The perils of micro-budget programming

Ken Fisher, developer for ActionMMORPG writes about the process and issues regarding ultra-low budget MMORPG creation. The intent of the blog is not to educate or force opinions, but rather to encourage thought on the subjects. Expect boring :-)

Author: XAPGames

A recent screenshot from WAC...
 

The world in WAC consists of four types of artwork, static shapes, dynamic shapes, interiors and terrains.

I have three of the four working with a modified version of relief mapping.  This creates a simulated high polygon rendering based on a low poly model and a combination normal map / height map known collectively as a relief map.  The relief mapping is shader driven so that it appears to move a texel (a point on a texture or model) location in 3D space including reacting to lighting.

Relief maps are normally very computationally expensive on the GPU, with each pixel rendered requiring multiple routines that step through 3D space to determine light intersections with the model.  My modification to this process eliminates several of these routines while still providing a reasonable representation of lighing and shading.

In application this can be seen in the above picture on the back wall of the placegholder dungeon (an interior model).  This is also working (but not shown) on static shapes (trees, rocks, etc) and dynamic shapes (character models and mobs).
 

Terrains are a problem

I haven't been able to apply this technique to terrains.  The stock terrain system doesn't store sufficient information to apply the shader.  Another issue with terrains in the engine I use is that LOD (reduction of mesh complexity with distance) is not strong enough to allow huge terrains in view without significant slow downs in rendering.

I'm looking into alternatives that would allow long view distances.  One alternative is switching game design to something other than fantasy medieval which is not heavily based on large open world scenics and avoids use of terrains except where they can be applied more efficiently or perhaps replacing them completely with a model that can do relief mapping and LOD.
 

Thinking about CyberPunk

I'm thinking cyberpunk, and I'm currently researching taking the game in that direction.  It's too early to announce that I'm going to roll development into a cyberpunk game.  There still could be issues show up that make doing so less than desirable.  On the other side, I do take the possibility seriously enough to have licensed several art packages for testing.

Looking at the current market.  There are tons of medieval fantasy games, a few urban super-hero games, a few science fiction games, at least one post apocalyptic game, and a few others that defy description.

Would cyberpunk work as an MMORPG sub-genre?  It sort of worked with The Matrix Online (RIP).  My vision is sort of a mix of styling as seen in Matrix, Blade Runner, and perhaps a touch of Aliens (but not enough to push  the genre from cyberpunk to sci-fi).

I'm not thinking full sandbox, nor am I thinking quest driven themepark.  I don't approve of heavy instancing or phasing because of them killing connected world feel.  I like zones / shards as a method of load distribution but I'd be looking at fairly large populations per shard, in the area of 500 players and thousands of mobs/NPCs.

We'll see how it goes.

Disclaimers:

Comments / suggestions are welcomed.  Keep in mind this is nano-budget, I'm the only programmer / dev / designer (I do have a couple advisors), there is no art team (only what I can buy from art vendors and a few nice people have donated artwork) and I'm fully aware that the end result won't even meet sub-standard "B" grade games.  Still, it's fun to push the envelope on DIY game development.  Maybe someday I'll have something worth making it beyond beta... then again, maybe not.