Since most everybody else has already mentioned the gameplay improvements, I'll go a bit to graphics:

• More visible damage to the ships. I wanna see the scars after the fight.
• More shaders. Add some reflection, self-shadowing, real time ambient occlusion.
  
• Procedural textures. If you use shaders to calculate them during rasterization, you won't need to pre-render it to a texture.
• Keep (and improve) the ship editor! That's one of the best parts of the game .

EDIT: The third edition of GPU Gems had an updated article.

1: Better UI. Civilization IV should be the minimum standard.

2: Combat should not be random. This is a personal pet-peeve I have with all TBS games.

3: Combat should not be reduced to simple "bigger number == win" type of things. Or at least, not always. Ship construction should feed into that; a fleet composed of the right ships with the right abilities should be able to defeat other fleets that are bigger/have bigger numbers/etc. Ditch the 3 attack vs. 3 defense nonsense in favor of something else.

4: Ditch the resource model entirely. Design a new resourcing model from the ground up that uses equations that are:

4a: Visible to the user.

4b: Don't require higher than a 5th grade education to use.

5: Trade needs to work better. A smaller nation should be able to specialize in "trade" (not the GC2-style of trade that just creates money for both sides) to the point where they can get another large nation dependent on trade with them.

6: Ditch StarBases. Replace them with something that isn't a chore to build.

7: Random galaxy generators need to be programmed to make sure that every side has some decent territory. Alternatively, expose (preferably via Lua script) the ability to allow users to add galaxy generator scripts, which should be able to do this themselves. They should also be able to expose UI elements for properties.

8: Planetary Invasions should remain relatively simple (in the sense that we should not have to build different types of ground troops, etc). However, it is not good that a space fleet is entirely useless to invasion. Ships should have modules that aid in planetary invasion (balanced by the fact that such ships are weaker in regular combat, since those modules won't be attack modules). Planets should be able to build defensive buildings that don't consume precious tiles. Tiles are a vital resource, and you cannot justify consuming them just for defensive purposes. You should be able to build a "fortress world" that is neigh-untakable without the enemy investing heavily in it. But at the same time, there should be a strong cost associated with bunkering a world so heavily. Just not a tile-based cost.

9: Develop a combat/ship construction system that encourages dual fleet building. That is, having defensive ships and offensive ships, and having these two kinds of ships be separate ships. In GC2, the same fleet you use to attack will also be used to defend. That is not as it should be.

10: Obey the Law of Small Numbers/Law of the d20. That is, a +1 bonus should always be significant. A +3 bonus should be substantial, and a +5 bonus should be monsterous. Anything that gives a bonus should give a significant bonus. That means that things like anomalies that give out lots of little bonuses have to go away. Giving out fewer, but more significant, bonuses is better.

11: Never use decimals. 5th grade math, remember?

12: Ditch the Twilight of the Arnor-style tech trees. Instead, have fewer tech trees (more races share the same tree), but have more distinct trees. As in, have some trees that are more like Civilization-style trees (lots of requirements and inter-dependencies), while other trees are more GC2-ish (more focused, but more techs overall). Best-case would be 3 well-balanced trees. Separate tech tress should have no techs in common; two races that use different trees have fundamentally incompatible technologies.

13: Game-changing techs. Gunpowdered in Civilization changed everything. If you didn't have gunpowder, you were fundamentally behind the times, and likely to get steamrolled. Same goes with Internal Combustion. Have certain techs that dramatically alter the balance of power, such that researching them quickly can be a viable strategy (the cost for doing so being that you're weaker in other areas). Something for GalCiv would be like the ability to create wormholes, allowing your fleets to jump to various locations, thus reducing your dependence on your ships' hyperspace movement speed. These should be tech-tree specific.

14: Ditch the espionage system. Replace it with something that doesn't suck.

15: Develop a planet quality system that ensures that every planet is viable. This isn't Civilization, where terrain is everywhere; if a planet is colonizable, it should be something that you can actually use. That is, there are no bad planets, merely planets that you can't colonize yet.

16: All planets should eventually be colonizable.

17: Follow the Civilization IV design mantra: reward instead of punishing.

18: Either remove government systems entirely or make it more interesting than just a bonus for researching certain techs.

19: Ditch the UP. Instead, institute a more generic system, where various factions can set up unions among themselves.

20: Tech trading. Allow the ability to trade the fruits of a tech instead of just the tech itself. This allows you to choose to prevent the other guy from whoring out tech. Of course, whoring is lucrative, so trading the fruits alone is not as valuable as the tech itself. Also, they can't build on fruits alone. Tech trading is only possible between races that use the same tech tree. For inter-tree trading, researching certain techs allows you the ability to trade special things that are not found on the other guy's tree. You can't use them yourself, but it allows them to gain some of the flavor of your tree.

