I'll pay for Alpha Centauri ai .exe mod

[dino]

I was experimenting recently with alphax.txt to provide a challange in single player, by leveling a field between human and AI.
Buffing AI and nerfing ( but not removing ) the best winning strategies of human player - borderline exploits ( crawlers, forest, satellites, popbooming, copters, etc. )
I believe I can find cool nerfs, where all these things AI cannot into are still worthy tricks, but instead of being I win buttons, they're required to match AI buffs.

Unfortunately in my experiments I've stumbled upon a roadblock that can't be solved with alphax.txt mod:
- Even if I encourage AI to research centauri ecology and build formers early, AI wont build formers later, nor replace destroyed ones.
- AI will keep building military units until support eat all production and it can't build anything anymore.

These two issues basically grinds AI bases development to a halt by the midgame.
If you survive an early game, you won the game, it's only a matter of time and your efficiency from that point.

Proposed solution procedure:

If a function that decides base production, choose a military unit to build than make a check:

x = 0,6 + 0,1 * ai-fight + 0,2 if at war

if ( base support cost > x * base production )
{
if ( number of formers <= ( base size/5 + 1 ) && number of formers < 3 ) build former;

else if ( base colony pods < 1 ) build colony pod;
// requires testing, maybe condition or alphax.txt variable to disable it, but often factions refuse to spread past initial few bases

else build anything else that doesn't increase support ( base facility, probe, crawler etc.)
// put a conditional checks into the original function to assure it has zero chance of choosing a military unit and rerun the function to achieve this effect
}
else build a military unit choosen by the original function;

I believe this modification would assure that AI will continue to spread, terraform, and build base facilities to the end of the game at good pace.

I'll paypal 100$ anyone with assembler skills up to task of implementing it on top of the latest Ytzi patch.
If someone thinks he could do it but 100$ is too little, please tell, maybe someone else will declare to join this initiative with some money.

[Vidsek]

  Yes, there are other folks that would support this.  Myself for one

  The AI emulating North Korea has been, perhaps, the most dissapointing bug in the AI for me, even to the point that I have gone into edit mode and manually cleared the deadwood from my enemies' jungles to try and (temporarily) improve their performance.

  One thing I think I've seen is that the overbuilding of sentry/guard units -seems- to start once the faction contacts another that is hostile, even if it is across the board and no real threat.  Before that the production is more balanced.

  *If* that is true, perhaps an adjustment to how it reacts when such a contact is made might be useful as well.  Say, by having it take distance, tech level, etc. into account before deciding to go into panic mode.

  There are lots of other tweeks that would be nice, but Dino's ranks very very high on my list.

  I'm guessing it would even make multi-player games more interesting and fun.

[bvanevery]

I agree that the AI overproducing garrison units is the single biggest problem with the game.  I currently don't know how to solve anything.  I'm investigating the feasibility of .exe modding.  The core infrastructure I need to see, is symbolic function entry points that aren't changing, so that I can just get to coding and not hunting and pecking for what's where.  I'm still learning about what's been done already.

I don't think you can actually pay programmers enough money to take on such tasks.  It's thousands of dollars worth of free market skills.  Better to spend your time to contribute to the infrastructure necessary to support such efforts.  What you want, is for a random ASM coder to be able to show up in this community, and be able to do something effective quickly.

For real funding, if I ever do a Kickstarter for a new 4X TBS project, I'll be sure to keep you guys in mind.   

[The_Reckoning]

Might be preferable to try funding an open source remake a la FreeCiv...

Or maybe to see how much it would take for EA to release the source and open the license? In theory it might be feasible.

[bvanevery]

Preferable for whom?  I'm supposed to be a professional game developer.  Instead I'm an amateur working on like the biggest SMAC mod ever, and literally starving were it not for food stamps.  If you want to organize an open source attempt at SMAC, by all means have at it.  It's been tried before and people have failed.  SMAC is a bigger project than Civ II, which has been done ala Freeciv.  That codebase is horrible IMO.  Last I looked anyways, you don't want to work with it.  I've done my time playing Freeciv to death, even modding it a little bit.  I went back to SMAC, because it's still better than anything GPLed open source is going to produce.  There's a life cycle to what people can actually get done in open source, and SMAC is beyond that life cycle.  I don't think you can point me at a single game in open source, in any genre, that's at the level of SMAC.  I think I spent enough time looking for extant "hackable" work with 3D game engines to know this, but if you've found otherwise, tell me where it is because I sure missed it.

[Mart]

It will probably be, that someone finally will create a niche game. Some small gamedev studio, that will be able to do it.
Maybe even a kickstarter or similar attempt will make it, although, copying SMAC success would not be achieved just like that. So assuming kickstarter attempt would end with release, chances are, it would not stand to SMAC quality.
But I think something comparable is possible.
I have heard Pandora is good, and I have to finally try it.

... I'm supposed to be a professional game developer.  Instead I'm an amateur working on like the biggest SMAC mod ever, and literally starving were it not for food stamps.
...

I work in software development, although initially I was a chemical engineer. I know difficulties of career change and I know something about work in development. They say, there is shortage in developers, although it is more like there is lack of skilled developers. But with years, one can get needed skills.
What technologies are you most familiar with?

World recently drifts towards web/cloud. Knowledge of .NET, .NET Core, Javascript, HTML, CSS, and on Java side their corresponding techs seems to lead. Yet, from experience, I know there is a lot of legacy code, so still knowledge of C, COBOL, even Fortran counts. C++ recently goes through revival, and sometimes they offer better money than in Java or C#. And recently there is this embedded stuff.

Anyway, what I mean, is that maybe you could find some very well paid job.

[bvanevery]

I have heard Pandora is good, and I have to finally try it.

They didn't do any of the narrative stuff SMAC did, making it feel hollow.  I demoed it for a bit without buying it.  The combat mechanics do remind one of SMAC.  I don't know if the AI got better or not.  Upon release there seemed to be an issue of the AI not being very good.  Then someone came along and modded the heck out of it, to the point of the original devs somehow incorporating this modder's work.  I don't recall whether he joined their development team or just consulted it or what.  I don't know where Pandora's gameplay or community life stand today.

What technologies are you most familiar with?

My background was originally in nitty gritty 3D graphics, down to the ASM level.  However I did not prove to have the stomach to keep that up over the years.  In 1998 I quit to find my star as a 3d game developer, and I got lost!  Been trying to figure it out ever since.  Got way too involved in build systems ala CMake and GNU Autoconf for awhile.  That was my career dead end.  It was product of all that open source wrangling, trying to get other people's broken crap to build.  I got paid briefly to do it professionally but it was an awful, awful task, I'll never do that again.  I have invested a lot of time trying to come up with a better programming language than C++.  In practice I've gotten nowhere, although I think a few of my design ideas are actually worth something. 

World recently drifts towards web/cloud. Knowledge of .NET, .NET Core, Javascript, HTML, CSS, and on Java side their corresponding techs seems to lead. Yet, from experience, I know there is a lot of legacy code, so still knowledge of C, COBOL, even Fortran counts. C++ recently goes through revival, and sometimes they offer better money than in Java or C#. And recently there is this embedded stuff.

Anyway, what I mean, is that maybe you could find some very well paid job.

I am far beyond caring about any of those things.  If I were paid to do any of those, I'd be miserable.  I am not miserable in my present poverty circumstances, just stressed.

Sometimes more than others, didn't help that my car died for good in November.  Sold it for scrap.  I'm an amateur auto mechanic, a skill I started on before I started running out of money, and perfected out of necessity.  I'm not going to do that for a living, getting exposed to brake dust etc. sucks.  I don't enjoy auto repair, I'm merely competent and disciplined at it.

Insisting on doing what I actually want to be doing, is why you guys got my mod.  The logical thing to do next is to suck it up, swallow some DirectX 11 that I hate, get a prototype of a 4X game together, and then do a Kickstarter or some such.  If I can financially survive that long.  Failing that, I'll have to figure out something in computers that I can actually stand, that doesn't make me completely miserable.

I've had some hope that Jonathan Blow might ship his Jai game programming language soon, but so far no rollout.

[Mart]

The logical thing to do next is to suck it up, swallow some DirectX 11 that I hate,

How about OpenGL? It is still possible to use it with success on Windows.
The application I work with now, it is CAD type, and it uses OpenGL and not DirectX.

4X would not probably need a lot of "3D power", I mean no race for frames-per-second world championship, etc.

get a prototype of a 4X game together, and then do a Kickstarter or some such.  If I can financially survive that long.  Failing that, I'll have to figure out something in computers that I can actually stand, that doesn't make me completely miserable.

I now just go for 8 hours to my job place, program what is needed and at home I can take care of the stuff I like. Maybe apart from the time I need to learn technologies to get a better job.
I cannot say, that I do not like what I do at work, but now after 2 years, I know it is just another episode in my programming career. In one book from a dev blogger I read, that according to his experience, one needs to change programming job every 3 years or sooner. It prevents boredom, professional burnout.
One outsourcing company I talked to, sends employees for 18 months periods to projects.

I've had some hope that Jonathan Blow might ship his Jai game programming language soon, but so far no rollout.

I haven't seen Jai yet, need to check it.
Something I have not yet checked is, that while now they allow to use 3D game engines free (and I think pay only when you successfully sell a game) like Unity or Unreal.
But I am not sure, how good they would be for a 4X game.

If limiting graphics to 2D, there is SDL (Simple DirectMedia Layer), and there is like little work to get something working well.
For example Civ3 does not feel like 2D, thanks to isometric map view and well designed "sprite" animation.
Civ4 could very well be made in a similar way.

[Geo]

Civ4 could very well be made in a similar way.

Ah. Good ol' cIV!
At least that version of Civ was still relatively easy to mod, on the graphical level as well.
Otherwise Maniac's Planetfall project would never have been so immersive.

[bvanevery]

How about OpenGL?

Oh God it's way worse.  The 'design' is so crufty, and commercially it's all but dead in game development now.  I tried to stomach it for 3 years when I was doing Linux, back when the Steam Machines looked like they could be a thing.  Well they didn't, they sucked.  So I gave up on Linux and went back to Windows.  Nobody wants OpenGL for game development now.  Apple's doing Metal, even the Linux crowd is trying to do Vulkan.  Not that anyone cares about Vulkan, Khronos is a joke at a lot of the things they push.  Only thing that caught my attention lately from them is glTF, a binary 3D asset format that might be useful.

4X would not probably need a lot of "3D power", I mean no race for frames-per-second world championship, etc.

People always make this mistake, thinking that turn-based games don't have to perform.  But they are actually complex visualizations and you do have to zoom around a lot to see stuff.  You don't want to be waiting between turns for slow animations.  I seem to recall some of the recent Civs really pigging the 3D and driving players nuts with that.  It probably doesn't have to be the best 3D app ever, but it shouldn't be slow and bloated with little attention paid to the design, as though it doesn't matter.

I now just go for 8 hours to my job place, program what is needed and at home I can take care of the stuff I like. Maybe apart from the time I need to learn technologies to get a better job.

If you can stand to work full time for someone else at computer programming, then do even more of it on your own, power to you.  I can't.

one needs to change programming job every 3 years or sooner.

I probably changed focus like that, but there was usually no job involved.  There are times when I wish I had.  Especially at the beginning when I knew the least.

Something I have not yet checked is, that while now they allow to use 3D game engines free (and I think pay only when you successfully sell a game) like Unity or Unreal.
But I am not sure, how good they would be for a 4X game.

As I have all the skills necessary to write a 3D engine, and have even attempted to do so 2 times in the past, I'm quite loath to use someone's proprietary engine.  I've been around the block with a number of the open source engines.  Most have problems sustaining their efforts over time.  If you're going to pick up someone else's code, you need to be prepared to do everything yourself, because their whole development effort may crumble.  Then it's all about whether they did something that was worth continuing with.  I was looking at Urho3d for a time, but their leadership went into shambles about a year ago I think.  Burnout. 

There's some other guy doing something called the Wicked Engine, need to look at that.  What I like about it is, he's not trying to do cross-platform much yet, it's just straight Visual Studio .sln files and DirectX.  A huge mistake these open source types make over and over again, is trying to run on all sorts of different platforms instead of doing 1 platform well.  I've made that mistake too, burned up a lot of my so-called career on it.

If limiting graphics to 2D,

There's not much labor to be saved that way.  Even 2D games started doing it via 3D graphics like 15 years ago.  It's also not hard to find a 2D oriented open source engine nowadays.  Godot seems to be the one that people look at the most.  It's also 3D, but I think it's probably oriented more towards 2D.