it is by what the decisions themselves actually were.  It cheapens games to be broken down in this...consumer reports kind of way.

So, that was my gripe for a long time.  But lately it's become clear that this isn't just irksome or frustrating, it's completely nonsensical.  More than any other medium, all the components of a game are deeply interconnected and interdependent.  I've been referring to this lately as "holistic" but that might be a bullshit term for it, I just like how it sounds!

A couple of experiences threw this into sharp relief for me.  For the last few months I've been working on a racing game for the iPhone.  I've never worked on a racing title before so it's been a constant learning experience.  But the main thing I've learned is that making one tiny change in the prototype would inevitably affect fifty other things.

For example, a few weeks into the prototype I changed from an accelerometer-driven control system to a touch-powered control system.  Both essentially took an angle as input, so I thought it would be a pretty simple swap.  Unfortunately, the accelerometer system had a bunch of cushioning and lag that was made much more noticeable by the more reactive touch system.  But as I removed the lag, the touch got too touchy, and I had to modify the physics system to be more forgiving.

Modifications to the physics system changed the way the AI cars were steering, and they started to miss waypoints.  When the cars started missing waypoints, the place tracking system went completely to hell.  All I wanted to do was change how the controls worked, and it broke the system that figures out who's in first and who's in second!

A couple weeks later it became clear that the racetrack and cars were just too small, beta testers were practically squinting at the screen, which is never a good sign.  It was a very simple matter to just increase the scale of the vehicles in the prototype, maybe five minutes of editing.  But the game was suddenly unplayable, because the racetrack wasn't wide enough to accompany the new larger cars.  So I added some extra tiles to the test track, and made it three tiles wide instead of two tiles wide.  That took maybe thirty seconds.  The game's framerate dropped by at least 50% on the device.

Turns out if you're trying to draw a whole second track's worth of tiles it affects the performance!  So, I increased the size of the tiles from 32 to 48, so that I could still have a two tiles wide track.  However, increasing the size of the tiles meant I could fit only 60% as many on our 512x512 texture atlas that we were using to render the track. We had to go through and simplify the tile design to make them more reusable, which in turn affected the types of tracks that we'd be able to include...

Maybe this was already obvious to everybody, but I originally got into game design and indie gaming through the Half Life modding community, which is basically founded on the idea of modularity and interchangeability.  You just take some generic tech, pick an art style, and go at it.  There have been some amazing successes and creations from that direction, but I think they all either played to the tech or substantially altered the tech to fit their purposes.

Anyways, I'm not really sure where I'm even going with this!  Just something that's been on the mind lately.  I guess the point is that it's useful and interesting to dissect games into categories like graphics, or mechanics, or technology, but only in a very general sense; when you start talking about a specific game, it's all about how well those things work together.