[RELEASED] Emerald AI 3.2 (New Sound Detection) - The Ultimate Universal AAA Quality AI Solution

Is anything new planned for this, the updates seem to have stopped.

 

1. It needs the Navmesh
4. Trial version? Buy the app, you wont be disappointed. Never in all my time have I seen a trial version on the asset store.

 

The team claimed that the Emerald AI would be out before the end of the year, which is obviously not happening now. I wonder though, maybe we can get an update here in the forums from a team member letting us know how progress is coming along? Looking forward to this next update and also will probably move away from TOD for day and night and integrate UniStorm so that the two can work together nicely! We also use UBER and I hear the two play together nicely via a little custom scripting.

 

@Razputin That's strange, the AI shouldn't generate waypoints that are too steep or hard to get to. We will look into it.

We are still working on Emerald. It is simply taking a little longer than expected. Our apologies for the lack of forums posts on this thread.

Due to the holidays, we will not be able to submit the update by tomorrow. This update ended up being quite a large update, but we are updating and adding quite a few things people requested and fixing issues. We can assure everyone that we are doing everything we can to get the update out asap. We will post here when we are close to submitting Emerald.

Great to hear about the migration to UniStorm. UniStorm integration will come, we are working on providing an example in the next version or on the Emerald Wiki site. What features specifically are you looking for with Emerald and UniStorm.

 

Hmm, unfortunately it is not for me then. Part of my world is dynamically generated. But thanks for answer!

 

I intend on using this for animals in my free roam mode, but I'm more interested in the random human encounters, like bandits and such, like the knights in the video.

 

I was hoping to use the new update for the core of my NPCs/Enemy, but as its not ready I've switched to another AI / Time of Day system - my purchase was based on the 1.2 update. Although I still plan on using this as it's good to have a few different AI systems in your game, so they do not look or react the same way. Hopefully it'll be ready before the end of Jan. Good luck and keep us updated.

 

All of our world is generated on the fly as well, but from my understanding the new update was supposed to include alternative options for pathfinding, besides being stuck with Navmesh.

 

Hello, Happy new year 2016 I send my best wishes

I have bought fhis for using with some dinosaurs I have bought, and please, I want to know if two dinosaur that are agressive configured can attack each other and also attack the player

Also I want to know here is if this could be possible to be used for moving vehicles, I have Edy's Vehicles physics https://www.assetstore.unity3d.com/en/#!/content/403
and would like to know if this Emerald AI could be used to attack other cars in a demolition derby way, I mean that I could attach the AI to vehicles so they could be agressive between each other and search for other cars and in for the player too, that would be the same as another car.

Also want to know why there is only 3 animations to choose when wandering and not possible to have 5 (this is what my dinosaurs have)

I will also be waiting for the new update to arrive that I hope it can be available and released soon

Thanks very much for any help

 

I've been using this with Realistic FPS and it works wonderfully, was very easy to hook into that system. If anyone is interested I can share the changes I made for that.

The new NPC AI is cool and I will definitely be using it but like others have said I would keep it limited to the basic behaviors you have right now. Personally I would like to see added:

-Animals set up with swimming animation / behavior
-Birds and other flying animals
-Animals set based on nocturnal or daylight behavior (sleep during day or hunt during day etc)

 

I would be up for you sharing the RFPS set up with this system ..not looked into it yet but would save me some time if you would be so kind

 

Sure!

