Get the Flash Player to see this player.
Bring home your new best friend today!
Spend many happy hours pampering, grooming and playing fetch with your new canine pal. Play with your dog, take him for walks and run through the park to perfect his agility and obedience. He needs your love and patience to grow into a trained pedigree dog that can compete in national competitions. Do you and your dog have what it takes to be named Best in Show?
“Fetch!” was the first title I worked on at Virtual Playground. I joined the studio three months into the nine-month development period, so to begin with I was helping getting a lot of the tech in place. Since most of the games produced by Virtual Playground are Tycoon games, there was a reasonable amount of work in moving from the robotic, quite predictable Tycoon characters to something that felt a little more alive. Some of the more interesting tasks I undertook during this project were:-
- - Implementation of an improved needs-driven behaviour system. Rather than storing needs as simple on/off states, each need was given a fractional value which could be affected by a number of factors. Actions were then prioritised by the value of the need they succoured, and an estimate of the distance the agent had to travel to perform the action. This system was later moved into the engine to be used in future Tycoon games.
- - Design and implementation of the gesture recognition system. I had considered that certain similar titles make use of the microphone to communicate with the dog, but that not all PC users have microphones. I decided that the best equivalent on PC would be mouse gestures, since a mouse was already required to play the game, but these are far more involving than simply selecting from a menu. The gesture recognition was implemented using a neural network.
- - I developed the two minigames featured: the agility and obedience trials. One involves running between obstacles in the best possible time, while the other is a test of the player’s ability with the gestures – with scores being given based on how close the gesture made was to that intended.
As well as this, there were of course a great many everyday tasks in a variety of areas. I really enjoyed working on this project – partly because it was my first professional work, of course – but also because it allowed me to think about AI and gameplay in ways I hadn’t considered before.