January 13 (A Little Bit Closer Now...)

Well, it's been an interesting week. With roughly two weeks remaining before my deadline, I've been in crunch mode trying to get things ready for the final closed testing session. My goal by the end of the month is to have all six maps in a playable state and all known bugs fixed. Essentially I want to get as close to beta as possible before the deadline. In real world terms, that means finishing two maps, some minor GUI work and doing a lot of testing.

With that said, I set out this week to get the next map, SegFault, to a playable state. This meant creating an entire map from scratch: terrain, level geometry, etc. Given how short on time I am, I had no idea going in how I wanted it to look or play. The only thing I did know was that I wanted to try something different gameplay-wise so that winning required some unique strategies. I try and follow this guideline every time I make a level but it's getting tougher both because of the lack of time and because there's only so many things to do differently without the risk of breaking gameplay or making something that's not fun.

I divvied up the week in such a way that I had roughly two days for each part of the mapping process: two days for heightmap/terrain design, two days for level geometry, and two days to put it all together. The last day I decided I would use for testing.

By Tuesday I had the terrain done but only after scrapping it several times and doing multiple rounds of revisions. What's interesting is SegFault resembles The Breach in that the terrain is very open with small, rolling hills. It's very organic and natural, unlike most of the other maps which almost resemble pieces of a circuit board. I originally shied away from that approach but it just seemed to fit with the overall theme of the map so I went with it.

Wednesday and Thursday I worked on the team bases. Again, coming up with ideas for these is really starting to be a problem. I think I've got one more design in me before I'll need to go in a completely different direction. One of the problems is I decided at the very start of Bit Battles that there shouldn't be massive interiors. The game was, and is, about players fighting it out outside. Large bases with four walls and a ceiling just don't lend themselves well to this type of game. Unfortunately, there's only so many unique bunker layouts you can make without getting a little repetitive.

Friday I put the pieces together. I played around a bit with the locations of the team bases. There are two bases for each team (for a total of four bases on the map), one in each corner. However, unlike the other maps, the teams spawn on opposite corners of the level. This means that the Green team spawns next to and within viewing distance of the Red team (and vice versa). I'm hoping this will result in some pretty large battles near the spawnpoints.

Objectives, etc. went in easily enough. Only three Capture Pylons in this map, one in the center, one at the north end and one at the south end. Another new thing I tried was putting all of the powerups near the center of the map. I'm hoping this puts some extra value on the north and south Pylons because capturing them will mean gaining access to all of the powerups which in turn will make defending the central Pylon much easier.

SegFault came together well enough and in record time (a mere one week). I will say it doesn't have as much polish as the other maps but that's scheduled for after February 8th. It wasn't the most fun map to work on simply because of the breakneck pace but it should be fun to play.

In other news, with SegFault at a playable state, I spent most of Saturday testing. In the process I discovered a serious issue with vehicle collisions that I had never noticed before. Ramming into walls a very low speeds catapulted the player's hovertank into the air. This was particularly unusual because hitting walls at high speeds resulted in the expected behavior (a hard recoil/bounce back effect). After a few hours of wasted effort I gave up on trying to fix the issue in code. Instead, I opted for replacing the hovertank's collision hull with a simpler design. This had the added benefit of improving the performance of collision detection because the new hull is around 10 polys whereas the old version had ~250 polys.

Next week I'll be repeating the process in an effort to get Access Denied up and running. Any time left over will, as always, be spent on testing and bug fixing. Until next time!