RFPS:
In WeaponBehavior.cs, add this to line 1117 to check for animals

Code (CSharp):

1. case 13://hit object is an NPC
2.                 if(hit.collider.gameObject.GetComponent<CharacterDamage>()){
3.                     hit.collider.gameObject.GetComponent<CharacterDamage>().ApplyDamage(damage, direction, mainCamTransform.position, myTransform, true);
4.                 }
5.             //NEW: Modified to support animal AI
6.                 if(hit.collider.gameObject.GetComponent<Emerald_Animal_AI>())
7.                 {
8.                     hit.collider.gameObject.GetComponent<Emerald_Animal_AI>().Damage(damage);
9.                 }

In Emerald Animal.cs find all these instances and replace them

Code (CSharp):

1. //currentAnimal.GetComponent<PlayerHealth>().DamagePlayer(attackDamage);
2.                             currentAnimal.GetComponent<FPSPlayer>().ApplyDamage(attackDamage);

In WeaponEffects.cs add these lines around 120 and 173 respectively

Code (CSharp):

1.             //NEW: updated for animals
2.             case "Predator":
3.                 impactObj = fleshImpact;
4.                 if(fleshImpactSounds[0]){
5.                     hitSound = fleshImpactSounds[Random.Range(0, fleshImpactSounds.Length)];
6.                 }
7.                 break;
8.  
9.             case "Prey":
10.                 impactObj = fleshImpact;
11.                 if(fleshImpactSounds[0]){
12.                     hitSound = fleshImpactSounds[Random.Range(0, fleshImpactSounds.Length)];
13.                 }
14.                 break;

Code (CSharp):

1.     //NEW: updated for animals
2.             case "Predator":
3.                 impactObj = fleshImpact;
4.                 hitSound = fleshImpactSoundsMelee[Random.Range(0, fleshImpactSoundsMelee.Length)];
5.                 break;
6.             case "Prey":
7.                 impactObj = fleshImpact;
8.                 hitSound = fleshImpactSoundsMelee[Random.Range(0, fleshImpactSoundsMelee.Length)];
9.                 break;

Then make sure to assign your animals to the NPCs layer and you should now have killable animals that will attack and hurt you!

 

Ditto.. I purchased for non-humanoid AI. I would love to see the development focused more towards animal AI and behaviors and human/other race behaviors be done in a separate product. Maybe they could use a same core framework or something but adding all the NPC type stuff to Emerald is adding clutter for those of us who purchased specifically for animal AI. I'd rather see fish, birds, and other types of animal behavior added before NPC crap. But then even those could be done as separate assets built off the same core framework. Maybe at least separate out NPC, land-based animal, air-based animal, water-based animal code into folders so we can choose which aspects we want to include in our project and only install code related to those aspects.

 

I agree. There are a lot of human AI things out there and this asset is getting very heavy.

 

I purchased it for both. If behaviors are separate to NPCs then I don't see that there is a problem, you have the best of both worlds - one solution, rather than having to try and get two AI solutions to work together which can sometimes be a nightmare. Not all of us are making games that are just specific to animals. The asset will be limited if just for animals, opening the scope up to other AI based characters will only make this product have a longer shelf life and support.

 

There are AI packages for human that we are already using. I did not Emerald for this and am not happy with the features for animals being ignored for humans. I must admit, I do regret buying this now. I didn't at first.

 

From what I can see is that it was lacking in certain features. The 1.2 update will hopefully fill the gap, so you'll need to see it before making that call. Some of the other AI solutions just don't cut it, this one has huge potential for being a one solution does all.

 

Well.. It was originally called Emerald Animal AI and that was why I purchased it. As it stands now, that is still how the thread title advertises it as well.. The NPCs and enemies crap was added after the fact. And that is great for those of you who want that but for those of us who purchased as it was originally advertised ([RELEASED] Emerald - Animal AI v1.1 (Dynamic Wildlife, Predators & Prey, Packs/Herds, and more!)) now have a lot of bloated crap we don't want. That was why I suggested maybe breaking it up. Even if it is in the same asset, we could choose whether or not to install all the modules or just the ones we need. To be honest, I was really happy to see what was planned for animals but feel a lot of that stuff is being held up by the added bloat of NPCs and enemy AIs. A tad disappointed myself. But hopefully things will at least get restructured enough so I don't have to install the extra bloat since I don't need that stuff.

 

I mean there can only be so many features for animals.. and I've helped a few AI solutions with ideas;

What more would you need for an AI solution?

For animals and NPCs: Behaviors and Animation States (Legacy or Mecanim) - you would have generics that would fit both: Idle (More than One) Wander, Attack (more than one), Flee to Destination or Random, Got Hit, Attack (more than one). Patrol Waypoints, Graze or Do a Task (at each way point or set marker). Death sequences, Healing / Sleeping.

 

We've not seen 1.2 yet? So we don't know how its going to be structured. Currently I've tried v1.1 with zombies and it works great, but lacks in death animations which 1.2 will cover. The herding (alpha) works the same for NPCs as it would with animals. Whilst my game is not just Zombies, but Aliens, Mutants and Animals having them both interact or fight each other as well as real human interacting is beneficial. If you have a drop down for each type of AI (Human/Animal/Other) then this would be good. My next project is Knights/Magic based, which made my purchase easy.

How do you know its bloated crap? To be honest from what I've seen I don't think it is, its giving you multiple animations, multiple sounds, eventual interaction with their Weather/Climate control on UniStorm. Breeding/Feeding and more.

 

In an fps where all you do is shoot everything and blow crap up, probably not much. In a much more evolved game with the addition of professions and richer story lines, a ton of stuff is still missing. Read through the thread. I'm sure you can find a ton of stuff suggested in the first five pages which isn't there yet. Some of it is coming with 1.2 but probably could have been released with 1.1 if the some of the non-animal stuff had been left out.. And again, I'm not pushing to get rid of that stuff. I'm pushing for a modular approach so those of us who purchased for animal AI can get what we paid for without the extra bloat of something we will never use. If you want to install everything, great. If you just want the animal AI, great. If you just want the NPCs and enemies AI, great.

 

Any bloat is crap if you aren't going to use it.. I'm not saying the code or functionality is bad.. I'm just saying it is stuff I don't need that is beyond the scope of what was originally advertised.

 

My game is not a total fps, its a survival game with many elements, animals being a part of that. Interaction with NPCs / Shops and trading, so if there is scope to interact with NPCs then I'm all for it, else I'll just bolt on another asset or make my own. I doubt you'll get everything in one solution, not straight away. Sometimes it can take a while especially if they have multiple products to support.

 

Then its there for those that do

 

Emerald AI first launched v1 in June 2015?, v1.1 came out in August (9th) 2015 (3 months later). If v1.2 comes out in Jan at some point, then that's only 5 months to put together an impressive update, in addition to a UniStorm update.

 

Exactly my point.. Which is why I bought Emerald Animal AI, not Emerald Everything AI.. And not sure why you are quoting dates. My point was if the NPC and enemy stuff wasn't added for 1.1 that more of the stuff coming in 1.2 might have been able to have been in there. Nobody was questioning or complaining about the time it was taking.

At any rate, we are just debating semantics at this point. All I was suggesting was to modularize it so if there are AI families we don't need (or never wanted), we don't have to bloat our projects with the extra code. Separating AIs between NPC and enemies, land animals, air animals, and water animals with a common core framework seems pretty logical and I don't think it is an unfair request. Nobody asked them to take anything out. Afterall, the only thing any of us want for our projects are ways we can get the best performance possible. Being able to get rid of something unnecessary is one of many ways to accomplish that.

 

I believe if you scroll way up to the top, I actually asked the developer if he was going to bloat this with stuff and he said no.

I also bought this for animal AI, which is how it was described. I have very good player and NPC AI. I don't want to switch.

My biggest issue is that the animal AI is not moving forward. We are not getting what we promised. I have given up, actually. I will write a review so others do not buy it for animal AI.

 

No I was merely pointing out that its good going for the developers, what they have achieved in such a short space of time, and that you can't just put everything in one release, it will take several. Sorry, it kinda sounded like you were complaining it was taking a long time for the next release, when all of the features you wanted should have been in v1.1 if they hadn't concentrated on the Non animal stuff, but really its not that much different! My mistake

 

v1.1 is not really bloated, I feel its not complete as such and v1.2 adds what I feel is missing (multiple death animations etc). Isn't all you asked for, being implemeted in v1.2 ? Products evolve to cater for a wider audience, it makes more sense to expand on what you have than start again especially if this is their bread and butter. Before you complain wait for v1.2, why push people away from a product that you want to see do well - doesn't make much sense to me?

 

I am sure the developers have a vision for Emerald and are working towards it.

I know from personal experience with Gaia that you are trying to deliver the best value you can for all your customers - and several of my ideas were seen as non essential - and when delivered - liked. Everyone has different requirements and perspectives.

Looking forward to seeing the next Emerald

 

Did they also specify a "YES" we'll add that in, or we'll look at that at some point. If they said yes we'll add that in, maybe they will at some point once the core of the product is done. If they said, maybe/we'll look into that then that's not a Yes answer. You shouldn't give up, its a great product.. they seem like the developers that care about their products and expand on them and appear to be around for the duration. I can understand that you may have been burned with other products/developers not adding features promised as you frequent many of the forums that I do read or contribute towards. This doesn't seem to be the case here, just wait for v1.2 and see what is next.

 

Yeah, I didn't really mean for my stuff to sound like complaints as much as I wanted it to suggest focus on the animal stuff and possible suggestions of ways to section it out. At the time Emerald was announced, the only decent animal solution I had found was HerdSim. And it is ok, but Emerald promised so much more for animal AI. Me? I'd love it to deal with specifics for animals, fish, and birds. You sound like you'd like it to handle those and any humanoid type AI. Someone else might want just animals or just fish. Down the road someone may want mechs or tanks or ships. Each and every one of those could really be their own asset containing the same core library and a folder that has that specific AI. Or there could be a master pack that contains them all or an animal bundle containing all the wildlife ones, vehicle bundle containing all the vehicle ones, humanoid bundle containing all the intelligent race AI, etc. I guess I wasn't trying to complain, but offer a suggestion of how to restructure to make everyone happy and maybe a way for BHS to put a little more cash in their pockets by creating not only a bundle of everything but also smaller individual modules as well.

I was just excited in the beginning to see such focus put on animals and now it has expanded to take some of that focus off of animals. So I'm disappointed to lose that focus and wonder what type of unique stuff could have been done with the animals in that time, but still holding out hope that BHS in the end has a solution that makes everyone happy.

 

Looking forward to the next version, but agree animals are a good focus for future updates. Birds and fish, which is what I need also. I would imagine that you could do this pretty much right now with fish, lowering the nav agent and making sure they cannot go above a certain height on the terrain? Same would probably apply to birds with the nav agent being set to the highest, so it still uses the Navmesh, but it would probably would remain as a set height and not swoop down to attack. Ideally my idea of this would be to have a sphere or square box and set path points in the sky for the birds to follow or random based on the terrain nav mesh reference below where they would follow a set path, either that or a spline to follow, and have landing points where they would break away from that to feed / rest or attack, then go back to the spline to follow.

Its good to make suggestions to help the developer move forward, it may take time and fit everything into the developers goals for the product, if viable.

 

"This update ended up being quite a large update, but we are updating and adding quite a few things people requested and fixing issues. We can assure everyone that we are doing everything we can to get the update out asap. We will post here when we are close to submitting Emerald."

I'm sure they will get there, and they are listening. Anyway back to work Look forward to the update soon.

 

The core product was supposed to be about animals.

We are moving on to something else. It might not be any better in some cases, but the developer is easier to reach and talk to about things. So who knows. I should have known better and will just chalk this up to another mistake.

Thanks for your kind response though, Julianr. I appreciate that you respected my opinion even if you didn't agree.

 

@Shawn67
@Teila

To offer some clarification, we aren't doing away with any animal features nor will it distract from the overall purpose of Emerald.

We wanted to create a system that offered dynamic wildlife with predators, prey, wandering, breeding, etc. We created a system that hadn't been done before, which is why many people loved Emerald. We will still keep these unique features. We are simply expanding support so our users can use Emerald for both animals and NPCs.

The editor won't change for the animal portion. What we've done is divide the editor into 2 sections using a pull down tab. This pull down tab allows you to choose Animal or NPC. The Animal setting will be the original Emerald editor you are used to (with the addition of the new animal specific features we've mentioned) and the NPC setting will be the new NPC specific features. All we are doing is adding additional support for NPCs. The original Emerald editor will look and function exactly the same as it does with Emerald 1.2.

If anyone has any more questions or concerns, please ask.

@julianr Thanks for helping out with clarifying. Animals will certainly be the main focus of our updates.

 

No problem - just don't want you to make the wrong choices, as if Emerald fits your needs now, waiting for the 1.2 update, or even 1.3 would be better than keep on switching solutions, it slows down development and could cost you more time in the long run, plus its a pain to keep switching products - very frustrating too especially with time constraints. BHS do not look like they are going away any time soon - they have a good set of products that would be sustainable for some time, as long as they keep up the good work and support then I doubt any other product would come close to what they have planned.

 

Waiting has already slowed us down so we are moving forward. No problems switching. Thanks for the concern though.

 

Great! I'm looking for that! Now I wait for the big update of Emerald Animal AI (and RFPS).

 

just wondering if there is update for ai companions thanks.

 

Companions will come in the update after 1.2. The 1.2 update is a big update containing a lot of fixes, improvements, and new animal and NPC features. The 1.3 update won't take nearly as long as our 1.2 update did.

We will be posting some new screenshots of what we've been working on sometime tomorrow.

 

We weren't able to get some screenshots of the new Editor, but we will be able to sometime tomorrow. There's a few more tweaks that need to be done before we do so.

As for a status update on our 1.2 work, we have a list below of all the stuff we have done since our last status update. We are getting close to finishing this update and we appreciate everyone for waiting patiently.

We tried to explain this the best we could. Hopefully everything makes sense because it's quite awesome. There's a few more things we will add to our next status update before we submit the 1.2 update. Emerald 1.2 functions quite nicely and is incredibly versatile. It should be able to find a place in pretty much any game.

We tried to design the editor/system to accommodate those who want only animal AI, only NPC AI, and those who want both. Note: That most of these features and improvements apply to both the Animal and NPC AI. We also didn't list everything we did. There is also a lot of fine tuning that was done.

• A pulldown tab of either NPC or Animal will separate the two types of AI. Each setting contains exclusive editor features and functionality to keep both types separated. All NPC settings and features will remain on the NPC setting and all Animal features will remain in the Animal settings. This is to ensure that those who want to use Emerald for animals only can do so without having to deal with NPC features.
• Both Animals and NCPs now have 4 Behavior Types: Cowardly, Passive, Aggressive, and Defensive. A Cowardly AI will flee when attacked by another AI or player and will not fight back. A Passive AI will never attack. Even if hit, it will wander around within its Wandering Range. An Aggressive AI will automatically attack both Enemy/Predator and Player tags that enter its Attack Range. A Defensive AI will only attack its Enemy Tag if it's within range. If an AI is attacked first, it will attempt to attack and kill the attacker. This is perfect for AI like guards who can patrol similar to Skyrim, for example.
• An AI will now pick a target that's nearest to it. This also goes for group fights, if their target is killed, they will pick the next nearest target to continue attacking.
• Animations have been improved and now blend better when changing actions.
• Hit Animations now work properly.
• Tags are now based off of Enemy and Player (NPC only). The Enemy Tag is used to determine who is an enemy or not. For example, a good AI with a tag of Ally can have their Enemy Tag be "Enemy" and a bad AI with a tag of Enemy can have their Enemy Tag be "Ally". This allows both AI types to work between each other. Their actions and fighting styles depend on their Behavior Type explained above.
• Player Tags are used for both Enemy/Predators and Ally/Non-Aggressive animal AI. Enemies, with an Aggressive behavior type, will attack a player as soon as they step into their Attack Radius. However, Allies, with a Defensive behavior type, will only attack a player using the Player Tag if a player hits them first. All this also applies to Animals. This allows users to have the option to have defensive animals that only attack if they are attacked first.
• The AI's trigger detection is now added automatically to an AI's Skinned Mesh Renderer. This eliminates the problem with the trigger detection detecting damage calls, which sometimes resulted in hits being detected from too far away.
• A new Attack Delay algorithm that sets an attack delay to match opponent's attack animations. Most animations need some sort of delay before they reach their target. This system handles it automatically, if animations are enabled. This feature can also be disabled if desired.
• The hit animation is also based on the Attack Delay. When AI are fighting each other, their hit animation won't play until the hit time is calculated. The hit animation will not overwrite an attack animation.
• Added support for Death Animations to the editor as an option for when AI die. (This is on the same gameobject).
• Wander distance is now based off of an AI's starting position. This fix the issue where AI would wander too far away from their original spot.

