DESIGN: CtP2 Source Code released - our actions [brainstorm & discussion]

[J Bytheway]

I think that for things much beyond bugfixing we should concentrate on expanding the capabilities of SLIC, and the ability of mods to alter the UI rather than adding the features in the game. This has the advantages that it shifts the onus of work onto the modder, and it is likely to provide extra functionality that was not otherwise anticipated (think of the things we've achieved with SLIC as it is, I'm sure they are far beyond what was anticipated).

[Solver]

Gee, had my connection lost for the last 20 hours .

Excellent input, Martin . And I do also agree that including the space layer could also be an option... I loved it in CtP1, and I am sure many others did.

I would also, though, agree that limit removal and bug fixing are out top priorities, before we proceed with any implementation of feautures. This is why I also didn't mention new feautures in my original posts.

Also, I believe that once we get a list of all bugs here (have we already?) we should start prioritising them by work order. Many are easy to fix, while some stuff, like PBEM, will take a longer time to get fixed.

Finally, on my wishlist, are options for CtP2 games. That's where Civ 3 definitely beats us, start up options. Like, we could go with No Ruins options. Make civ-specific stuff, and as an option. If we add space layer, maybe even make that an option, und so weiter.

[MrBaggins]

Repeatable randomness... that is getting the same result in an event, even if you reload directly before performing it.

Maybe the option to specify a game seed too...

[Immortal Wombat]

Originally posted by Solver
Finally, on my wishlist, are options for CtP2 games. That's where Civ 3 definitely beats us, start up options. Like, we could go with No Ruins options. Make civ-specific stuff, and as an option. If we add space layer, maybe even make that an option, und so weiter.

And make every single addition and alteration(that isn't a direct bug-fix) a user option in the start-up screen.

[Pedrunn]

DP

[Pedrunn]

I think that for things much beyond bugfixing we should concentrate on expanding the capabilities of SLIC, and the ability of mods to alter the UI rather than adding the features in the game. This has the advantages that it shifts the onus of work onto the modder, and it is likely to provide extra functionality that was not otherwise anticipated (think of the things we've achieved with SLIC as it is, I'm sure they are far beyond what was anticipated).

I second that!
Fixing some Slic functions like IsContinentBiggerThan, add some power over the UI and converting som GEA_ variables to slic would be the best to have new features to CTP2.

[Martin Gühmann]

Originally posted by stankarp
1) expanding the limits (units, buildings, tile improvements etc. should be easiest).

First the are two types of limits, on the one hand a database limit and on the other hand a graphics limit, for units there is no database limit at least not one that can be reached too easily. But you have the limit of 200 sprites. And 200 is not a reasonible number for not a 2^n, so that should be easy to change.

Actual there is also no database limit for buildings or wonders, it looks like the problem is that there is a limit of buildings a city can own, you can build the additional buildings but the city want get them. Not a problem for units as they are not owned by the city, but also for wonders. And by the way there is also another way to store the wonders in the city than arrays:

quote: Originally posted by Mr Ogre on 11-16-2000 05:37 PM There's no practical limit I know of on the number of units or advances in the databases (That is, there's a limit, but you'd run out of memory hundreds of millions of entries before you hit it). There is a limit of 64 buildings and 64 wonders (same as CTP1). Those should be the only databases with such a limit. It doesn't have anything to do with arrays though, it's due to 64 bit integers being the largest size that's convenient to deal with.

So to increase the limit of buildings and wonders a city can own, have to use two 64bit integers to make clear that a the city owns that building and wonder. Obviously each bit of the number can be one or zero, representing true or false, first bit stands for ownership of the first building/wonder in the database and so on. So whenever the savegame is reloaded the game looks up in the database and adds the according features of the buildings/wonders. Well this would explain why it is not a problem to change the attributes of a building during the game, but it is a huge problem if you insert a wonder for instance, you see that all the wonders are shifted. Well this example illustrates that it is not just increasing the size of an array, well this one might be easy, but can also be tricky.

