FeXoR

Well, here's a start: iber_wall_civ_bonus2012-5-3.zip I added a new starting unit placement method to rmgen misc and adjusted all random maps to it. Iberians now have walls added to their starting entities automatically. They can be set (e.g. to towers only) for maps where walls doesn't fit by an optional keyword argument so the function can be easily adjusted to any further needs without adjusting all rms again (which sucks). Rms no further need to get the starting entities - the place function does it now. I didn't test if the trees caught in the walls cause trouble... The fortresses are not randomizes yet but will be. Some variation might perhaps be nice (sometimes elliptic, sometimes irregular polygons, ...) See the wall_demo rms for some examples. Most naval maps only have towers added.

The naming convention that is used for civ building entities ('structures/' + civ + '_' + type with type in [wall_short, wall_medium, wall_long, wall_gate, wall_tower, outpost, fortress]) could be used for palisade entities as well like 'structures/palisades_wall_short', 'structures/palisades_wall_medium', 'structures/palisades_wall_long', 'structures/palisades_wall_gate', 'structures/palisades_wall_tower', 'structures/palisades_outpost', 'structures/palisades_fortress'. This would be nice to have.

Well, lets see how it works out when order priority is in place and maybe stances are redone as well. I think I made my points clear and perhaps it needs some playtesting for further investigation after those are in.

It's a matter of taste of cause and it's surely unique. In general I like the variety of the walls styles. The gate ending with towers fit better to a wall directly IMO. Never mind. With the information I got here I made most things generic in wall_builder for rmgen so I'm happy

Thanks for the reply. I'll start with demo bot to get the concept as a whole. Tried with qbot but gave up half the way on. I seam to be somehow handicapped in matters of reading foreign code Was a bit stupid of me to ask for the code I imagine I'll try harder...

Yes! Really good point I forgot to mention this one.

Just wanted to say that the Iberian gate is slightly to shot as well. With towers in between with the given width of X = 6 using the concept explained above it fits but not to another wall. In addition with a tower next to it is looks not that great since both are rounded. Not an important matter but I just wanted tell.

Ok, thx

Assuming here we only speak of attack units: I setup some control groups for attack units so I only have to manage about 3 attack groups which reduces the amount of attention needed for them... as far as they follow my orders and don't run everywhere. So they should be able to move a bit (lets say 1.5 to 2 * max attack range of any unit so that melee cavalry has a chance to chase down ranged units) and attack if enemies are close but otherwise approximately hold the position. Don't know which stance that is... I never want them to do that automatically. If I want them to attack I'd use attack move if available. That is indeed a good point for stances. However, I'd try to build defensive structures in such a case and gather my units behind. An patrol order is IMO not that much needed but helpful and stances doesn't cover it as well IMO. Just to make sure we get each other right: I assume patrol (as attack move/auto explore/search and destroy/gather) is an player given order with vast help of the unit AI. In that cases effecting them by stances might seam useful in theory but as far as I thought about how that should be in detail (with the help of some others like e.g. feneur) in most cases it turns out it isn't. So the units with a patrol order should IMO act like said before as well (be free to move and attack but approximately hold the 'position' - in case of patrol don't run away to far from the 'patrol line') That way ranged units will rip your units away without much danger to themselves. Same in my case. I don't have any problem with stances to be implemented as far as they don't disturb my style of play. But 'as is' they mainly reduce my influence on the units (mainly because of the missing priority system and the stances implemented 'prematurely'). I agree that stances are more useful then formations. But AFAIK formations are planned to 'cheat' by disabling enemies to use focused fire (kill one unit after the other) and make them 'attack as one' (whatever that means?). My hairs raise thinking of such things... Perhaps open a discussion about formations as well and let the PPL explain their thoughts and ideas? I'm not entirely sure I get you right. Are you agreeing that stances shouldn't interact with player given commands? If so I'm fine because it doesn't interacts with my style of play I'm open to discussions about stance usage for commands making vast use of the unit AI like attack move/gather/auto explore/search and destroy/patrol (if those are implemented) but as far as I see this is only helpful in edge cases and mainly makes the player to click more (the stance buttons) so I'm not thrilled with that idea either.

Thx. I adjusted it by hand (and eye ofc ) and since I think changes will most likely be made to fit the standard I will have no problems. Are the Carthaginians walls planned to be changed until the next alpha?

I know, but there has to be a basic length unit. AFAIK 4 units in the 3D engine is one tile (the size of a texture mainly) in RMGEN. That approximately matches with 'Footprint' and 'Obstruction' within the entity XML files as well. Thx! I didn't meant the functionality but the width of gates though perhaps it's just best to avoid them at all, even for style reasons ATM.

Thx for the reply I thought it would be to much work to remodel that and with towers considered in between it is not useful at all Good to know the general idea but what is X? 6 units (1.5 in rmgen tiles) and palisades_rock half the size or different for all 'wall styles'? Are they planned to be of the same width for all civs in the end like Mythos_Ruler said (if I recall correctly)? This would be really useful for generating wall styles generic for rmgen rather than adjusting them by hand... which is terribly annoying. Not all entities seam to be adjusted to that system though (especially gates): Carthaginian and Celtic gates are not as wide as their long walls for example. Not that important (perhaps more a style issue) is that some gates still have towers at their sides that cannot shoot or be garrisoned AFAIK (Iberian and Roman siege). Have a nice day

Oh, I always tried 'print' Wouldn't this be more intuitive? Well, at least I know now Thanks!

Updated roman siege wall and adjusted width and indentation of the Iberian fortress when build inside a wall. Additionally 'entry' and 'entryTower' now have the same width as gates of that wall style. Here's an update (only SVN): wall_pack2012-5-1.zip EDIT: An even newer version is already in SVN. Iberians now get walls as starting entities by default (as their civ bonus). In some maps that doesn't provide enough space (mainly naval maps) only towers are added.

Well, than it makes sense. Huns and Mongols where quite capable of rampaging through the roman empire with massive amounts of mounted ranged units

This would be fixed by preferring player given orders over anything else as well. So I don't see the need for stances here neither.

Glad to hear that Since a priority system is coming I think it could easily include towers. So you might be lucky here . However, I think it should be the nearest target the tower can deal reasonable damage to. If it's a ram and there are other targets e.g. a war elephant prefer the war elephant. If melee units are attacking the tower they are closest anyways. Edit: On the second thought I rather like the random targeting as it is.

As far as I can see the length of the civ's wall entities (e.g. rome_siege_wall_ short/medium/long) seams to have lengths adding up to each other. As far as I can see it's going to be 12/24/36 units (3/6/9 in rmgen tiles) for short/medium/long wall entities (I use rome because I was told by Mythos_Ruler all other civs would get walls of the length as Rome already has). I thought the purpose of walls of differing length was to be able to build walls of any length without much overlapping. But since you can build the 'medium' wall out of 2 'short' walls and the 'long' wall out of 3 'short' walls or a 'medium' and a 'short' wall it somehow defeats that purpose I saw (could be totally wrong with that though). Using lengths with differing prime factors would do much better for that (like 12 = 2*2*3, 20 = 2*2*5 and 28 = 2*2*7). I don't know how the models are build but it would be easy to use repeated parts in the middle with a length of 8 and endings with a length of 2 for each side (so they add up to 2*2 + 8 = 12, 2*2 + 2*8 = 20 and 2*2 + 3*8 = 28). That way one could still build walls without overlapping of length 24 and 36 (2*12 and 3*12) but also of length 20 (1*20), 28 (1*28), 32 (12+20), 40 (2*20 or 12+28), 44 (2*12+20) and so on. So though the shortest wall is of length 12 composed walls of length greater or equal to 20 can be build without overlapping of any length that is a multiple of 4 (the prime factors in all of the walls lengths: 2*2). Not sure though what purpose walls of different length had in the first place so please tell me what was the purpose or what you think of it. Thanks in advance.

I'm a bit confused. I thought ranged units should be able to run out melee units due to the lighter armor like said e.g. in this post. Perhaps you mean melee cavalry to chase down ranged food units but what about ranged mounted units?

Yes, this is useful as said in this other topic. I thought this topic however does focus on stances.

Agreed though a cancel order is only useful for player given commands since the unit AI will start right away with it's previous behavior afterwards anyways. But its good to give full control to the unit AI/stances as well.

