Announcement

**RonHiler** · May 20, 2000, 20:31

Hi Mark,

I'm sure much of this was covered in your forum, but I'm afraid I don't get there as much as I should (I do lurk sometimes, but I'm a bit busy on other projects, so I don't read everything, I'm afraid), so forgive me if I am going over old ground I am interested in some of your thoughts about the project (as one developer to another, heh).

quote:

Originally posted by Mark_Everson on 05-20-2000 11:00 AM

One of the main goals of this project is to require less micro-management from the player (it will remain optional)

This is a very tricky proposition. If MM imparts any benefit to the player over those not using MM, you will have universal MM usage (if that makes any sense, heh). IOW, people will be compelled to use as much MM as they can if it gives them some benefit, because if they do not, the next player (who presumably is using MM to their full advantage) would have the upper hand, right? Do you have ideas yet on how to combat this problem? I ask because we have struggled with this problem in our own game quite a bit.

quote:

Further setting Clash apart from others in the genre, truly our "Holy Grail", is a sophisticated AI far superior to that available in the commercial games.

You'll have to beat me to it What kind of AI system are you using? Neural net? Expert? State machine? Some other type I don't know about?

quote:

Current Status:
The Team of over 20 people so far, is continually making progress on both the design and coding of game models for Clash.

How does it fair organizing and controlling the output of such a large team? Presumably, you have a virtual office (or am I wrong?), do you think that makes it more difficult or easier?

quote:

The available coders (More Needed!) are working up a series of demos showing what we have to date. We are committed to releasing a demo at every major milestone so we can get a Reality Check on how the game is working so far.

A good plan. We are using the same model ourselves.

quote:

Coders, Designers, Graphics, & Interface Gurus wanted:

The Clash project is still looking for people to join.

Even with 20 people, you are looking for more? Why such a large team?

quote:

If you're interested in participating

Ah, if I didn't already have a major project of my own going on, I would probably volunteer. So much coding to do, so little time I don't know much about Java anyway, I'm a C++ coder, though I understand they are quite similar.

I hope you get everyone you need.

Ron

**Mark_Everson** · May 21, 2000, 19:53

Hi Ron:

How's it going? I'm not sure I'll get through everything you brought up... we'll see.

I think your premise about any game being unstable with respect to micromanagement isn't right. IMO the correct way to look at it is that people will use micromanagement when the trade-off between micromanagement and playing time is beneficial. For instance, in civilization 2, I would gladly use the automated settlers if they were only 10% less effective than doing it all by hand. My reason is that the time savings from not having to do all that stuff would compensate me for the loss of effectiveness. Quite simply, I can play more games! Of course, I consider the penalty to be much worse than 10% so I don't use them. I think most of the problem with micromanagement really disappears if the AI is reasonably good, and the player has a decent way to guide the AI. (Just having the AI go along using its own notions of what the right thing to do is is clearly not acceptable.) Then the player can focus their time micromanaging just the things that they feel are fun... or possibly, as we hope to do in Clash, micromanage nothing. I think it will be great just issuing the big strategic orders and seeing what comes out of it without needing to push all the pawns by hand!

In terms of the groundbreaking AI, we're using an Hierarchical AI, with essentially an expert system running each level. Just for spice, if the AI gets enough time to think, it will also use a genetic algorithm approach to modify the solutions that come out of the expert system. If you want a brief overview you can get it by reading the bulleted list at the clash AI page. There is also a detailed description there of our general approach.

So far, the organization of the team is all done "by hand". Twenty people sounds like a lot, but in general usually only five or six at a time are actively working. Real Life tends to take almost everyone off actively working with amazing regularity. And that's why we need more! We aren't doing too bad in the design department, but I have had a really hard time finding dedicated programmers. I think we have a couple now in addition to me, and hopefully we can add some more. But we have about 10 game models all in need of coding, so I could easily use 10 or more programmers.

Good luck to you and your project also!

Mark

**VetLegion** · May 22, 2000, 17:46

Hi Mark

I wanted to send you some questions about clash but I figured out most when I dig through your web site deeper.

