Last Friday we went code-complete. That means, no touching the code except to fix something.
We had spent the past 2 weeks in the final QA process -- fixing bugs, typos, grammar, optimizing performance, and adding usability features. We had also used that time to continue improving the AI here and there along with tweaking the campaign missions, game mechanics, etc.
So over the weekend we went through the really hard tests. That is, making it as solid as it can possibly be. Running it on the absolute worst systems we could find. Trying to crash it. Having it run on its own (forced turn mode) to see if it would crash.
This is one of the reasons why I'm glad we're self-publishing. Total control over when and if it goes to manufacturing. No one to tell us when it has to go to retail. We had the luxury of being able to make sure we were happy with it. Plus, our experience in testing and debugging is pretty significant. After all, most game developers release one product every 2 or 3 years. Stardock releases new software it develops every few months in the form of Object Desktop and related components.
Of course, even with all this testing and testing and testing you know there's still going to be someone who runs into some problem. I recall reading on forums about Civ IV being "buggy". For me and I suspect for Firaxis's internal testing people, Civ IV was rock solid other than the ATI issue that was quickly resolved. That's the thing, with so much hardware out there, you never know what you're going to run into.
Since I'm an engineer I'm going to tell you the horror stories we ran into during testing. The marketing people when they read this will probably have a heart attach...
Win98, 2K, and ME..
"Why are there so many, songs about about compatibility.."
Most games these days only support Windows XP. And for good reason. If you're running Windows 98 or ME you should upgrade. Really. Stop reading this and do it. But we said we'd run on Windows 98 and ME and we meant it. But what a pain.
First off, the December Microsoft DirectX SDK says it supports Windows 98/ME. It doesn't. Not as far as we can tell anyway. We had thousands of beta testers and NONE of them were running Windows 98
But it did mean we had to make a special build just for Windows 98
Meanwhile Windows 2000 had a different issue with regards to how it handled critical sections. It too required its own EXE for installation in which we removed a background thread and put it into the main thread. It's a bit redundant now anyway. At one point in development, GalCiv II took like 7 seconds to go from clicking on the icon to seeing the title screen on a decent system. Beta 5 was slow like that. So we added a background thread to speed that up. But through performance optimization, we got the time down to about 1 to 2 seconds (i.e. very very fast). So the background thread isn't really necessary (though I must say I think some people will be impressed at how fast the game comes up).
Only about 5% of the beta group has Windows 2000. But we think plenty of buyers out there in absolute numbers are still running Windows 2000 so we wanted to make sure they were taken care of. The last problem we had to solve on Windows 2000 may not have necessarily come up because of the artificial conditions of the testing box. On one of the test boxes running Win2K, the game would crash if you clicked on New game too fast (the thread timing problem). Of course, how many 3 Ghz machines are running Windows 2000? On slower machines, it didn't come up or at least a lot less often. That's why it was hard to track down. Once we found that the issue was loading of models -- that the main thread zipped through to the main screen so fast that the semaphore controlling it wasn't working on Win2K, it was just a matter of making it not threaded. But definitely a weird issue since Win2K and XP are so similar something like that was unexpected.
XP, it's good to be the king
The people who really benefit from this are Windows XP users (well technically everyone). Because while we were futzing to make sure Win98/ME/2K were working well, the rest of the team was able to keep tweaking things.
For instance, once we nailed down making sure it was bullet proof on Win98/ME, I used the weekend to keep tweaking missions. My favorite campaign mission is called Apocalypse. In it, you have to take on the Dread Lords directly. And it had to be balanced just right. So I was able to tweak the Dread Lords to be just the right level of nastiness without it being overwhelming. I spent most of Sunday playing that mission at different difficulty levels to gauge it.
It also gave me time to tweak a bit further the AI weapon/defense adaption code. That 2 days between "code freeze" and "gold" made a big difference there. The AI is much better at adapting to the weapons and defenses of its enemies than it was before.
The AI is a never ending journey
As I get better at the game myself, I think of new things I want to put into new AI personalities. I would like to add some additional AI personalities in so that when people get their game and bring it home, they can look forward to the first of many free updates that keep adding new stuff.
