Control Point (animated)
by Subby on Feb.13, 2009, under Pre-Production
No behavior yet, watch it spin…
Simple movement, now with camera
by Subby on Feb.11, 2009, under Pre-Production
Boulder trap brainstorm
by Subby on Feb.08, 2009, under Brainstorm, Pre-Production
Idea: Position that player issues command to lay trap will be the position the boulder will eventually roll over. The problem here is where best to place the boulder such that this eventually happens.
yunxu: didn’t go too well. Boulder may not roll towards desired direction if only using free fall physics. Will be implementing one which has a initial force acting on the boulder. Results will be posted here.
Update: Trap works a lot better. Refinements needed as problems might arise if player decides to build at the peak of a region.
Terrain + Physics
by Subby on Feb.04, 2009, under Pre-Production
Added some cubes to Anselm’s terrain to demonstrate the physics engine.
Terrain generation
by anselm on Feb.01, 2009, under Pre-Production
At first we thought of using terrain view of google maps to generate terrain, so the player will feel more fun playing the game. After some discussion with Allan we finally dropped the idea of using google map’s terrain view since it is not a height map nor a gradient map, so it will be hard to gather useful information from terrain view map and generate a game world out of it.
Secondly we thought of using the GIS(Geographic Information Systems) data. However, since we are intending to implement the game onto Xbox 360, and a good set of GIS data are usually quite huge, memory on xbox might not be enough.
The third suggested option was to use diamond-square algorithm to randomly generate a terrain. Basically it is an algorithm to randomly generate a height map that the neighboring points are not very far off. After a few parses the height map was quite a decent one as show below.

The last option was to randomly generate some points at fixed intervals and using bezier curvesto interpolate the points in between + Perlin noise to generate the terrain. This is an interesting option as it could also be used along with Geographical Mip Mapping where terrains further away are interpolated with less points making real time rendering a lot cheaper.
the result of using diamond-square random map generation. Everytime it generates a new map that looks natural.
Weekly Discussion – part 2
by boro84 on Jan.29, 2009, under Meeting Outlines
Exploration (Anselm)
Generate game map from real satellite images
Timeline
Stage 1: Image processing on Google Maps image to generate 3D terrain
Stage 2: Generate control points based on generated terrain
Stage 3: Populate the map – trees, building, monsters w influence map (w Yunxu)
Stage 4: Process user input using Google Maps API
Exploration (Yunxu)Terrain analysis for trap placement, high level Squad AI decisions
Timeline
Stage 1: decide best location for 1 trap type – Boulder Trap – high ground facing possible path.
Stage 2: 2 more types of trap, e.g. the powerful Bear Trap – low visibility, high foliage
Stage 3: intelligent squad deployment AI (w Boon Pin)
Stage 4: TBC
Exploration (Boon Pin)
Intelligent Combat AI algorithms for NPCs (Monsters)
Timeline
Stage 1: Basic pathfinding and monster AI
Stage 2: High level Commander AI based on control points (w Anselm)
Stage 3: Squad AI
Stage 4: Advanced squad AI for squad and monster
Exploration (Shao Chong)
Engine development, UI and enhancing game play through camera and controls
Timeline
Stage 1: Engine up w combat capability and fly through camera
Stage 2: Game play coding, camera research
Stage 3: Camera development and combat control
Stage 4: Fine tuning of camera and combat controls via play testing
Weekly Discussion – part 1 (online)
by boro84 on Jan.29, 2009, under Meeting Outlines
Agenda
———–
1] Update
2] Reflections on what transpired last week
3] to-do
Update
The team read through and is in the process of compiling the Innovation Engine into XP-Dev
Abit of research on individual parts
Reflections
Input from Allan:
3D Game is wide (complex) and difficult to implement
individual’s parts are dependent on each other, ie. Map Representation -> Pathfinding -> AI, etc
advise is to scale down on implementation so as to fit into project timeframe
features to show must be obvious (eg. clever AI must be able to be observed by player)
To-Do
Meeting the next day to sort out product backlog and plan sprints
Follow-up on relevant people for update on xbox loan, creators’ club premium account, schedule for meeting with Claude and Dr. Ashraf, updating Allan on meeting timings
part 2 coming up after tommorow’s meeting








