Sandbox Stuff

by Darius Kazemi on July 31, 2006

in Uncategorized

So I attended the ACM Sandbox Symposium this weekend. For the first year of a conference, it was pretty good.

Greg Costikyan‘s keynote was fantastic (even though he was preaching to the proverbial choir that I belong to). The gist of it was that games need more “process intensity”, which is Chris Crawford’s term meaning that games need to spend more CPU cycles doing gameplay calculations, and fewer CPU cycles pushing pixels to the screen. It very loosely maps to the procedural content stuff that I love.

The second day’s keynote was from Ian Shaw, who runs the EA UK studio. This was interesting, because it contrasted greatly with Greg’s talk. It was neat to hear about their use of agile methodologies, and particularly the cell system, which I’d never heard described in detail. One of the things Ian said was that “software is the container, and the media is content.” This seemed at odds with what Greg was saying (not suprising), and it also seemed baldly stupid to me at first, which was suprising because Ian seems like a very smart guy. Isn’t the software what makes the game interactive to begin with? Isn’t the interactivity the content? I asked Ian about it after his talk, and it turns out, yes, he considers the interactivity to be part of the media/content. Another interesting thing he mentioned was that he likes the use of scripting languages to describe game behavior, rather than C++, because it means the interactivity bits aren’t mixed in with the engine code. I can see the use of that–you can then treat the interactivity as an individual asset (a Lua script or something), and then indeed considering interactivity to be the media/content becomes a much more sensible analogy.


Craig Perko July 31, 2006 at 4:21 pm

I definitely agree with what Ian said. Or, at least, what you said he said.

It’s almost impossible for me to think of interactivity or gameplay as “software” any more. Software is the engine which packages my scripts and pics.

Is that the summary of what the “cell system” is?

Kinda wish I’d gone. :)

Ian Schreiber July 31, 2006 at 5:04 pm

Take an example game like Civ 4, where there are multiple layers of “content”. Even a beginning player can figure out how to use the game editor to make maps and scenarios. A computer-savvy player can change unit strengths and other basic stats in XML. An expert user can start messing with the actual game rules in script (I forget if they use Lua or Python or whatnot, but whatever it is it’s made accessible to the end user). So where’s the dividing line between software and content? Is that line even meaningful anymore as a distinction?

Darren Torpey August 9, 2006 at 12:23 pm

The comment about CPU power going to gameplay-related calculations reminds me of the ASSISTments tutoring system that I work on at WPI.

The system is starting to meet some of its early potential as and intelligent tutor by keeping track of student progress and using statistical methods (informed by content creators) to decide what kind of problem to give the student next.

The actual “graphics” of the system are dead simple. Very rudimentary HTML output is all we really need, but eventually there will be tons of processing going on behind the scenes just to decide what problem (or even what hint text) to show next.

It reminds me of how Matt Ward, my animation teacher in undergrad and graphics teacher in grad school, always reminded us that the entire sum of graphics processing all boils down to choosing the right color for each pixel. No more, no less.

It seems like many threads of game development (design + programming, that is), if not all of them, will start to head in the same direction, but with respect to gameplay, as you and your quoted sources suggest.

A million gagillion cycles, and all to decide what challenge/story element to present next. ;)

Comments on this entry are closed.

Previous post:

Next post: