Its that time of year again! Tab and I have our tickets for the Saturday qualifying sessions and the Firestone Indy Lights and IndyCar Series races at the Kansas Speedway. The race is at the end of the month and you can catch the broadcast on Versus or online at Indycar.com.

Anyway, I’ve updated the Kansas Fan Guide from last year to include more details about the Legends shopping district, and update the schedule of the Royals games the weekend of the race. The Royals are in town and this year, it looks like they might not suck! YAY!

Let me know if I need to make any corrections to the guide, and we’ll see ya at the track!

I know, its already Tuesday, and I’m just now getting around to posting the happenings of GDC on Friday. Its been busy, cut me some slack. At any rate, Friday at GDC was like most final days of a conference: lots of people with suitcases in tow, gradually diminishing population throughout the day, and exhibitors tearing down their booths and trying to give away as much remaining swag as possible. Unfortunately, I didn’t have too much time to make it over to the Expo hall on this last day, but I did manage to secure some good swag. Perhaps the best of all was available at the NASA booth. Yes, NASA was at GDC this year promoting their library of 3D models that are free for all to use. They had some very nice pins, big meatball stickers and meatball temporary tatoos. (don’t know what I’m going to do with the tat)

The math and physics discussions were over for the week, although there were a few session on how physics was integrated into some animation packages. The session I attended was High-Definition Physics with Clothing and Vegetation by NVIDIA APEX. I’m not an animator, but the presentation was very interesting. The APEX animation engine incorporates the PhysX engine as well, so animation and simulation can be combined in one tool with no coding required. Very nice. Check out NVIDIA’s APEX website linked above.

Attending the Math and Physics tutorials the first two days was certainly worthwhile, but I did regret not being able to also see the Serious Games Summit. Fortunately, Ben Sawyer of the Serious Games Initiative gave a great summary of the two-day summit. One of the biggest consumers of serious games currently is the health field. Through games such as DDR and WiiFit, health organizations are starting to realize that video games can drive physical activity. You don’t always have to play video games on a couch! In order to reinforce this trend in the health fields, the Serious Games Initiative created a group called Games for Health which is looking at not only how games and game technology can help patients, but also how it can aid nurses and doctors improve their ability to deliver quality health care. There were many other presenters as well, including the more traditional education-based serious games from Hidden Agenda Games. HA Games uses an interesting model of providing financial incentives for college students who develop educational games for high school students. An interesting model, and one that would benefit students at both levels. For an example of one of their games, check out SlinkyBall which demonstrates the function of simple machines.

The last talk I attended for the conference was a historical look at the game development challenges of the Atari 2600 by Ian Bogost of Persuasive Games. It wasn’t exactly what I’d expected going in, but it was an entertaining talk which tied into Kojima’s keynote from Thursday regarding how hardware limitations can influence game design.

With the conference over, Russ and I headed down to Fisherman’s Wharf and Pier 39 to enjoy the sights and seafood. Good times and great food were had, and many pictures were taken. Here’s a small sample. The rest can be seen in the GDC09 set I have on Flickr.

(yes, I needed to clean my lens and image sensor. it was a long week)

More physics and fluid dynamics! YAY! Ok, actually less physics and fluid dynamics theory, and more on their role in game design. The first talk was How Sackboy Learned To Love Physics by Dave Smith co-founder of Media Molecule which produced the triple-award winning Little Big Planet. The talk wasn’t so much about the physics used to create the game as it was about how physics can influence design. Any game that has onscreen movement must include some level of physics, but how much? Is more always better? The example that Dave used was Donkey Kong. Of course there is some physics involved since there is certainly movement, and there has to be some level of collision detection. However, now we have more sophisticated hardware and can perform full 3D rigid-body simulations at 60 Hz, would Donkey Kong be a better game if it were rendered in 3D will full physics? Of course not, so rather than applying mad amounts of physics to every single game, a designer needs to consider how much physics to include. There is a trade off, the more physics and dynamics one includes in a game, the less control the designer will have over the direction of the game play, and the greater the risk of a kablamo resulting from unanticipated paradoxes or crashes. As a physicist, my default reaction is “Yay more physics!”, however, after hearing Dave’s talk, I can certainly see that, in a game environment, high-end physics simulation can sometimes get in the way of game design. Its back to the overall theme I’ve been seeing all week, know when it’s “good enough!”

