RPGista

Just for the record, as in this topic: http://forums.thedar...post__p__272122 AI should be able to eventualy detect he was successfuly pickpocketted when he tries to access his "should be closed" locked door and is unable to open it due to not having the keys with him anymore. This raises several problems though (expressed in the topic above), but is still a realistic/common sense feature that is lacking right now. AI could then either fall back to the remaining path corners after a light suspecious reaction (problem: wouldnt make much sense in missions like Tears of Saint Lucia where the AI in question only has two path corners, one inside and one outside the locked door, effectively breaking his route); or go to full alert mode, raise the alarm and go into active searching behaviour (but would eventually have to get back to the route again, thus going back to the previous problem? I dont know enough about it to really give an good opinion here).

Yeah, very cool, and Im liking the whole fantasy feel to the mission too, very creative indeed.

Yep, now that you mention it, it doesnt seem easy to solve... The only reasonable way to avoid the "blank stare at door" scenario would be to have mappers always creating alternative routes for AI to fall to once they have been pickpocketed (after a "Oh, I seem to have lost my key" kind of reaction), but that would brake already existing maps like Lucia, and thus highly problematic... The only other way I see for this would be to have the AI going into full alert whenever he realises the theft ("There's foul play here" kind of reaction, like spotting a weapon or being hit by an arrow), enough to send him running for help, bring reinforcements and the whole show; but even then they would have to fall back to normal behaviour eventually (right?), and that would make it very unforgiving, if not impossible to ghost a mission, specially when you pickpocket a key that you dont know where the respective door is, the AI is bound to get there first and then raise the alarm, and there's nothing you can do to avoid that but to KO or kill him. Still, IMHO, I'm all for realism and challenge, and in an ideal situation I would still choose that second alternative over having the AI ignore something so flagrant as having their keys stolen. I will probably take this suggestion to the AI topic though.

Deifinetly, I'm feeling some good vibes from the revamped AI, more testing is needed, but I was surprised by some sharper responses already! I really want to test the new general alarm capabilities, as I feel this is one of the areas which could have added a lot more challenge for sure.

There's one more dimension to this, as you can see in Tears of Saint Lucia - if you steal the cardinal's room key from his pocket, he'll still get inside his room no problem (he goes back and forth from it in his route), even though the door is always left locked. So in the future it would be also good to have AI finding out about the keys that allowed them to open doors have been stolen, when they actually reach the door and are unable to open it.

I have translated the training mission and Thief's Den to portuguese, but all I did was translate the text (replace the actual strings) and images (in Thief's Den case, manually editting), I dont think I even changed the original names of the files, not to break any links... I believe they are the ones listed on that server you posted? In this situation, if you play my translated missions, regardless of any language setting, you would get the portuguese texts right? I'll need to do it the right way when I have time... Either way, I have to get back to you Tels, I'm afraid we still have some work to do on the portuguese translation as I'm finding several things we need to improve/correct. Sorry man, youre having tons of work!

Pretty cool. But how do you envision the Pagan village? I always saw them as a kind of iron age settlement, like this: http://www.tollundman.dk/landsby.asp http://homepage.mac.com/venitar/Travels/StFagens/StFagans.html Anyway, good luck!

Unfortunetly no. There's no way to delete the vote, if the poll maker didnt allow it when he was creating the thread (as is the case here, for example).

Stumpy, if it's indeed a village you are aiming for, from my experience, they are rarely walled like that, specially by such a tall wall, those would be actually impossible to make. There are many little towns that were incorporated by the castles' walls, but medieval urban enviroments (even in TDM's time that is roughly equivalent to baroque) would be pretty much organic, "improvised" kind of settlements, so I would suggest less strict geometry and a healthy dose of disorder to your map, Tears of Saint Lucia town is a beautiful example of that, also a Score to Settle, but they are full blown cities... What kind of village is that? I can see a lot of effort on details and the cemetery is intriguing, I think some references would make it easier for you to think of what to build.

This is amazing, I love that list of features, can't wait to check out that translation system, all the new AI improvements, animations... Brilliant work.

Nice to hear! The only thing I would point out at this point is that filters stay active between sessions, I dont know if that's intentional, but I got a few scares there before I remembered to check them. Im pretty sure this happens on other versions too, not just the latest one.

So you have these fantastical creatures flying around and all you can think of is to kill/wipe out all of the world's fauna? Don't you guys have a realm or a princess to save or something? Joking, but I remember when I bought Dragon Lore 2 and what you actually wanted was to earn the respect of everyone and the right to have a dragon ally (to be chosen by him). My beef with most free roam RPGs is that the whole point is to go from this lowly squire to "the meanest m*f* in the valley" (trademark - Sean Penn in Causalties of War), killing for points, and finally, sporting the uber sword twice your size and the black magical armor (like the WoW Simpsons episode spoof), rule the land by force. It's kinda pointless for me, I really like tight storylines and the npc interaction a lot more. But yeah, I finished my rant! I should check some more videos to see what's this game all about.