I think stances should only effect behavior of units without any player given order. Maybe there could be stances explicitly overriding commands like 'violent' (attack) might as historic_bruno said in the list and an opposite stance 'evasive' (run away). Arguments for and against that can be found in this topic. Then the question arises 'What is a 'player given order'?'. IMO any order not given by the unit AI itself after a unit was idle after fulfilling a player given command. There is a gray area between player given and unit AI given orders however like: Gather/return resources and some future commands 'auto explore' and 'attack move'. But I think they still are player given and the unit AI only helps him to accomplish them. In case of wild animals I agree though elephants will most likely just kill gatherers or even citizen soldiers. Making a single unit attack or run away will have no big game impact anyways: A single unit attacking a bunch of enemies will most likely die without dealing much damage. A unit running away for a few seconds and then return to it's duty will most likely be attacked again - and die a bit later. If they would just for example gather further or premature return resources if attacked they have better chances to survive or draw the enemy into better defended parts of the own base and so help killing them while still gathering resources. If the aim of stances is to make citizens stay alive a 'return resources if attacked' or 'auto garrison into a building with an attack if attacked' behavior would do better IMO. For a bunch of attack units standing idle gathering somewhere to attack or as a defense force however it's a different thing. Stances could be useful here. Then the question arises: 'What exactly should stances effect?'. IMO it should only be the distance they walk away trying to attack or evade the enemy from the point of order of the last given player command (or where they where when the player command became invalid like attack a unit that dies while moving towards it. In this case it might be good to let the unit walk to the point of death of that unit to make it gather with other units with the same command). In this case evasive behavior is quite useless because maybe the units will die slower but still won't do any damage to the enemy. BTW: The units should first check if they can get into attack range following the stances restrictions before starting to move. My opinion in short: - Stances should only effect 'idle' units (don't interact with player commands) - Defensive stance is only useful for units without an attack and so may as well be the default and only behavior of those units (if idle). When gathering/scouting/garrisoning they should still do that and only that (while defending against wild animals might be a part of gathering in the wild) - For units with attacks an aggressive stance is the only useful so it may as well be the default and only behavior (if idle). - To act more aggressive the 'attack move' order can be used. In the end there is no real need for stances IMO. I didn't mention the problems arising when having attack priorities, formations, minimum range and attack restrictions in addition to stances here because it seams a topic mostly focusing on stances. This all might be considered before reducing the number of available stances like planed by Mythos_Ruler. Since I'm obviously not a friend of stances perhaps someone more delighted about the idea could explain what stances should achieve in detail.

1. I totally agree! 2. That might be good in some cases but not in all. I don't like features assuming something that is not always true. In this cases the player should have to decide/act IMO.

Thx much for the reply. I think my problem is made clear now and maybe you can discuss it further inside the team especially those who are desperate for formations just to make sure one can get what the others ideas/goals/interests are. But it was really important to me since I think it is a somehow very basic part of the games behavior and I saw it running the (for me) false direction. I played that many RTS games and most of them had stances/formations OR attack move. I'm hopeful that it will not be like that for 0ad and attack move will be available since it is more useful than stances/formations could ever be IMO. By the way: Attack move doesn't need an own button IMO. Just an attack button should be added to all units with an attack (static entities like buildings too, IMO). If no unit but a point on the (mini)map is chosen as target the attack move order is applied. The problem with target priority/ability to attack some unit types at all and the resulting 'chaos' on the battlefield is an issue still present of cause but I pointed it out which was my main concern. I'd be glad to offer other ideas (e.g. more detailed priority system) as well if that is wanted. I think it's better if the player has to do things to make his units act more effective in battle (though they are in basic handled by the unit AI) than the unit AI trying to do it better but fails and in addition disables the player to make corrections. As I said before: If the unit AI CAN do better than the player (without cheating ) it's more than welcome! I'm sure now I can sleep better

Thx! I thought so since I tried to fill a siege tower with them but it didn't seam to have any effect. Was not sure though. That makes a sensible town bell function difficult though. On the one hand side you don't want to loose your females but citizen soldiers are more expensive and so less expendable. - Most realistic would be garrison females first to secure them and let the citizen soldiers attack IMO. - Most effective would be garrison citizen soldiers first if enemy non-siege weapons are in attack range of lets say the civil center and let the females that have no space to garrison into the building gather around it trying to avoid attacks. That way you deal full damage and loose less valuable units.