I am interested in two more technical things I havent found out:
1) Is AI having same interface as the player does (no cheating)?
2) Is the multiplayer model that was discussed in threads accepted? If it is, and it is peer-to-peer model, can I read more about it somwhere (I read everything on the forum)?

Hi Ron

I am familiar with MD (I have a little older demo on my disk, I will get newer one when I have time).
We are also making a civ game in c++, that is OpenCiv3. I have the impression your game is almost done, so I am interested:

-how many people do you have involved (coding)?
-how long it took to design the game before coding, and do you think it is good to do, or just adapt as needed?
-is your ai detached from code, and what technique are you using?

**RonHiler** · May 22, 2000, 21:55

Hi Vet,

quote:

Originally posted by VetLegion on 05-22-2000 05:46 PM

We are also making a civ game in c++, that is OpenCiv3.

Ah, yes. I have only recently learned of your project, but am following with interest. Us turn based strategy game developers need to stick together (BTW, if you'd like a link from our page, send me a small pic and/or a brief description, ala the format that our links page has, and I'd be glad to add you guys in. I really need to update that anyway, I think Mark's game link is still pointing at his old page).

quote:

I have the impression your game is almost done:

Yes and no. The initial Alpha is nearly complete (the first playable version). I suspect it will be a while (a year?) before we finish betas and I call it gold. But while we are in Alpha and Beta stages, the game will be free and playable (though a bit feature incomplete and buggy at times), and multiplayer games will (hopefully) be being played (all the better to get bug reports!)

quote:

-how many people do you have involved (coding)?

Coding? Just me. We have another person who does Art, and a sort of unofficial person who does our web design, and that's the entire team. We're pretty small (he said, making an understatement..)

quote:

-how long it took to design the game before coding, and do you think it is good to do, or just adapt as needed?

Tricky question. For MD, we purposely did not create a design doc, for I wanted this game to evolve into being, rather than try to lock the entire design down from the start. If I had done that, MD would have been a completely different game in a lot of ways. That's a pretty dangerous thing to do, though (for various reasons) and I wouldn't necessarily recommend it for other developers and/or game types. It worked for MD (mainly because it's a turn-based game) and for our team in particular (partly because we are so small). If both of those items weren't true, I'd likely have mostly feature locked the game before starting coding. Otherwise, you're just asking for team burnout, heh.

quote:

-is your ai detached from code, and what technique are you using?

Yes, it's completely detached (in fact, it's so detached, it's an entirely seperate program right now, I haven't integrated it into the main program yet). They will be linked via two classes (the player stack class, which controls each stack of units, and the player units class, which controls individual units for every player). Other than that, they will be completely independent.

As for techniques, our AI is a combination of Fuzzy logic and Neural nets. The AI is a work in progress, to be sure. I think when our first Alpha hits, it will be none too intellegent (right now, it makes some pretty questionable decisions for attack target picking). But I will be tweaking it quite a bit during subsequent releases, as I decide what factors and weights need to be considered for each decision the net has to make.

I should point out that when I say 'AI', I mean the combat AI (it decides what each unit ought to do during combat), which is the most complete of the bunch. Computer opponents will use the same techniques (except we are also going to add learning algorithms to their nets to modify the weights on the neurons as the game progresses). Right now (and for at least the Alpha 1 release) they are completely brain dead. You can put in computer players, but they won't move, they won't build cities, they won't research. Basically they will just sit there. Computer AI programming probably won't happen until Alpha 2 or 3.

For the pathfinding, we're simply going to use A*, somewhat modified so that it works on a spherical world. Those are our three main AI's in the game. A fourth, sort of helper AI (to take over some of the more menial tasks of running a large civ at the player's request), is being considered, but that's a ways off, and might not even make it in at all (it's a sort of "wait and see if it's needed" kind of thing).

Whew, long. Probably more details than you wanted

Ron
[This message has been edited by RonHiler (edited May 22, 2000).]

**Mark_Everson** · May 23, 2000, 20:33

VetLegion:

Yep, the plan is to have the AI not cheat. We'll see if the team can pull it off.