The second big keynote talk was given today by none other than Hideo Kojima, creator of the Metal Gear series of games, the first of which gave birth to the genre of stealth games. Although he wasn’t as dynamic of a speaker as Mr. Iwata, he had a lot of really great insights about game design. The major takeaway that I got from his talk was the three basic elements to overcoming obstacles in game creation: hardware technology, software techniques, and game design. When limited by the hardware, one must be more creating with the software techniques that one employs and the design of the game play. The reason why the original Metal Gear became a stealth game was because of the limitation of the MSX2 platform for which the game was being built. Each iteration of the game over the past two decades had to overcome similar challenges. As the hardware improved, more and more was possible, but also more and more was expected, so clever game design was still very critical to the commercial success of the series.

The poster presentations were pretty sparse this year, but there was one very good one from Dr. Michael Gourley at University of Central Florida on the modeling of fluid turbulence. Check out this video demonstrating his simulation methods.

The last interesting session for the day was titled Beauty of Destruction. I should have read the abstract instead of just the title. I saw the title and thought that it was going to be a discussion of the animation and simulation techniques for modeling destruction of game objects, but rather is was a discussion of using the C++ destructor function. This is NOT what I had expected, but it was certainly interesting.

After the end of the day’s sessions, I met up with my good friend Steve who now works at Berzerkly and we ventured out to Johnny Foley’s again and had fantastic cottage pie, Guinness, and a desert of Bushmill’s 16yr single malt. YUM! I won’t bore you with all of the incredibly nerdy conversation we had regarding fluid dynamics, stress tensors, and the pitfalls of SPH, but that wasn’t all we discussed. After seeing my haul of swag, he was super jealous! The AGU conference doesn’t have swag anywhere half as cool as GDC. We also spent a lot of time discussing how excited we were about the coming IndyCar season and especially the Indianapolis 500 Mile Race. It was a good evening. I’ll see you in May, Steve, and I have your uber piece of swag!

Wednesday marked the beginning of the main part of the conference, which also means that the population density surged significantly! The local cell grid also took a serious hit! Service for me was spotty all day long. 20,000 geeks all trying to hit the same cell tower all at once probably had something to do with it. The crowds will likely thin out over time as people sift their way out into the city to explore inbetween session. Its still likely to be crowded tomorrow, though, …and Friday.

The day started off with a keynote address from none other than the president of Nintendo, Satoru Iwata. Last time I was at GDC, the keynote was done by Shigeru Miyamoto, and it too was fantastic. Is being a dynamic speaker a requirement for employment at Nintendo? Needless to say, the queue to get in was LONG. At least we got there early enough to only be a block away from the front door. At any rate, Mr. Iwata’s talk was fantastic and started with a discussion what makes for a successful personal and business philosophy with respect to project development. The traditional “deadline” model tends to lead toward decay in what Iwata called a “death spiral”. I can see his point to an extent, but as cool as free-form and open-timeline development may be, business loans, rent, and other bills DO have a deadline. Point taken, though, that short deadline times do lead to shoddy development and inferior products, which lead to weak sales and poor growth. The last half of his talk was spent pimping Nintendo and some of the new stuff that’s coming. There are some upgrades for the Wii that will allow it to use SDHC cards which will benefit those who’ve already filled up one 2GB card and their main Wii Menu. He also introduced the new version of their handheld console, the DSi, which incorporates an SD reader, and a built-in camera that performs much like the EyeToy for the Playstation. Best of all was a new game that gets released in a little more than a week, Rhythm Heaven. However, I won’t be waiting in line for it, nor will my colleagues. We already have a copy! At the end of the keynote, they gave away copies as we were headed out the door. w00t! …and no, I don’t have an extra!

There were two fantastic talks today. The first being Real-Time Deformation and Fracture - Finite Element Simulation and its Use in STAR WARS: THE FORCE UNLEASHED. In most engines, the fracture points for an object are predetermined when the artist creates the asset. This can lead to a simplistic and roughly believable behavior, but James O’Brien of UC Berkley and Eric Parker of Pixelux demonstrated a method for determining fracture and deflection behavior used for the Force Unleashed game that makes use of a more physical representation of the material properties with the stress tensor. It was a very cool formulation, and I’m looking forward to getting their slides once they’re available.