As a newcomer I always saw polls as merely a tool for map authors to get a concrete feel of the success/failures of his/her own missions, a kind of "numeric review", if you will. Whenever I play a mission, I try to understand what the author really wanted to achieve, the intentions, the overall feel, the execution, and thats my only criteria to vote - how well he achieved what he wanted. So missions like Heart of Salvation, Score to Settle or Thiefs Den were all pretty much perfect in their own different scales and scope. Creeps is brilliantly executed, but has slightly simplistic architecture on parts of the map - by voting you can tell the author what he got right and what wasn't as great, and thats all there's to it, for me. I made the mistake of giving it a sufficient score on visuals because I was giving missions that were a lot more complex (architecture wise) "good, but not perfect" ratings. After thinking for a second, I realised that the atmosphere of that map alone easily surprass that kind of nitpickng, so thats a perfect example of why missions shouldnt necessarily be compared to others. Missions with high ratings would basicly mean their authors achieved exactly the gaming experience they were going for, and not necessarily that one mission is "better" than the other. Crude missions could very well be kept from the suggested or the downloader list (freely available in the community, of course), but to tell you the truth, I plan on playing ALL missions available and havent checked ratings or reviews for any of them before playing (no spoilers), Im now going in chronological order, actually.

Yeah, I think it goes without saying that the AI is already pretty fun and sometimes scarily unpredicatable, awesome job, otherwise there wouldnt be any fans here, like me. TO Springheel and Grayman, really nice to hear about those modifications, can't wait to play the new version.

Great find Baddcog, good luck with the porting work. To Theothesnopp, liked that castle and the stone detail on the windows, but I suggest you consider adding a brighter, dark blue color to your outside, give it more of a moon shine feel. Loads of nice work around here.

Ok, here are some more suggestions: + Collision = AI should be able to feel/notice whenever a ) the player touches them with a held object; b ) they are hit by objects thrown by the player. AI should be vulnerable to damage from certain types of thrown objects: heavy, solid ones like rocks or bottles - this takes me to another request: a new attack dimension, thrown objects as weapons; if you hit a guard with a rock on the arms or chest, a pain anim could play and he will become alerted, but if you hit him right in the head, chances are you might KO him; thrown daggers could actually kill (new small weapons could be incorporated, like throwing axes, or a spear?), etc. But this is a different scope, for now, I would only suggest AI can, not only be alerted by the impact, but play pain animations when hit by certain kinds of objects. + Combat = Like suggested, AI could be more active in combat, strafing, assuming temporary offensive stance (2-3 attacks in fast sequence when very hurt for example), or defensive stance (blocks with little counters), etc. This would keep the duels more varied, not so much A attacks, B blocks, B attacks, A blocks. A lot depends on animation, like leap attacks, trips or taunts, so I wont mention them here. AI sometimes tend to run right against the player (they will keep running against your chest), sometimes when charging an attack (seems to be a case of wrong timing of the blow). As mentioned, there seems to be a bug where AI will swing at you without drawing their weapon. + Awereness = AI should be able to notice moving visible objects and basicly treat them as they would a visible player - get suspicious. So if the AI is in a room, and the door suddenly swings open, even if the player is hidden, the door itself might be in plain view and thus be "spotted" by the AI as he would the player himself (with different suspicious levels and sounds though). This would be true for held objects moving about in full light (again, even if the player is hidden). Noticing moving objects would be also useful in neutral situations, as the AI would notice the player moving objects that don't belong to him and that could increase agressiveness levels. If moving objects awereness is not possible, AI should be able to identify changes in state for tagged objects, and react - a locked door that is "open" (when it shouldn't) would be noticed by the AI when he sees it (or interacts with it, if it was left shut by the player). Food that was eaten would be noticed, etc. The point is, not only the state of lights but of various other objects should be sensible for the AI. Most important for me, AI should be able to notice when "tagged" companions are not where they are supposed to be (they are missing = were stealthly taken out by the player). This was talked about here: http://forums.thedar...ce/page__st__50 . As it is now, it is easy to exploit their lack of awereness of each other, and just take out one by one while they are alone in their routes. It should be possible to define "teams" of linked AI that would notice each other's absence, and that would make them alert or even go about looking for the missing one(s) (going in the direction of their path corners perhaps?). A possible way of doing it, as discussed in the mentioned thread, could be to set a timer on the AI's path nodes that would start an alert signal after X time the AI's state changes (from alive to dead or KOed), which seems to be similar to the way the valauable items awereness works. The timer would simulate the time the other AI is willing to wait for the victim to come back from his route. After it runs out, the path node would alert any other AI that its "owner" is missing, when seen. Sorry if this description is complete crap, I'm not a coder at all. The point is "If guards were to be alerted when certain companions are not where they should be, it would effectively make it much harder to kill them with abandon, as long as you do it quietly".

I'll be probably travelling close to xmas, but untill then, no problems for me, can be whenever you guys prefer.

