It’s Test Build time! We recommend you uninstall your previous test builds before installing this!

Quite a bit of new stuff in here, but much of it isn’t immediately apparent. Several big changes happened under the hood; examples include the way maps are loaded and the complete reorganization of the material palette.

Some of the many other changes and improvements:

• SLOW CRUSHING OF ANY GIBBABLE OBJECT

• Vastly improved collision detection and response, both against terrain and other objects. Still not perfect, but much better
• Physics simulation and graphical framerate decoupled! This means no matter what your FPS is, the physics will behave EXACTLY the same. No more sinking into terrain because things are slowing down.
• Completely new sets of small gibs and particles
• AI mode control – hold down the switch actor button, and then press a direction to select a different mode for the Actor you’re switching away from
• Your team’s actors are now ordered by their horizontal positions in the scene. Press right switch button and you’ll get the closest guy to your right, for example.
• Much fewer intersection problems with the terrain. Things are not as likely to get embedded and weaken bunker walls.
• ‘Tunneling’ through terrain much less likely to happen
• Enemy AI ships won’t appear directly over your base any more, only on opposite half of scene.
• Long wielded devices are prevented from sticking into terrain; can’t poke a long cannon through walls anymore
• Scene terrains are now simple RLE Bitmaps instead of using weird contour files. Paint your terrain with the material palette and save as 8-bit bitmap with RLE compression, and presto you have a terrain
• Only loads all data once on startup of game, so it’s much quicker to reset or restart the scene.
• …Lots more, going to fill out later by looking at source control logs

What this does NOT have:

• Campaign mode.. saving it for the real game.
• Scene Editor.. ditto.
• Gibbable ships and new ships, working on these, some may appear in next test build
• New guys and guns.. may save for real game too
• Ability to force reload of all ini’s without restarting game, will be in next build.

Also, old mods are unlikely to work with this build! We appreciate the huge amount of effort put into all the modifications for previous test builds, but we just can’t afford to spend time dealing with legacy support of not-even-beta versions! The good news is that mods should be relatively easy to bring up to speed with TB14. Later we will try to post some helpful technical outlines of changes between 13 and 14.

I know some of you are really getting anxious about a new build; here’s the situation:

Up till now the test builds have been a great way for us to evaluate what works and what doesn’t, and with the great community feedback and support we have an idea of what you like too.

Even though the tests have been fun to play in their own right, they were little more than half-finished game engines with some temporary data thrown in to try the different systems out. They did not really represent the larger gameplay design we have planned for the final product. Sure, the gun battles and basic elements are there, but the base building phase and more emphasis on strategy and tactics over twitchy action wasn’t.

Now that we have a framework and and idea of what’s fun, it’s time to build the real game. The project is at a point where we’re taking apart and rebuilding a lot of areas, and replacing most of the old temporary test data with fresh new art assets. It’s just not conducive to putting out tests as frequently as some of you have become used to.

Another reason for not putting out new builds while we’re changing things so drastically is that each one would break all the mods that are already out there.

So, while I will try to keep this devlog a little more frequently updated on things we’re working on, don’t expect test builds nearly as often as before. There may only be one or two left before we put out a version that really starts to look like the final product.

If you can’t handle the waiting, then why not submit your email address above, and give it a break. You’ll get an email every time anything new is announced.

Soooo… this whole week has been spent tearing my hair out over the core physics code. Everything else is being held up by these issues that I’m trying to resolve with the lowest level collision code. Being a physics-based game, Cortex Command’s gameplay is entirely dependent on having this robust foundation, or things really go to pot.

A clear example is the ‘tunneling’ problems, where actors could glitch themselves through the terrain, no matter how hard or thick the materials. The AI-controlled skeleton actors loved to do this in previous test builds. But no more! Although there are still other ways to compromise the terrain through glitches, the major ones have been fixed:

Note how there’s no more tunneling, so the AI enemies have to pile bodies to get over those walls

Also, a lot of time and effort was spent experimenting with having invisible physics springs pulling things apart when two objects started to intersect. The initial method I tried failed miserably, resulting in things being very squishy and mushy feeling. I’ll work more on getting the MO-MO collision responses working better, and see if we can’t get a new build out soonish.