One of the things I want to work on further is the planet building up stuff. But to do that, I'll have to wait until I see how the top players are playing the game. From the Metaverse, I'll start emailing the top players and see how they play. What do they build first? When do they purchase (quick build) something and when do they let it build normally? Which resources should be gotten first? Should I trade with minor races more who aren't likely to go to war with me or use trade as part of my diplomatic arsenal? I have opinions on these things -- strong opinions. And the AI reflects much of this. I know from the gamma testers that they certainly find it pretty challenging.
"It's cheating!"
I know I'm going to hear that a lot because I kept saying it when I'd see the AI do something that I didn't think was possible. One of the changes that has had a far greater impact on the AI's effectiveness is the new economic analysis system. Previously, (including GalCiv I) the AI looked its economy as a matter of how much to spend. It didn't touch taxation. But now I've plugged in the overall approval API into its analysis and it raises and lowers taxes based on a bunch of criteria. The result was unexpected -- the AI focuses somewhat more on getting high approval ratings so it can raise its taxes which in turn gives it more money.
But when I was playing it out of the debugger, all I saw was the AI with a larger economy with a similar population. That meant loading up the game in the debugger and trying to figure out the "bug" to keep it from "cheating" in the future. The result was a higher tax rate than anticipated and the AI going after those morale resources. Let me tell you this -- if you let the AI take and fortify either the morale or influence starbases all the way, you're dead. If you see the galaxy suddenly turn to their color, that's why. Controlling those galactic resources matter a lot in the long term because when they get built up, they can literally double your ability in a given area.
Now, if you set the AI intelligence to "Genius" or higher then it IS cheating. I make no bones about that. If you beat the AI at "intelligent" then you are beating me. Congratulations, you're better than I am at the game. At that point, play some Metaverse games and I'll likely contact you to learn from you. But in a single player game, the player needs to be challenged and if I can't beat you, then there's the option of just handing the AI more money. But that is the key -- the AI isn't getting free units or special attacks. We also don't start you in crummier galaxies or give you poorer planets. The highest levels are just AI gets more money. The reverse is true at lower levels. If you play at beginner, we're giving you money.
The Marketing really ramps up
We've had full page ads going in the major PC game publications in the United States for the last few months. Next month that'll expand to on-line ads, in-store ads, and an expansion on the existing ads with 2-page spreads and such.
Now, to gamers reading this, that's all to be expected. But the costs are amazing. A full page ad in a major game publication can cost well over $10k per month. An in-store ad, a good one, can cost $30k. Even the major gaming sites can charge pretty big bucks for this. There aren't very many indie developers out there. There are even fewer indie developer/publishers. And we're trying to be certain not to cut any corners.
Meanwhile, in Europe, we teamed up with Paradox to make sure the game is available in Europe at the same time. Paradox, makers of Europa Universalis and Hearts of Iron are kindred spirits. Initially it'll only be available in English but German, Italian, Russian, Polish, and other languages are in the works. It should also be available down under (Australia and New Zealand) around the same time but I'm not as up to speed on what's happening on that front.
So far, we've got orders from WalMart, Best Buy, CompUSA, EB, GameStop, Jack of All Games, Circuit City and a few others. Users who pre-order get the collector's edition which besides having a cool looking box get the tech tree poster, the GalCiv Desktop, and extra game content (additional pieces for the ships, etc.). Cost is the same as the regular. Users who have pre-ordered it from us directly can also get the same thing.
The initial availability is going to be 3-times what the first GalCiv was. We're very excited
The Reviews
The reviews always make me nervous. Very nervous. Two years of work and I can tell you the reviews on an indie game can make or break it. Now that we're not as obscure, we tend to get the better reviewers. Back in the days of Entrepreneur (our first Windows game) we'd be lucky to get assigned to a reviewer who would play the game. It was discouraging to have worked so hard on something and have some reviewer not even give it 10 minutes before writing the review. But like I said, things are much better now. By GalCiv I the major game magazines knew us enough to make sure that their solid reviewers checked it out.
Good reviewers don't necessarily mean positive reviews though. It just means they'll give it a fair shot which is all you can ask for.
GalCiv I got 4.5 stars from CGW and CGM and an 82% from PC Gamer and an 8.4 from Gamespot. GameSpy gave it an 87. So that's our baseline to see how we've done. 3 years have passed and the bar has been raised, so there are other factors to consider as well. This time, it's not "Hey, it's either this or MOO3". Civilization IV, the best turn-based strategy game I've ever played, came out 4 months ago. So in many reviewers minds (I know it would be in my mind) the question is "Is this a game that is worth looking at with Civ IV there?" That's a question I can't answer. I'm too close and too enthralled with Civ IV anyway. 