The multiplayer model discussion died a long time ago. The stuff on the forum is all there is... I don't think I'd consider it an accepted model. If someone came along wanting to code that part, we would start a new discussion, but refer back to the old one.

For now multiplayer is Way down on the priority list. That may be a mistake, but I'm hoping someone that really knows what they're talking about shows up to take it over.

**VetLegion** · May 24, 2000, 19:07

Mark

I see. Anyway, the info I got from threads is good. I wish you find a network programmer.

Ron

I will note our webmaster to exchange links, as you said, we are in the same buissiness

First, I cannot run the latest demo. I got the funniest error message win2000 ever gave me: "error: 03.7190 is not a valid floating point value". last time I checked it was...

About number of coders...only one? You must be pretty expirienced yourself, to cover everything from UI, to network code and graphics.

Which brings me to the network (my favorite design topic right now). I noticed in that help file (the ad in it says "..when the author is honest enough to register.." this gave me quite a lol ) there is talk about different files being transfered around. I think it is simmilar to what we are likely to do, just without files (send just the necessary data, not whole game state). We will try to implement simultaneous turns also.

AI
I can see about fuzzy logic (CTP uses that), but neural networks...does it work? I argued with a friend about NN in civ-like game, and we agreed it cannot be used for stratgic purposes (civ planning), but only tactical (combat). If I understood right, that is what you are doing. I cannot wait to see it in action.

**RonHiler** · May 24, 2000, 20:24

quote:

Originally posted by VetLegion on 05-24-2000 07:07 PM
First, I cannot run the latest demo. I got the funniest error message win2000 ever gave me: "error: 03.7190 is not a valid floating point value". last time I checked it was...

Yeah, isn't that ridiculous. Sigh. I hate that error. As far as I'm aware, there are two (and hopefully only two) problems that will cause the game not to run. That is one of them. I've already fixed that problem for what will become the next version of the game, but if you want to get it running now, simply do the following:
1) Open up your control panel (Start|Settings|Control Panel)
2) Double click on Regional Settings
3) Click on the "Number" Tab
4) At the top, you will see a "Decimal Symbol" box, and it will likely have a comma in it (or some other symbol?). Change it to a period, and hit Okay.
5) The game should now run without that error.

I've written a routine that checks that setting, so in the future, it can be set to whatever you want. But for now, it needs to be a period

In case you or anyone else is having the other problem: The game, on some machines, won't run in 24 or 32 bpp mode (It'll start to load the map, then give an access violation error). Switch to 16 bpp (High Color), which may require a reboot, and it'll load. I'm working on that one right now.

quote:

About number of coders...only one? You must be pretty expirienced yourself, to cover everything from UI, to network code and graphics.

Yeah, I've got a fair amount of professional programming experience (done a lot of business apps to pay the bills). But even so, I've learned tons doing this game (far more than I ever did programming a thermocouple interface or a graph plotting program, heh).

quote:

Which brings me to the network (my favorite design topic right now). I noticed in that help file (the ad in it says "..when the author is honest enough to register.." this gave me quite a lol ) there is talk about different files being transfered around. I think it is simmilar to what we
are likely to do, just without files (send just the necessary data, not whole game state). We will try to implement simultaneous turns also.

Heh, that little ad will go away in the next version. I always register any shareware I intend to use once I've evaluated it (we have registered versions of Paint Shop Pro and WinPatch, for instance). I was still checking out various help file writers when the last version was released, but I've decided on keeping that one (Help Scribble, a good program for it's price. There is one better, but what they are asking for it is simply silly).

Anyway, the file transfers will be our "network code" so to speak. There will be no option for direct connections in MD, it'll all be through e-mail. I considered it, but there are too many headaches for not much gain (since our game is all simultaneous turns, there would be no advantage, except for chatting (which admittedly, would be nice) instead of messaging). Do you plan to have direct connections in OpenCiv?

quote:

AI
I can see about fuzzy logic (CTP uses that), but neural networks...does it work? I argued with a friend about NN in civ-like game, and we agreed it cannot be used for stratgic purposes (civ planning), but only tactical (combat). If I understood right, that is what you are doing. I cannot wait to see it in action.

What led you to that conclusion? There is nothing fundamentally different between controlling a civ and controlling a unit in battle, if you can quantify all the factors that need to be considered for a particular decision. Of a certainty, this is more difficult to do in the latter case, but assuming it can be done (a task much simplified by the use of fuzzy logic, BTW) then it breaks down to the same sort of problem in either case: Given factors A,B,C,...Z, (with corresponding magnitudes or weights), what is my best option for such-and-such task. This is exactly the sort of thing NNs excel at

I'm being awfully verbose on this board, aren't I? Sorry about that

Ron
[This message has been edited by RonHiler (edited May 24, 2000).]

**VetLegion** · May 26, 2000, 19:04

Ok, I changed that setting in regional options and it worked.
The map is very good. Is it generated flat, and then wrapped/adjusted to a sphere, or is it made on a sphere and then streched to the screen? The map in civ2:ToT also has a globe minimap, but it is prodeuced by mathematicall approximation from flat map. It cannot be zoomed, because, simply the civ world is not round. I think you did excellent job at your map.

Yes, we will try to make OC3 have direct connections in multiplayer. Your system (files) is good for play by e-mail, but it is simmilary structured to what I thought of for OC3. Can you tell me what were your headaches for direct-connection MP? I think we are nearing a workable model, and I want to see if it is good.

Also, I havent read that your game will be simultaneous turns. Can you please refer me to a document about it, or explain? ( as you see, I dont mind you being verbose )

As for the AI, well, I have never done any so I am speaking about theory. The problem we face in OC3 is a future evolving set of game rules, and under such circumstances it would be difficult to adjust the factors all the time something changes. Indeed, it will even be difficult to write an expert system for it. I was thinking that NNs can be used therefore for simpler tasks that involve a limited number of factors, that are easily quantified, and changed. Strategic decisions involve so many many factors...
Again, I am maybe talking nonsense due to my lack of expirience . But I am confident that some good AI will emerge since it is going to be detached in a dll (like in civ2evolution).
Someone, somwhere will write a good AI

**Mark_Everson** · May 26, 2000, 21:41

Um, Guys, if you are going to have longwinded discussions about MD, could you please put them in the MD thread? I keep returning, thinking that new posts are about Clash, but it just goes on and on . And if we ever do get real Clash discussions going here, people will have to wade through several pages of OT material...

Mark

**RonHiler** · May 26, 2000, 23:06

You're right, Mark. My bad. Sorry 'bout that.

Vet, I'll answer your questions over in the MD thread.

Ron

**VetLegion** · May 27, 2000, 04:24

Ups, Mark sorry about that

Should I go and edit the posts? That way, it can be same as before.

**Mark_Everson** · May 27, 2000, 19:56

Gents:

Thanks much! No editing of previous posts needed...

**Mark_Everson** · August 5, 2000, 18:44

We could really use some people to check out the military models and comment/work on them. The two guys who were on it have developed a case of Real Life...

Please check the Mil page on the web site if you're interested in commenting, or even better, want to help develop the model.

------------------
Mark Everson
Project lead for The Clash of Civilizations ( clash.apolyton.net )
(That means I do the things nobody else wants to do ;-) )
This Radically different civ game needs your suggestions and/or criticism of our design.
Check our our Web Site & Forum right here at Apolyton...

**Mark_Everson** · February 21, 2001, 08:44

Demo 5 will be out in a few weeks. It will include extensive government and economy models, and the first cut at our new military model.

We could still use artists (none currently) and at least a few more Java coders (three currently workin' hard! and a few more joining in...).

We can also use your comments and criticisms on our design so far. Good ideas stand a great chance of being used!

-Mark

Announcement

Clash of Civilizations - More Designers / Coders needed for Civ-Type Game Project...

Clash of Civilizations - More Designers / Coders needed for Civ-Type Game Project...

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment