[C4:AC] Keeping Civ4 changes?

[Chaos Theory]

It doesn't have to be that bad, but rule-based AIs are fundamentally simple to program. It could have a short list of strategies, and pick one based on current conditions, re-evaluating conditions every so often. You could even implement independent strategies for offense, defense, recon, building, etc.

[binTravkin]

I imagine you suggest using whatever routines Civ4 does for forming attack groups and attacking, and otherwise giving the CPU a braindead route to pop-booming with tree farms?

Senethro, please don't be offensive.
I am confident CT knows what he is talking about, because I've been doing that myself.
I've been writing code for AIs ranging from simple chess programs to coding AoE AI (which is not any advanced one, but has quite a bunch of options and possibilities).
In the last example I was not even chaning the way AI 'thinks', I was just changing the way AI prioritises things and changing it's action patterns.

For a game where we all know what are the best strategies it's pretty easy to write an AI especially if you have a base AI from which to start with (the default cIV AI).
Of course the AI might be a bit straightforward in some situations, but it can still be made more flexible than some of the human players and what's the most important, it can be mad to do two fundamental things:
1.Not make fatally stupid decisions (like stacking units next to the base which is being attacked).
2.Use victorious strategies and exploits.

I am confident that if someone gave me SMACX base code, it wouldn't take me much to teach AI build and use crawlers and build SPs in a single turn.
The crawler use just needs making a formula with which to evaluate cost effectiveness of another crawler and cashing in crawlers for an SP really needs very little programming IMO.

Furthermore I can bet we will be able to teach AIs concepts as popbooming (in all the different ways it can be done), proper SE switching (not just sticking to Fundy/Power for entire game), ICSing, REXing (rapid early expansion), specialist usage and more, with not a HUGE effort.
All those things seem very simple from the algorythmic/programmatic viewpoint as most of them don't even need to be considered upon, they just need to be done in almost any game where one wants to be successfull.

[Senethro]

Wow, more talents emerge.

Originally posted by binTravkin
Senethro, please don't be offensive.

I wasn't trying to be offensive, sorry if it came across that way. :/ I was describing the approach I would have taken. Pop-booming without any considerations about it is probably the first thing I would teach an AI, and probably double its performance in doing so.

[drekmonger]

From there, just write a learning AI (yes, I know how), and allow it to learn.

That's a rather high estimate of "learning AIs". I assume your talking about genetic algos, genetic programming, and/or neural nets. I've never personally seen such an AI that was proficient in a game more sophisticated than connect-4 or video car racing. (...on the other hand, it's not a subject I've looked into since circa 1999...)

Anyway, the point isn't that it's *impossible* to construct an AI for Smac, which would clearly be incorrect. I'm just saying that it would take significantly less time and yield an acceptable result to keep as many Civ4 systems intact as possible--for one thing, you'd only have to rejigger the exisitng AI, instead of rehauling it.

The more ambitous the project, the greater the chance of failure. Personally, I'd set the 1.0 goals to something I knew I could achieve, working alone, in three monthes worth of spare time. In fact, that's exactly what I aim to do with the smac-ish mod I'm planning in my head.

Anyway, good luck, have fun, etc.

[binTravkin]

I'm just saying that it would take significantly less time and yield an acceptable result to keep as many Civ4 systems intact as possible--for one thing, you'd only have to rejigger the exisitng AI, instead of rehauling it.

Well, I guess noone said here we'll be doing something anew (at least I skipped it).
Noone writes code anew unless there's no example/base code available or the example/base code is too faulty/lousy to make efficent use of it.

I am not the best and most expierenced programmer here (CT is I guess), but in my work I've not been writing anything completely new since the days when I was making my first unofficial/noncommercial projects.
Right now a project rarely contains more than 1/3 of new code.


Illuminatus, if you're reading this, there's an organisational issue:
1.We should start splitting in groups of speciality - Programmers, Designers etc.
2.The people would then ask questions specifically for that group.
3.There will be no such mess as it is now - the design questions are answered by non-design people, same for programmers, resulting in a useless discussion just how we will be righting code.
4.While we do not have clear speciality groups, I'd like to admit that:
- programming questions should go to Chaos Theory, me and Darsnan *
- design questions should go to Wgabrie, Blake and GFC*
- sound questions should go to Platypus Rex as he was the first to start working on it (you can correct me here if Im missing something)*
- organisational questions should go to Illuminatus, Senethro, me.


* - any other person who has signed up for that taskgroup is eligible to answer, but should he arrive at doubts, he should not mess the discussion.


Let's not make mess by answering to questions we don't know and are not responsible about!

[binTravkin]

My suggestion: have one tyrant, your most experienced, valuable coder perhaps. He makes a checklist of what is to be kept and what is to be changed in the "official" mod. In other words, One Linus to Rule them All.

Suggestion noticed.

[Rasbelin]

Originally posted by Darsnan

What with you being a veteran of the Stella Polaris project I respect your comment: you tried to make a SMAC2 from scratch. Although this endevour ultimately failed, why in my book the second most noble thing is to have tried, but failed.

Actually your statement contains various mistakes. First of all Stella Polaris wasn't about making "SMAC2", not by miles, but something which was beyong it and had concepts not familiar to SMAC, but which were unseen or came from other games. Second of all we didn't ultimately fail in my and some others' opinion. We did make helluva lot of progress in terms of game concept designs for a space thematic TBS/RTS mixture. However we just couldn't get it coded, as people like Blake just walked away. Stating this to sort out the facts as they were.

Now onto what this thread was all about: I second what Blake has in mind in terms of this mod for Civ IV. He's right about keeping at least some Civ IV aspects of the game. The mod can't never be 100% SMAC, so indeed there will be flavours from Civ IV. You anyway have a very daunting task ahead while getting this mod actually into a playable format. Remember: you can have a lot of forum discussions, but ultimately you also need to deliver something. Some aspects of the Civ IV legacy features can be get rid of at a later date, as long as there's first something that can be tested.

[Rasbelin]

Originally posted by binTravkin

And this is getting too hot IMO.
Let's face two facts:
1.We have limited resources (which means we can't copy+paste SMACX, we will have to use at least some of the cIV base code).
2.The more you ramble here the more displeased with the project people get. Let's just learn Python and C++, read reviews and do other 'while we are waiting' stuff, but not shout at each other just because our opinions differ.

Two very important points. As I said in my post above, get first something coded that utilises Civ IV functions to a quite high percentage and then start venturing out on your own with crunching "SMAC'ish" code to replace the Civ IV functions you don't want to keep or need to have changed. The very first thing could be getting the very basic game data such as nametags corrected. Then continue to the next level and so on.

[Darsnan]

Originally posted by Rasbelin

Actually your statement contains various mistakes. First of all Stella Polaris wasn't about making "SMAC2", not by miles, but something which was beyong it and had concepts not familiar to SMAC, but which were unseen or came from other games. Second of all we didn't ultimately fail in my and some others' opinion. We did make helluva lot of progress in terms of game concept designs for a space thematic TBS/RTS mixture.

My appologies: I only knew Stella Ploaris from second hand conversations. I meant no offense, and thank you for clarifying this for others who came later after this project.

Originally posted by Rasbelin
Now onto what this thread was all about: I second what Blake has in mind in terms of this mod for Civ IV. He's right about keeping at least some Civ IV aspects of the game. The mod can't never be 100% SMAC, so indeed there will be flavours from Civ IV. You anyway have a very daunting task ahead while getting this mod actually into a playable format. Remember: you can have a lot of forum discussions, but ultimately you also need to deliver something. Some aspects of the Civ IV legacy features can be get rid of at a later date, as long as there's first something that can be tested.

I am going to quote myself in reply here:

Originally posted by Darsnan
Ideally speaking, we should attempt a complete clone of the game.
Practically speaking (that there will be some cIV aspects in the game) this is what we are going to end up with.

At this stage (i.e. before the game is released) we set an idealized goal. Once the game is released and we can then start empirically digesting exactly what can and cannot be done (as well as what should and shouldn't be kept), then there will probably be many discussions on these items. And I will close here by saying that the discussions in this thread are going to seem like a walk in the park compared to what will occur once we start discussions on the final form the mod will take.....

Originally posted by Rasbelin
However we just couldn't get it coded, as people like Blake just walked away. Stating this to sort out the facts as they were.

Just curious, but why did the coders walk away from this project? If at all posible if its something we can avoid on this project, why I'd like to hear the reason.


D

[Chaos Theory]

I would agree that we need Civ 4 with SMAC flavor before we have a total conversion, but I consider the partial conversion a milestone, rather than a goal. I imagine we'll have many intermediate versions, since we're lucky enough to be starting with a fully-functional game. I would expect more problems bringing together parallel changes from multiple programmers than problems implementing SMAC features.

[binTravkin]

I imagine we'll have many intermediate versions, since we're lucky enough to be starting with a fully-functional game.

Yes, and we should strive to not disrupt this full functionality in any version of our project.
Each version should be playable and more close to SMAC than the last one.

would expect more problems bringing together parallel changes from multiple programmers than problems implementing SMAC features.

If we can divide work so that each of the programmers works in his own area/module/library, it shouldn't be very problematic.

[binTravkin]

Just curious, but why did the coders walk away from this project? If at all posible if its something we can avoid on this project, why I'd like to hear the reason.

I suppose it was because of leadership.
It's the most often reason why coders seek for other job/project.
Leadership often envisions irrational and impractical things or on the other side sees coder's innovation as something unneeded.
I had just case in my company just yesterday after I delivered forum software similar to which I haven't been able to find thusfar.
Im now searching for a new job..

And I will close here by saying that the discussions in this thread are going to seem like a walk in the park compared to what will occur once we start discussions on the final form the mod will take.....

Have expierenced this..

[Mart]

It is only one more indication, how important is skillful leadership. I haven't took any specifically oriented in leadership courses, however, I happend to have some real experiences with leadership. starting from being a gamemaster in paper-and-pencil roleplaying games through teaching assistant, to directing others in my student job, where I had to have "broader look" onto who is doing what and what is more important to do first.

And here I would see this project to have a Leader. Maybe also possibly a 3 person Leadership Council. One Head Leader, to have final say. As for me, I will follow this leadership for the sake of the project being completed. Any my ideas will be written as suggestions and submitted to the leader for consideration. A good leader takes seriously any suggestion in my oppinion and if rejects it there is always a good reason given to the author of the proposed change. Well, just my idea of how it might look. If I think of something new, and someone is capable to show me a better way than this I think of, then I will agree.

[binTravkin]

Exactly my thoughts and I suggested it to Illuminatus just the other day, still waiting for an answer or a broader discussion to be started.

[Mart]

As for me, I think the person who gives direction should be someone who knows SMAC very well. And for the Leader I would support Darsnan. I like the approach, that you have to this project, Darsnan, and it is seen already in the mission statement.