Agile Game Development: Tales From The Trenches
By Noel Llopis
12 November 2006
Still curious about how to apply agile development to games? These presentations from Gamefest 2006 and the Montréal Games Summit 2006 give an overview of agile game development, and then get into some of the practices we use at High Moon Studios.
- Gamefest 2006 presentation. This one is a gentle introduction to agile game development, covering both process and technical issues (audio also available).
- Montreal 2006 presentation. This one deals only with technical agile game development topics, and gets into more detail about when and how to use them.
Page 5: White text on white picture? O_o lol
Posted by: wursel at November 14, 2006 07:56 AMYeah, sorry about the white on white. There was an animation in that page, and the text wasn't supposed to be turned on when the picture was there. I guess those are the things lost when you export them to pdf. I figured that was a better format than OO Impress (or MS Powerpoint).
Posted by: Noel Llopis at November 14, 2006 09:57 PMThe Gamefest '06 audio link currently appears to be pointing at a very handy zip file containing XNA and Visual Studio. Any chance of linking to where the audio actually is?
Posted by: Dave Mariner at November 16, 2006 07:46 AMDave,
The zip file includes the audio and slides for all the sessions in the XNA and Visual Studio track. My presentation was listed as Intro to Agile Game Development. I know it's a big download, but I'm not sure whether I can make that file available for download by itself.
Posted by: Noel Llopis at November 16, 2006 12:41 PMHi Noel,
I've been reading your blog and appreciate the effort you've put into bringing new ideas to the wider audience of the games industry. Thanks.
Perhaps you can help me understand the UnitTest++ philosophy, though? I find myself struggling to believe how you can set up a system that runs several thousand tests in a fraction of a second and tells you anything you didn't already know.
I mean, it's either dealing with real game data to do it's tests, in which case, how is it managing to load all that data in fast enough? Or, it's not, in which case it's not telling you anything useful, surely? It's not telling you if your code will behave as it should with real data, it's telling you that your code will behave as it should if provided solely with contrived test data.
Secondly, UnitTesting appears to approach the problem from the bottom-up in comparison to the typical QA approach of tackling if from the top-down (see http://www.romsteady.net/blog/ for interesting perspectives on this). My guess is neither method really catches the "middle-ground" issues that can plague development - large scale quirks that only become apparent when multiple systems interact and are notoriously difficult to describe succinctly in a bug report.
Finally, how would you design a UnitTest to verify that an idle timeout is occuring correctly (e.g. front end menu times-out to play an attract-mode video), without it taking (e.g.) 30 seconds to run? And if it doesn't take 30 seconds to run, what exactly are you testing?
Posted by: Dino at December 4, 2006 05:30 AM