The last good talk of the day was an introduction to the Kaboom code from Intel, Kaboom: Real-Time Multi-Threaded Fluid Simulation for Games. This system for modeling fluid dynamics permits a more physical treatment of the fluid than typical SPH used for gaming environments. Kaboom includes enough detail for changes in temperature to affect the density, dynamic pressure and bulk flow of the fluid. This results in very effective modeling of hot smokes and cool fogs. Best feature of all is that this code is being made public by Intel. Its not out yet, but look for it soon at Intel’s Visual Adrenaline site.

The last event of the day was the annual Independent Games Festival Awards and the Game Developers’ Choice Awards. This is a fantastic event! You can see all of the finalists, and play a demo of the games, for the IGF awards from the IGF website. Sadly, PixelJunk Eden received no awards, despite being nominated in three different categories.

Although Fallout 3 won the Game of the Year award, it was really a big night for Little BIG Planet. LBP took home three separate awards: Best Debut, Best Technology, and Innovation Awards. It was a bit funny when the developers came up to accept the Best Debut award. The first comment when they got to the podium was, “Best debut? We’ve been doing this s*!^ for 20 years!”

Tomorrow is more Expo goodness and swag gathering.

We’ve been very lucky this week. Its been sunny and warm every single day, and is likely to remain that way for the duration! This also means that there’s been some terrific lighting for pictures in the morning (the one time I’ve actually been free). Here’s a picture of the Moscone Center where the conference is being held.

Moscone Center in the Morning

Today was hit-n-miss. Some of the talks on physics as applied to game programming were very good, others were so much chest-beating and not so informative. Most of this morning was spent discussing methods of collision detection. The talks had lots of good information delivered by really bad speakers in many cases. Guys, this is why they made us take Public Speaking! At any rate, good collision detection can make or break a good physics engine, and its more than just detecting and preventing object penetrations. How an engine responds to a collision is equally important. The two main speakers on this issue were Gino van der Bergen of DTECTA and Erin Catto of Blizzard.

Gino’s discussion on collision detection led into a seriously deep discussion of the GJK Algorithm and how useful and robust it is for detecting collisions via Axially-Aligned Bounding Box (AABB) trees. Very neat stuff, but it makes use of configuration space, which I still don’t fully understand. I need to spend a lot more time looking at this formulation. Fortunately, Gino’s slides are online at DTECTA.com, so I can go over them later.

The next really good talk of the day unfortunately straddled the lunch hour. Erin just started to launch into a great discussion of Jacobians and Lagrange multipliers when we broke for lunch. Fortunately, he did spend some time after lunch going back over the formulation leading up to where we were at the break. The end goal of Erin’s discussion was to arrive at a good method for treating and evaluating constrained motion without inducing instability or jittery response. Very cool stuff. Check out his slides and demos at Gphysics.com.

The last worthwhile talk of the day was by Kees Van Kooten of Virtual Proteins. He described a particle-based method for modeling fluid dynamics. Its actually a very slick and effective method, although there were many times during the talk that I had to bite my tongue a bit. I have to keep reminding myself that the intent is to arrive at believable behavior, not physically and thermodynamically correct behavior. As my old adviser is fond of saying, “perfectionism is the enemy of good-enough,” and this method is certainly good enough. The method is called Smoothed Particle Hydrodynamics, and can be used to model fluids in a very physically accurate way. Its a great tool for space plasmas, in fact! The basic jist is to smooth each particle with a kernel function, model the nearest-neighbor interactions to determine the internal pressure and viscosity forces, then apply the external forces to the system, and grind the particles through an Euler-style integration. To render the fluid, a mesh wrapper is applied around the particles (metaballs). Sadly, the slides from this talk aren’t up on the net yet, but in about a week, they should be available at EssentialMath.com.

Well that’s it for the two-day Math and Physics tutorials, now on to the rest of the conference. Photowise, there will be less purely architectural photos since Wednesday beings the main part of the conference and the huge onslaught of attendees. For comparison, here are two pics I took today. The first is of a recessed area just off of the main classrooms on the third floor of Moscone West, and the second is of the Intel Booth. Tomorrow I’ll take pics from the same two locations to highlight the serious increase in the mass of people.

Negative Space

Intel Booth on the 3rd Floor of Moscone West