21: Each of the tech trees should have a unique superweapon. These superweapons should all work differently (one can work like Terror Stars, but the others shouldn't).

22: No tactical battles.

23: Ditch the gampaign. Instead, integrate the game's storyline into the gameplay. Look at Alpha Centauri and how they integrated that plot into the normal game.

24: Ditch asteroids. There should be colonizable asteroids, but those should functionally act as regular colonies.

25: Revamp the event system. Events should not take place instantly. The Dread Lords event is really the ideal; it happens, and you need to deal with it, but not necessarily right now. AI needs to be better able to deal with events.

26: Ditch influence. Replace it with a better alternative-planet-capture mechanism that isn't dependent on being able to militarily protect StarBases while the computer rolls dice to see if a planet flips.

27: Ditch diplomacy. The stat, not the concept. Replace it with nothing, as gaming the AI is a fundamentally bad idea.

28: Ditch surrender. Instead, allow civilizations to become vassal states. These are states that are supposed to do what you tell them to do. You tell them how much trade they must do with you (and where to do it, if location matters). If they fail to fulfill their obligations, they lose their vassal status and become enemies again. This allows lots of cool things, like the Drengin becoming vassals as a temporary measure to rebuild their army. If you don't "tax" them enough to keep their army small, they'll grow strong again, and possibly overthrow your rule.

29: Ditch fighters entirely. Any ships you build should be of "corvette" size or bigger. "Carriers" would just be ships with special carrier modules that spawn a "Fighter Fleet" ship at the beginning of combat. Even if the ship is destroyed in one battle, it is considered reconstituted in another (just as ships don't run out of missiles).

30: Ditch range. Replace it with nothing. It was a bad idea.

31: Avoid techs that are "What you had before, only with a +1 bonus." Instead, develop a resource model that allows later "manufacturing" techs to improve manufacturing in different ways. Not just a better building that does the same thing only slightly more.

32: Ditch morale and population entirely. Replace them with nothing. That certainly has the virtue of never having been tried.

Um, FYI: in space, there's no "ambient" light.

In real life, there's no "ambient" light, either - it's simply the effect of light bouncing around a gazillion times.

Many of the ships from the Show off your ship designs posts are fairly complex, and at least a rudimentary illimination pass would, IMHO, improve the accuracy of the lighting.

I don't know if you've noticed it, but when you add a "light" to a ship, it does not illuminate the panels and surfaces around it - it's pretty unrealsitic. Even if they don't do anything very complex, I think they should at least do something to enable the lights to actually light up the ship.

Totally agreed. The anomalies, as currently implemented, just seem to be a hassle, and since there is a lot of anomalies just adding a small effect - and often a small effect to a fairly useless ship, I might add - they just don't add any variety or depth to the game.

Here's my suggestion: All starbases are highly specialized - and all starbases are already upgraded.

The influence starbases are already max influence, and that's all they do, period.

The military bases are already max weapons and armor, and that's all they do, period.

Same with all the other types of bases.

That's my suggestion.

They've tried three times. If you have a suggestion that works, we're all ears.

Agreed. Frankly, the current system feels as if there's no control. Which is bad.

• It's too centralized. While a centralized governing body is a great utopian vision - GalCiv 2 isn't simulating utopia. The various factions are fighting each other, and it doesn't make sense that there is only one entity and that all factions are a part of that entity. In real life, we have the UN, EU, NATO, and all sorts of other organizations and treaties.
  
• Apparently only the AI can create proposals. I think the player should be able to create proposals.
• More proposal types.
• There should be the option to vote "refuse proposal," or some other "keep the status quo" type option.
  
• It's silly that you can't trade without it.

Agreed.

Agreed.

This sounds like a great way to implement carriers .

You mean the combat viewer?

Disagree, I like it. It gives me an overview of the battle and some useful statistics. You can already disable it if you wish.

Agreed. There's nothing more annoying than launching a battle with several ship types only to discover only half of your fleet can make it to the battlefield.

Yes, trial and error diplomatic options are only going to make the player abuse them, since the AI wants sooo much for everything, especially at higher difficulties.

Instead of making the AI rip the player off, make trades even at any level... but with a few exceptions. If you want a tech, the AI will come back and say what it wants of yours. If you offer a tech, the AI will say either it doesn't want it, or offer you a tech it feels is fair trade. If you need 500 BC, the AI will tell you what it wants in exchange.