I can say that they're very different.
There and back again
GalCiv II took a major turn this time around. We wanted to make a game that could be a definitive 4X space strategy game. That meant incorporating features from the genre that are popular. Sometimes, games leave out feature -- good features -- from other games simply because they don't want to acknowledge that other games exist. I see that in RTSs and it drives me nuts. After playing Rise of Nations, it's hard to go to another land based RTS that doesn't have a concept of land control (at least for me). It's such a good idea. After playing Total Annihilation, it's hard to deal with units that don't fight back (or run away) when attacked.
Similarly, GalCiv II couldn't just pretend that Master of Orion didn't exist. Or that Homeworld didn't exist. Or that Ascendancy didn't exist. Or that Stars didn't exist. Or Twilight Imperium. Or Space Empires, or Starships Unlimited, or even the unreleased Stars Supernova. I don't claim to be a real game developer. Or even a real programmer. I'm a gamer who just happens to know how to make game ideas be turned into actual games.
So we could either pretend that GalCiv is the only 4X space game that's ever come out and expand on its existing features. Or we could try to reach out to the fans of all these classic games, many of whom are probably looking for a new game to try and see if we have something to offer them. That's not to say we just took features X, Y, and Z and crammed them in to have a bullet point. We didn't do that. Instead we looked at features we thought might make the game better and then looked to see if other games implemented it, how they implemented it, and how that might make sense in our game.
First GOLD screenshots
Here are the first GOLD screenshots..
This is a fairly late stage planet. In GalCiv II you can have your planets focus on a particular type of production (a lot of players asked for this). And if you're not building a ship, you're not charged for it which is why the (2) at the top (2 potential production but you're not being charged because you're not building a ship). That was one of the most common complaints from GalCiv I.
Planets are not just visually unique (we generated the high quality planets each game, so you'll never see this planet again). We also added things like artifacts, mineral rich areas, fertile land, and other things that give you bonuses (BIG bonuses) if you build the proper thing there.
There are probably three major technological innovations in GalCiv II that I can think of off the top of my hat. The first is the unlimited smartscaling technology. Play at any resolution (at or above 1024 x768) and have it all used intelligently -- 16x9, 9x16, you name it, the smart scaling stuff will let the game be relevant years from now. The second is the vector based texture development so that the textures can keep getting better as technology changes. So 3 years from now when our video cards have 2 gigs of video ram on them you'll be able to plop in GalCiv and hopefully have say 4096x4096 textures on things.
And the third is the per-race dialog system. I don't know how big of a deal it is to have each race have its own dialog. I do know that Paradox's translation team won't be sending me any birthday cards though. But we really thought of having a lot of this stuff would be fun. And it's done in such a way that we can keep adding more easily.
Some would probably argue that this is why user-designed ships aren't such a good thing. This is my "Awacs" ship. I used the ring to throw on tons of sensors.
Speaking of ship design, I wonder how many polygons this ship has. Heh. The game (as far as I know) has no limit to the number of things you can connect to it.
Here it is in action:
BTW, little thing but the display settings in the game let you control ambient light. So you can make space very dark. The lights you see on the ships aren't just texture noise. They're literally lights:
I turned down the ambient lighting so that just the ships themselves have their own light sources doing their thing.
I think I mentioned this before but we felt this was important:
You can click on a ship and press the details button (or hit I) and see what is actually on the ship and how much it cost.
Don't like the colors of the UI? No problem, change them.
GalCiv I's UI wasn't quite so nice..
Speaking of which, I've heard some people downloading the GalCiv I demo. I wouldn't bother. It's nothing like GalCiv II. It would be like trying to figure out what Windows Vista is going to be like by trying to fire up a copy of Windows 3.1.
One of the other things I think people will like is that you can zoom out with your mouse wheel until the map turns into strategic mode where the units are icons. You can play it like a war game from afar. It's not nearly as pretty.
I'm going to try to put together a quickie animation though because you can zoom all the way from this to looking right up close at a planet or ship.
So anyway, tomorrow the crew will start to go on vacation. I plan to take a day or two off, then catch up on email and then try to catch up in general! Plus spend some time with wife/kids.
In about two weeks, it'll be out. February 21st at a store near you.