There’s a reason for the unusually long gap between builds: wedding here in the tropics tomorrow, then honeymoon somewhere else. I’ll be back in business in a couple of weeks – although I’m still sneaking in some work on the laptop when ideas for the new collision response algorithms strike!

Until then ~

UPDATE: Thanks everyone for the well wishes and very kind comments! We’re done and married, and will enjoy the relaxing latter half of our two weeks now. 😉

It may seem like things are slow going, but I assure you that we’re being more productive than ever behind the scenes here at Data Realms HQ. Arne has been cranking out an incredible amount of new art assets for pretty much every aspect of the game.

I’m talking several new craft (with complete gib sets), humanoid actors, entire sets of bunker modules, etc… It’ll take a while to get it all implemented (read: not by next build), but that won’t stop us from teasing you in the mean time! 😉

So we thought we’d push a build out this last weekend, but then we came up with this and that to add, and then all of a sudden there’s a serious need to overhaul the core physics code! In essence we’re going to try to squash some of the old core physics engine problems like sinking objects and inconsistently mushy terrain.

So, along with a whole lot of other major improvements (e.g. scenes are very easy to make now, no more silly contours), we’ll take some more time to really get into overhauling the physics integrator and collision response code.

In the mean time, here’s a little surrealistic teaser:

UPDATE:
For the curious, this describes one of the things we’re actually doing to the physics engine:
http://www.gaffer.org/game-physics/fix-your-timestep/
It, combined with a lot more bugfixing and tweaking in collision response code will help prevent MovableObject-Terrain sinking, as well as MO-MO interpenetration and consequent explosions. We’re also considering swapping out the integrator, but not sure yet if it’s something we can retrofit into the current way the engine works.

Things are getting better in terms of proper control over your team:

Before, when you pressed either of the actor switch buttons, brain control changed immediately. Now, when you press and hold either, you get all that stuff you see here overlaid on top of your current and adjacent team members.

When in this mode, you can use the directional keys/pad/stick to change your currently selected AI mode before the actual control switch happens. Only when you let go of the actor switch button you pressed earlier, does the actual switch happen.

This way you can still very quickly cycle through your guys (by tapping the switch buttons) – or conveniently change their AI modes as you go through them, by briefly holding the switch button and hitting a direction corresponding to the new AI mode you want.

Also, all your team members are now sorted by their horizontal position on the map, so it’s more intuitive to switch between them. When you hit the right switch actor button, you’ll get the guy to your right, and vice versa.

We’ll, we’re back to programming after the post-conference fatigue. Here’s the photos from last week. Good times.

Anyway, work is well underway on the next test build which will focus on the different selectable AI modes for both the humanoid and flying actors.

Before switching away from the currently controlled one, you get to choose what behavior it will have when it is on its own. Some example modes are sentry, patrol, brain hunt, and gold dig for the humanoids.

If we can make the interface intuitive enough, it should be fun to play too!

Spending the week in San Fransisco at the GDC, making friends with fellow game developers and and absorbing the wisdom. More photos to follow!

Over the weekend we had some friends over to really try the four-player game. We learned a lot and the to-do list grew accordingly. Here are some pictures:

Note the crazy cyberpunk nightmare on the right… the Tablet PC is what CC was running on, but it only has two USB ports, and we had no standalone USB hub, so we had to use the built-in one in that bottom monitor, along with a PS2 PC adapter (which allow two Dual Shocks to connect through a single USB port). This way we managed to get two wired Xbox 360 controllers, and two PS2 Dual Shock controllers going on the same laptop.

The great news related to this is that Microsoft has finally released their new wireless PC adapter for its regular wireless controllers that come with the Xbox 360!

With this, you just have to plug it into a single USB port on any windows machine, and suddenly you can connect up to four wireless Xbox 360 controllers, for the perfect Cortex Command multiplayer setup! I wish we had one this weekend; it would have made half of that mess to the right disappear. Alas, our orders from Amazon didn’t arrive in time.

Finally, here are a couple of short videos from the weekend. The first is from our local test at Data Realms HQ, and the other is from our tester Nate Sabanski and his friends’ LAN party up in Canada. Good times!