I would much prefer the offer and ask method. Because the way the diplo options are now, it's basically a screw-fest. Trading worthless Influence points for everything valuable the AI has, who would actually ever make such a trade like that, no one.

By thunder...I will figure out a better espionage system.

When I do so I'll post it.

Yes, but due to the empty nature of space, you're only going to one one or two bounces before the light simply disappears. Hardly worth the effort of modelling. The ambient occlusion stuff is designed to deal with an environment that is creating diffuse interreflections. Having a ground plane that reflects light upward is a big part of that. Space has no such thing.

In space rendering, ambient is about being able to see stuff, since space is very dark. It isn't about modelling any real-world phenomenon.

Well, the lights aren't actually lighting anything now, so having lights act as light sources would be a necessary first step.

What blasphemy is this?!

Oh, my bad.  Carry on.

Sole Soul ist Grammatik Führer!

Does anyone else find it a little odd that SS just happened to have that picture laying around for just this moment?

Just because something is simplified does not mean that it isn't deep or is "dumbed down". The point is to be able to make an intuitive economic model that the player can actually predict the effects of, without arbitrary unknown rules.

Take the tax structure for example. The tax vs. morale function is a piecewise function; increasing the tax rate across certain boundaries suddenly and dramatically increases morale. This is bad game design because the player was not informed of this (the function is hidden) nor is it a function that is something the player can compute. The only way you can find this stuff out is with a lot of guessing and checking.

A player should not have to poke at values and see what happens in order to make plans. If a player is trying to balance out morale to get the highest taxes possible, the above kind of function exists only to thwart him. How many morale buildings it takes for a planet to sustain a tax increase from 58% to 60% is not something that should be hidden from the player. The player should not have to build a couple of morale buildings, change the tax rate, and see what happens to morale; they should be able to apply the various functions in their head and predict the outcome.

The game then is about executing strategies and finding out which strategies work and which ones don't. To predict things, the player should not have to get out graph paper and plot some stupid arbitrary function that involves advanced mathematics. If a linear-least squares fit is necessary to predict the effects of various changes, then the game designers have failed.

It also turns out that GalCiv is a game. A game must be fun and enjoyable. And it is much more enjoyable to get meaningful bonuses than to have this nonsense with small accumulations of negligable bonuses to add up into something more by the end. And the reason I suggest limiting it to fifth grade math is because that's what you can do in your head. That makes it something you can predict without having to break out graph paper.

I'm not against terrain, but asteroids as they currently are implemented are stupid. To use them, you have to build a ship that has no other purpose than to upgrade them. This ship takes a set quantity of time to upgrade them to a certain level.

Further, asteroids are starting to dip into Civilization-style terrain management territory, which is not the right way to do GalCiv resource management. GC2 already has tiles on planets; having these off-planet resources just over-complicates resource management and city production.

Besides, in a space-based game, there are far better ways to create terrain.

Nonsense.

What is population in most TBS games? It's a limitation; a restriction on the capabilities of your empire. Basically, it forces you to spend time waiting before your cities can gain their full capabilities. In Civilization, it is a direct limitation; population = number of tiles worked = more resources produced by the city. In GalCiv, it is an indirect limitation; population = money = more resources produced by the city without it going into debt. But in both cases, it acts as a time-based friction on your ability to use cities.

But we already have time-based frictional forces in GalCiv. Namely, the time it takes to build buildingins on a planet. That takes a while. If you make buildings more complicated, possibly with multiple buildings per tile that have certain interrelationships or something, you then increase the time-limiting force of building stuff on a planet. So then, what good is having population as a time-limiting force? That's two forces doing the same thing; it's needless.

That doesn't mean you shouldn't have money. But money shouldn't be based on a time-limiting force, since we already have a time-limiting force to begin with.

Not only that, in GC2, population just doesn't matter for very long. It's an early-game force, but once you've passed the colony rush phase and have manicured your empire properly, every planet is at the population cap. So for the vast majority of the game, each planet gets X money (modified by tax buildings and taxes, of course), period. At least in Civilization, city growth and population is a meaningful mechanic throughout the entire game.

It will have a negligable impact on the rendered object compared to pretty much anything else. Remember: most of the time, these ships will be maybe 64x64 pixels in size. No point in spending a lot of time on something so small.

On the contrary, that is exactly what simplified means.

Or they could do the sensible thing: adjust the tax rate first, see the effect, then plan morale buildings to fit the drop they already observed. And yes, many/most of us CAN do that in our head.

The tax model is a simple linear relationship with preset jumps at the 9-0 threshhold. A linear regression is possibly the worst thing you could try to predict the outcome.

Yes, there are two time constrictions, but both are easily circumvented. If population is a problem you can ferry people from more populous worlds. This requires production elsewhere in your empire, and dedication of resources that could be doing other things. Buildings can be bought, not built, so if money is freely available when a colony is settled, the colony will be at max production at (PQ-1) turns. The bottom line is that neither time restriction is as effective separately as they are together. Unless you are willing to get rid of rush buying, of course.

I disagree. Smaller incremental bonuses - especially empire-wide bonuses - prevent massive game-changing events. Getting a 20% econ bonus from one anomaly could instantly change an insolvent empire to a fully funded one. Getting the same scattered over ten bonuses and 30 turns is not nearly as game-breaking. Perhaps the bigger bonus model might work on small maps, but even there it would be pretty unbalanced.

But the function for morale is even more ridiculous than the piecewise tax function. You can't just look at an planet's 80% morale and go, "I need 4 morale buildings that eacn provide 5% morale" because the equations are non-linear. You can only keep building morale buildings until the morale problem is better. Or you solve an equation that you don't have access to, likely by plotting it on a graph and drawing lines between the points.

It's still non-linear over the range of input. It is not a simple function that you can learn quickly, nor will you be able to predict what happens at 61% if you've never gone above 58%.

But massive, game-changing events are good. Massive, game-imbalancing events are bad.

The point is that having 2 time-based frictional forces is redundant. Time is one-dimensional; if you need one of them to last longer than it currently is, make it last longer. If you want a player to have to do multiple things to speed it up, then have each method cap out at halving (or so) the time constraint.

Yes. And this is a good thing. Especially considering how much money that would take.

It is good because it allows the game to play differently depending on how far you are into it. In the late game, you can afford to skip the time-delay step entirely and go straight for full production. But even then, it ain't cheap.

And I would point out that if you can afford to bring a late-game planet up to full power in PQ-1 turns, you have enough surplus money to pay for that planet until it starts paying for itself. That makes the population constraint redundant.

Yes. That's why anomalies should be removed. Or there should be far fewer of them, and the bonuses they give should be one-time things (extra money, extra research, etc), not persistent bonuses.

And to be honest, I'm not a big fan of global persistent bonuses to begin with. Like world wonders, they create "Rich get Richer/Poor get Poorer"-type sitations.

Unlike games like Civ III or even Civ 4, among others, GalCiv seems to go on the route where you can be only 1/5th (or less) way through the game and you can know with 99.9% certainly if you are going to win. I played Civ III quite a bit back in the day, and generally it was quite a rough ride, since as was mentioned earlier, gaining a new military tech like gunpowder was not crazy powerful, but would provide that civ with significant advantages. It seems that GalCiv2 is never a tense-'will i be able to stay alive and come back' type situation, it's always a landslide victory. It is slightly not so much in my current Suicidal Game, but I already know after 1.4 years that I will definitely win, with no chance of defeat. And I'm not even playing with the x-factory or x-research strategies, I'm playing normal build w/e on my planets.

Something needs to be done to ensure it is not soooo easy. Such as, DONT let IP's be tradable, period. That way, you can't get tons of tech and gold for worthless IP's.

Hopefully they will fix the AI and make it expand to the fullest extent. If they ever fix this in GC2, it would make the game much much more difficult than it is now.

Make it more difficult to take worlds. Right now, taking a planet is crazy easy. As far as the 'click the mouse button to get a good roll' during the planet invasion screen goes, that is stupid. It should be even odds (or close to even), and depend on the civ's soldiering ability that will decide. So to take a planet with 20 billion people (even if they have a low soldiering), you better bring lots of transports.

GET RID of the ridiculous ending weapons, they are way overpowered and unbalance the game to a ridiculous level. Or keep them and reduce how powerful they are, because the damage difference e.g. of the black hole eruptor from the prev weapon is too large to even make sense.

The whole tech tree (or the AI research algorithms) need to be rebalanced, to make all the civ's stay somewhat competent during the game. Having certain AI's research absolutely no weapons until attacked is awful. It also, in itself, unbalances the game, because the first civ to steamroll that civ becomes that much more powerful. If all civs attempted to keep up militarily, the game wouldn't be so easy.

And finally, if the AI's actually build improvements on all their planets (hoping it will be fixed, and GC3 won't suffer the same problem), that would help alleviate the rich get richer/poor get poorer problem as well.

As far as graphics go, I could care less if they stay the same as they are now, I would much prefer a more 'solid' AI. I hope they dump resources into this aspect over way fancier graphics that you can see only when you are zoomed all the way in.