Well for tile improvements, there is no limit of practical concern in the database either. Also the limit of graphics is not a problem. With the current originasion of the tile file there is space for 255 minus 2 or 3 terrain types, will with a reorganisation aka CTP1 the limit could be increased, as a terrain graphics consits of more then one tile images the number of tile images is higher, you can assign IDs 65535 maybe a few less, same is true for tile improvements, rivers, and border tiles. So actual the limit is high enough.

And for the database limit, for tile improvements there is also no relevant limit, for testing purposes, actual to have a look on all the nice tile improvements, I did a tileimp.txt with more than 900 tileimprovements.

The only problem here is to make shure that it is possible that the player can select them all from a menue, preferable from a list box like in CTP1 instead of a button bank.

Originally posted by Immortal Wombat
And make every single addition and alteration(that isn't a direct bug-fix) a user option in the start-up screen.

Actual not a good idea to have a very long list on the start menue:

Do you like to play with space?
Do you want to have diplomatic photos in the game?
Do you want civ specific features?
Do you want Great Leaders?
Do you like to play with workers and slaves with nationality aka Civ3?
Do you want to play with more city styles?
Do you want to play with more goods?
Do you like to play with strategic ressources?

Well I think all these options should be left to mods. The goal should be to adapt the text files so that all these features can be implemented easily by the moders. Make a space terrain possible. And so on. The solution should be to find a setup on that we can all agree as a base, and select our mod. So basicly I like to keep the current tech tree, but of course for modders to add space, but on the other hand if we want bring space to work we need an official with space version.

Originally posted by MrBaggins
Repeatable randomness... that is getting the same result in an event, even if you reload directly before performing it.

Actual this is already in the game, open a goody hut and you will find everytime the same stuff in it unlike in civ2, were I reloaded as long as I got a new city. That was rapid expansion. For battels I don't think the result should everytime the same, and actual even if you could reload RL I doubt that the events could be repeatable exactly. It is a difference if you look first into the one direction then in the other direction or the way around. For battles the result depends already much more on other things than on tossing a coin or rolling a die. So I think the degree of repeatable randomness is already balanced correctly.

Originally posted by MrBaggins
My thoughts on the first three things to deal with; resync, rushbuy, and then neutral pillage bugs.

No idea where the resyncs are located, but I think the players whose turn it is should be considers as the most up to date game, so that slic code at least wouldn't fail in MP. The rushbuy bug should have the highest priority, I think we can fix it very fast by just disabling the button when it is not the players turn, but actual more consideration is needed here to find the optimal solution, because some of the rush buy buttos allow rush buying when only one turn is left for the completion of the build item and other don't allow it. I think there is a sense in it to rush buy in item when only one turn is left. In a case of riots for instance, your temple just needs one turn, but can't be build due to riots and you can't pay the workers, is a little bit rediculous.

If the neutral tile pillage bug is easy to fix it has high priority. If it is difficuilt it has low priority, we have a slic workaround.

So there is a class of bugs with workarounds and a class without workarounds. So if the bugfix is easy for those bugs with workaround it has a higher priorty as it has a workaround and the fix is difficuilt.

Originally posted by Pedrunn
Fixing some Slic functions like IsContinentBiggerThan, add some power over the UI and converting som GEA_ variables to slic would be the best to have new features to CTP2.

Well the IsContinentBiggerThan is another story, I think we need a new AI model with a better AI map representation, etc...

Of course fixing some slic functions is also a good idea.

-Martin

[finfahey]

It's a little premature, given that we haven't got our mitts on the CTP2 code as yet, but what about getting hold of the CTP1 code too? I mention this because:

1) People seem quite well-disposed towards restoring the space layer (I poo-pooed this on another thread, but thinking about it, maybe it could be made fun and important to the game, instead of slow and boring as it is in CTP1).

2) Having CTP1 code for comparison could help in debugging. An example is the non-working veteran troops (works in CTP1, not in CTP2)

[Martin Gühmann]

I think the problem with this non working fetures of CTP1 in CTP2 is that the programmers wanted to rewrite the whole engine but weren't able to finish their work.

An example are the disapeared diplomatic photos they are in the *.zfs files but aren't used. In this case it looks like the programmers screwed up something in the last minute.

Another think I am missing is the pollution graph.

Originally posted by hexagonian
As a non-programmer, but a Modder, I'm wondering if the proposed fixes and changes are going to cause a host of conflicts with the existing major Mods?

Let's just say I'd hate the thought of rebuilding 'Cradle' from scratch...

As long as we don't alter any text filese that are modified in Cradle a new ctp2.exe would be compartible with it. But just fixing bugs is not much, so for instance if we want the redirect trade route feature in the game then roads for instance need the Freight field that is not available in the current version. So to make Cradle use of this new feature you would add this field to your Cradle tileimp.txt. So you just need to add the new fields to your text files, no need to start from scratch. It would be just an update.

-Martin

[MrBaggins]

Re: repeatable randomness, I'm fairly sure I remember that it was demonstrated in the CTP2DG that results were NOT consistent based on a single save game. People got different results, very often.

[Martin Gühmann]

Originally posted by MrBaggins
Re: repeatable randomness, I'm fairly sure I remember that it was demonstrated in the CTP2DG that results were NOT consistent based on a single save game. People got different results, very often.

Yes this was demontrated there, one reason for this is that the AI has the right to come to different decision, does it more its stack West or East South or North, does it attack or does not it attack. So differnt reactions and actions of the AI leads to different results. Well you could argue that is because the AI can't think strategicly. So this is not a problem of randomness this is a problem of the bad AI that has no idea what is a best for the next turns. But even with a better AI small variations must be allowed otherwise the AI is not smart and can't react properly.

On the other hand the differences are also caused by the different action of the players, not every step was done equally on all machines for instance forgetting to add an entertainer here.

Well and finally random influences. It is bad luck if you lose in one game one more archer then in the other game. For instance how do you want to determine the result of a battle with two stacks with exactly the same strength, which one should win. The attacker or the defender, because they attack or they defend, or because of the terrain boni. If the advantage of one stack is just small over the other one then there must be change that both can win. Otherwise the AI could be programmed so that it only attacks if it is sure that the AI win. There must be some uncertainess in it otherwise some of the fun would be missing, and this wouldn't be real live. And actual the randomness is not such a big, mostly it were just three units more or less maybe six, but not the difference between losing and winning. And actual in a normal game this phenominon doesn't play a rhole. And I am shure the level of random applies to Civ2 and Civ3, maybe the randomness of the outcome of a battle there is bigger then in CTP1/2, because of the less units stats there.

-Martin

[MrBaggins]

My point with repeatable randomness is that CtP2 apparently uses a "real" random function, using a timer to set a random seed, this means any reload will cause differences.

What I'd suggest instead, as per Civ3, is a pseudo random function, where a seed is set, then a mathematical process performed, to generate a "random" result (which is the seed for the next "roll".) Then the AI will make the same "choices" after a reload; each roll is essentially random, but consistent.

[Maquiladora]

The Ranking Screen should be like it was in CTP1 IMHO, just a simple overall ranking. In CTP2 you get far too much information which makes it easier to choose who to attack and who to ally with, which is information you shouldnt really know, like knowing how powerfu;l a military all your enemies have? Thats strange and really takes away some of the mystery of the game.

And like Martin said in one of these threads, bring back the pollution screen at some point.

[SMIFFGIG]

Minor Civlizations

yea maquiladora is right

that sort of info should only be available through spying

EDIT:
ok i just had to add this
but another nice feature would be

Minor Civilizations, these would have 1 or 2 cities and couldnt expand and the cities had there own unique style

they wouldnt come up or be part of the 7 or 6 civs already on the map, but you could ally with them and each minor civ would be powerful in some way for there size and have something unique about them
e.g a Technology that was unique to them that allowed a unique unit or City improvement for example

you could either gain this by killing them and hoping to steal it whilst looting there cities (wich wouldnt be a 100% chance) or by slowly gaining good relations with them then allying and eventually merging them into your civilization so all there cities become yours

this peaceful way would gaurentee 100% that you would gain watever unique advantage that this minor civ has to offer.

Im sure this would be difficult but with the source code it would also be possible

[Maquiladora]

What about increasing the 12 unit limit on one tile?