Holy cow are things going fast now.  Up till now, it's been engine building. And even now, it's still engine building. But now we're starting to get into gameplay features which are far more fun.

Fleets are in. Totally in. Of course, at the AI guy, I have to get the AI to effectively use them.  I just finished working on the AI's handling of the planetary improvements.  The planet surface code is all done with STL which I'm a bit sketchy with.  So my first crack at it was a bit weak.  This time, it's better. Not great. But better.  The AI doesn't know how to upgrade its planetary improvements yet. I will have to go through and do that.  That's going to be interesting to do.

It's easy to write fast AI code. And it's easy to write really smart AI code.  But it's really hard to write fast and smart AI code.  And that's the conumdrum I always have.  For instance, in GalCiv I, my biggest beef was that the AI didn't update its ship orders between turns. That is, if you moved your ship, the AI couldn't attack it because it "moved".  That was because of the time it would take for the AI to "update" its ships.  This time, I'm going to fix that.  CPUs are faster and I'll use that extra processing power there.

Another thign I'm going to do is completely re-do the way the AI goes to war.  Fleets change things a lot and make things potentially easier.  I have my fears about the balancing of the weapons techs still but since I do that too, at least I can tweak that myself as I discover exploits. 

The way the AI did war in GalCiv I was based on a military concept of projections of force.  I would identify a sector I wanted to militarily control and then figure out how many ships I needed to control it and send them in there.  The problem came in when trying to coordinate my groups of ships.  In a tiny or small galaxy, the AI ships woudl come in groups quite nicely.  Since I was running in the debugger (Which is much slower), I tended to play the game on very small galaxy sizes and the result was nice.  But as you moved up in size, the ships had more and more of a problem getting to the target at the same time.  As a result, ships would come in dribs which looked like, to me and our critics, as the usual "death march" of lame AI that we've seen for decades. 

But with GalCiv II, I've got better tools.  First, Altarian Prophecy introduced the Rally Point system.  The AI, thanks to the "not cheating" design, can make use of the same stuff as humans and vice versa since the engine doesn't make a distinction between players.  So the AI can set up rally points on ships to move together.  And now with the fleet code, I can group them together more easily.  I am also going to add in code so that even if I don't build fleets (due to lack of logistics units for instance) I will set a "slower" speed for ships that can't keep up so that when you're attacked, you know it. They come in waves of massive fleets.

I'm also going to put more work into having the AI clue the player in on what it's thinking.  There's a lot of calculations the AI does that humans don't even know about that could make the game more exciting.

We're starting to move forware more on the tech tree.  I hope people like our philosphy on it.  In GalCiv I, it was based on the OS/2 version of the tech tree which was based on the tech tree not being tightly coordinated with the rest of the game.  In GalCiv II, the techs  tie into something in the game. So it's much more linear.  You have techs called Plasma Weapons IV for instance (as opposed to "Advanced Phase Induction" or something like that).  You research a technology because you want something specific out of it.  So there's a lot fewer essoteric techs in there.

I sure hope people like the fleet battles. I think they're cool. But I think some people are going to think "Why can't I control my ships?".  It's a playback of the battle, not a tactical battle element.  I hate tactical battles in strategy games.  Hate them.  The reason is that they're so time consuming. I am the guy who always pressed "Auto" in Master of Orion.  But I also knew that I wasn't playing the game optimally by doing that because the AI is never going to be as good as a skilled player at tactical combat.  Our combat system is essentially like a forced "auto" of the MOO 2 thing except with 2005 3D graphics.  If you watch a battle in cinematic view, it's a lot like watching a battle from Star Trek TNG.

Speaking of MOO, people who only played MOO need to remember -- GalCiv isn't Master of Orion.  GalCiv for OS/2 was a contemporary of MOO 1 for DOS back years ago.   GalCiv I was in public beta back in 1993.  It's a different game.  As much as I like the MOO series, I don't want GalCiv to be MOO. 

There were 3 points from MOO1/2 that I particularly don't want to emulate (and bear in mind, I'm a huge MOO2 fan, I played MOO2 in the delivery room waiting for my first born -- true story):

1. MOO 1/2 always ended in a genocide race. Each side collected huge fleets and then just wiped out each other's planets so it was a race to see who killed off the other.
2. MOO 1/2 had tactical combat which I really don't like in a strategy game.  If I want tactical combat (And sometimes I do) then I'll play Homeworld which does it very well.
3. MOO 1/2 had starlanes.  I like free-form maps. It worked for MOO since you're parking fleets around.  But in a free form map like GalCiv, you have to limit fleet sizes and distribute up military power.

So while I sympathize that many MOO 1/2 fans are looking for a MOO 4, GalCiv isn't designed to be it.   It has its own set of strengths and weaknesses that make it a unique gaming experience.

Anyway, Beta 4 should be out before Halloween.  It'll be still a beta (read: buggy and incomplete) but it'll be well on its way to final by then. It's not quite fun IMO yet, but it's getting there.  Beta 4 will be open to those who pre-ordered the game.