On beta-testing this mission I was definetly amazed by the amount of effort Bikerdude put into this, we would bombard him with bug reports and he kept his cool and got an immense amount of things done in no time at all (even helped another member finish his own map in the middle of the contest), we were all aware of what was missing but the point was to put it out anyway to honour the competition and that was made clear right from the beginning, that the mission was playable but wasn't what you would call finished. Of course criticism is always welcome guys, and Sotha makes many valid points, but I think it's pretty cool that we have one more good mission with good potential to play, and even better, Bikerdude intends to expand on it and create the complete version it deserves, so that has to be celebrated. And I wont spoil, but I know some interesting sections had to be cut out of this release, so there's even more to the map besides what you can see right now just waiting to be added.

Thanks a lot guys, that did help, I only had to manually align a couple of textures after doing the copy from a parallel face. Oddly enough, I didn't realise before that the orientation of the original texture is important when applying it to patches...

Brilliant, nice work and thanks a lot for the explanations, I was confused there for a second. That reparent tool and the selection set will help for sure,

So I have a patch wall composed of many different sized 3x3 patches. Even with world coordinates and "natural" texturing, the patchs simply don't match side to side, as a normal segmented brush wall would. Is there another tool to auto-align textures on a group of patches that could solve my problem? Though there are info about texturing a single patch, there isnt a lot about texturing a series of them, and if Natural fails, what else is there besides manually aligning it (nightmare in my case)? The Texture Tool doesnt seem to be particularly easy to use in this situation either (aligned textured areas in the tool will still be offset in the patch group), or maybe I just doesn't know how to use it properly.

Yeah, I would recommend re-installing Doom3 and TDM on a "Games" or some such folder OUTSIDE the "Program Files" default one, as it is the most protected. Right click on the Doom3 exe later and check the admin rights on it. The whole thing really is a pain, and has kept me wondering many times why things dont work well (specially on older programs and games).

After some trials I was able to render the image, maybe it had something to do with file size, since they were all JPGs... So far, no new bugs. I'm also just doing basic geometry building, so I guess a more experienced mapper would be needed to check for advanced features; I was sure there was a problem with grouping brushes into a func_static, the new "grouped" entity would not carry the original shape's layer, and I would need to set the static to that layer again. But then I ran some more tests and it worked, it retained the layer of the original brushes (regardless of the default layer, which is what I wanted), so I was probably just being paranoid. It also seems to work the other way around, if you set a new layer for the grouped static, when you revert back the brushes also keep the new layer, which is perfect. Just need to test these things methodically, there are too many layers and little objects in my map right now... EDIT: Now that I read this, it doesnt make sense, if you want to group a set of objects all on different layers, the func_static should probably override those and set them all, along with itself, to the default layer? Could you please explain how it works exactly greebo? A question/suggestion: I mostly use func_static as a way of grouping many related objects (I'm also aware they have other technical aspects, but I'm just not there yet). I'm experiencing difficulties when trying to convert a series of objects that include grouped statics already, in this case you cannot apply group changes to the whole selection, such as reverting them all back to world spawn or making it all a func static - you would have to manually select and modify each object so you can then apply a group change to the now homogeneous set of objects. Would it be useful to be able to overide existing differences between world spawns and func statics inside a selection and just apply a group change to all of them at once, keeping unnaltered those that already possess the wanted state - ex: a set of 10 objects selected, 8 brushes and 2 func static, you want to overide that and make them all a single func static, without having to manually revert back both the existing statics and THEN applying the change. The problem I already see is that if you do the opposite and apply a revert back to world spawn on a selection of func_sttics that includes statics that can't be reverted, like models and etc... Unless they are already immune to the command. Maybe I'm thinking of func_statics the wrong way, as too much of a grouping tool. It would be nice to be able to group objects, then group groups of objects on "layers" (a group that includes other groups), and then be able to apply simple changes to all objects inside the group - textures and the like. Unless you create a layer for each set of related objects (so you can insta-select the whole bunch), there's no simple way to select them all to apply group changes, specially when objects in the same layer are scattered around the whole building (like a "wood structure" layer, for example). I do that by grouping them with func_static, but since you can't do any changes to them as you would with a normal selection of a set of objects, it gets a bit risky to be reverting them back to change something, you better not let go of that selection!, otherwise you'll have to hand pick all the related objects to be able to make them a group again. These are just ramblings, you CAN overcome all of these difficulties with patience and time, so not a crucial feature by any means.

Reporting: * Func_static issue: has been solved, in the sense that undoing and then redoing a "revert to func_static" operation doesn't makes DR crash to the desktop anymore, but strangely whenever I redo a func_static, it is rendered as invisible on the viewports - you can select it, and they will show up, but will only display correctly if you revert back to world_spawn. * The Background image still doesnt work for me, all I get is a square plane with no shader texture.

I was talking with Thor about this one, as he's also working on open areas, have you guys seen this video? http://www.youtube.com/watch?v=vWkhyIeEfjY I'm picturing something like this for your forest Pranqster! Joking, but it would be pretty cool to see such detail, I think we could get similiar results with Tels' SEED system? I'm working on learning it too, I'll use it on small sections of my first map as training.