Modelling for CA

CA is designed to be an open-source, free game inspired by OTA- but it is not OTA. In order to be free and open source, it needs to be free of OTA content, and that means new models, new story.

Models can be inspired by OTA designs, but they should be different enough as to not infringe on OTA intellectual property. OTA is 10 years old and we can afford more complex models following less simplistic designs. More importantly, they should reflect the function of the unit as it exists in CA.

Technical Specs

Spring uses .s3o objects, made with the Upspring program. s3o has no mesh deformation, animated parts must be seperate objects. 3ds (preffered) and .obj files can be imported into upspring and converted into .s3o.

Animation is done via .bos/.cob script (we have people who can do this).

To give you an idea of the polygon budget, the current Spherebot model is one of the most spammed units in the game, and is 1k polygons. Nobody suffers slowdown from too many of these on screen. You shouldnt use any more polygons than you need to, but there is no reason to use 6 sided barrels and wheels, or to not include more detail.

Textures are a single combined UVMAP, typically at 512x512 pixels. We tend to use DDS image format, though .png and .tga all work- if you arent familiar with .dds we can convert your textures easily.

Design notes

Models should reflect the units stats and role. Fast units should be sleek, high hp units armoured, expensive units large etc. Weapons of the same type should look similiar. A weapon that fires twice as fast could have two barrels, longer ranged weapons longer barrels, etc.

Everything on a unit should reflect a purpose. IE, a unit shouldnt have a radar dish unless it has radar.

Models should always have some kind of animation, a moving part, or at the least an effect such as that on the fusion. The movement should reflect function: A spinning dodad for no reason is rather superfluous, but wheels should spin, flaps can move, thrusters can rotate, solars can open/close, etc.

Some examples:

Metal Extractors should show extraction rate through the speed of rotatation/drilling/etc. They should also show their overdrive through a secondary effect, such as a glow.

Storages should show their capacity. I.E Metal storages could be open topped and have a shiny metalic object that moves up and down inside depending on the players stored metal. Perhaps even an overflowing effect when excessing. Energy generators might have a glow, electricity effect, etc.


Since there are no t1/t2 distinctions in CA, each lab instead represents a different propulsion method. Units from each factory should follow a relatively similiar style though there may be some exceptions (say, tick). The factories are...

Infantry Bots

Human-shaped, bipedal. Ensure that the joints have a good range of motion without clipping, and that the top half of the unit behaves like a turret, capable of rotating 360.

Light Vehicles

Wheeled vehicles. Ideally we should take advantage of suspension scripts, so units should be made with suspension in mind. This can be either ensuring the wheels have plenty of freedom to move up and down or even having the wheels come out on axles that rotate. Steering gives an extra option for animation if you wish to include it but its fine if units use independent drive to turn (IE, like a tank does).


Doesnt need much animation beyond turrets, but you may want to include propulsion systems such as a propellor (probably internal and jet-engine in appearance).


Ideally, could have some kind of visible propulsion: levitation pads that change direction with movement, thusters, fins, etc.


TBD. Amphibs arent finished.


Should have rotating thrusters that change orientation depending on whether the unit is hovering or moving. Can include flaps. Probably should not have pronounced wings to avoid confusion with Planes.

Planes (Aircraft?)

Since planes also do VTOL, should probably have engines capable of rotating. Can include wings that fold away when landed, flaps, etc.

Heavy Tanks

Tracks. There is the swapping-out-3-objects style, to give the appearance of animation (see mr.ds models), or if you want to go really complex you can use individual objects for each tread piece.

Spec-Ops Walkers

More mechy, bipedal. For arm, spiders, which should ideally follow a similiar design to the bipeds in the factory. For core, jumpjets.

Ultraheavy Mechs

Even larger bipedal mechs (we'll need two distinct styles for this and the spec-ops).

Some older, waffly stuff you dont need to read!

The graphics, sound, and all aspects of the interface exist to convey information to the player. Not only does this increase the players understanding of the game but it also increases his connection to the art itself (IE it means something concrete). The game should use a common language that is:

  • Intuitive. A player uses ideas from outside the game (the sci-fi genre, other games, real life) in his attempts to understand it. When designing the game itself there must be a way to represent a given gameplay element. Any 'invisible' parts of the game (special damages, etc) should be cut down or ways found to represent them. Numbers are helpful to the player and should always be accessible (unit help, tooltips), but colour, size, movement (a rotating windmill = wind speed, moving legs = movement speed etc) are more immediate and understandable for the player.
  • Consistent. Where information is reflected in a certain fashion in one place a player can grow to easily understand it if it is reflected elsewhere in the same way. As much as possible where it does not diminish meaningful complexity should be standardized- similar weapons and units should share similar aspects of design and appearance. This extends to the intuitive aspects of the game, but beyond them. Even where a player has no previous experience with an element of the game representing it in a universal fashion will make the burden of learning easier- they only need to learn a single universal law.
  • Differentiated. In contrast to the game being consistent, it is equally important that a player see the differences where those differences are present. It is important to think about what makes something unique and then find a way to show it- tracking, arcing, and guided missiles, for example, should all need variations in their effects to convey these differences. On units, three kinds of differentiation are important:
    • Background differentiation. A unit must be bright and visible on the map, among corpses etc.
    • Side differentiation. A unit must have enough teamcolour to allow you to tell who they belong to at even the most cursory of glances.
    • Unit differentiation. A unit type must look different enough to be easily distinguishable from other units.

You can see how these blend into and cross over with oneanother, towards the single purpose of giving the player meaningful information.

To these ends:

  • Large things should cost more
  • Fast things should be sleek and graceful.
  • High HP things should be bulky and armoured.
  • AoE's should be reflected in the graphical size of explosions.
  • Damage should be reflected in the intensity of the weapons fire and the size of the weapon itself
  • etc

Note that these guidelines in some areas also apply in other aspects of the game, including balance and design.

Some waffling on the nature of video games as art

Art in any medium relies on a set of symbols or a language to convey meaning, be it the English language, the understanding the red means danger/anger/passion/aggression, or a pig representing gluttony. Art that fails to use an understandable language fails to convey a meaning. Some art is emergent, provoking a deep and not fully understood response that the viewer is unable to articulate, that proves to be deeper and more complex as the viewer examines it- but this still relies on the language the artist is using, consciously or otherwise. Other times art (especially abstract art* and music) relies purely on an aesthetic sense, and conveys no information.

Games as an artistic medium are defined by their interactivity. All aspects of the interface, UI, graphics, sound, etc should be in service to that interactivity. In the end, even primitive graphics and sound can elicit positive responses if they reflect the meaningful gameplay data of the underlying game. The art in the game should be beautiful in an aesthetic sense in its own right, but the beauty will always be in large part due to the underlying simulation. A nuclear explosion is spectacular to a large degree because you understand that it does absolutely massive damage over a huge area (I always felt disappointed by nukes that do piddling amounts of damage in many games, even if they were graphically spectacular- TA's nuke excited me because it really did massive damage, despite being graphically simple).

The most attractive games are ones that can be beautiful in the fashion in which they convey their information. Homeworlds ship trails are a perfect example of this, adding breathtaking beauty as well as allowing you to see the speed, direction, location, and team of units that would otherwise be impossible to see.

* A game based on the principles of abstract art would be fascinating, if it were even possible. Not for CA though. :P