 

A couple of the things that were previously showcased or talked about that I don't see mentioned here that I have been looking forward to are the breeding and companion systems. Are those still slated for 1.2 release?

So we can no longer have different animals behave differently to NPCs? Do these tags have to be Ally and Enemy or are we free to configure based on our selection of configured tags? Any way we can make these selections to be an array of selected tags? Instead of working off of tags, I really wish this could be driven off of an internal variable configured with each species. It wouldn't be so bad if we weren't limited to a single tag assigned to a prefab.

Thanks! So this will mean those of us not using rag doll who have models that have both a dying and dead animation to use both now? I think currently we have dying but I remember asking for the other. Having trouble remembering but I think this is what you are saying, right?

 

These aren't all the 1.2 update notes. This is simply what we have done since our last status update. The breeding system for farm animals (as shown in our demo video) is still in place for 1.2. However, dynamic mating and the companion system will be saved for the 1.3 release.

The tags don't have to be labeled as Enemy or Ally. This was simply used as an example. However, we are looking into having multiple tags for each category.

Yes, death animations will now work as a way of dying. You no longer have to use a replacement object. Emerald will also automatically disable all components on dead AI to help with performance.

 

Awwwww... but.... but.... Seriously though, I'm happy at least the farm animal one will be in 1.2 so we can play around with it and maybe spark some ideas of ways to use the dynamic one down the road. I'm assuming it could still be used on other animals too, but just requires player interaction? If so, maybe it could still be used as part of questing or some animal husbandry/naturalist type professions dealing with wildlife.

Bummer on the companion system as well. To be honest, I kind of figured it was more of a 1.3 thing since it isn't something you've shown any work of in videos. Only so much time in the day, right?

Other responses sound good. The whole multi-tag thing would be great. The way it is now is good for the basic system but lacks in the ability to further define animal relationships between species. Of course this might be something I could do with @TonyLi's Love/Hate system as well. Still wrapping my head around integrating that into my project and the possibilities it opens up.

Oh.. Speaking about disabling components on death, might be asking too much but is this something where we could add components that would also be disabled? I can't think of one offhand at the moment as an example, but there might be a time that we would want to disable additional components after death that aren't actually part of Emerald but may not want to just do a disable all kind of thing. Not sure how that would work but at least something to think about when you are working on it.

Thanks for the quick responses!

 

Awesome set of updates.. can't wait to try it. Suggestion for a future release .. what is lacking in AI systems is the ability for the NPC to change weapons. My own AI system in another engine had the ability to change from ranged weapons to melee based on a range. You could add as many weapons as you like and the AI would choose them randomly / in sequence, or if the AI had any ammo left in the weapon and no reload clips left. It would make it more realistic.

 

Forgot to add for the Animals; they could also run a sequence / random attack animations (bite / claw / swipe) with different damage levels.

edit: forgot to add - not only attack, what about defensive measures... blocking/defending for NPCs or animals with some form of armor (like horses)

 

Does the AI autodetect your player just entering their range, or can line of sight be used too, as I'd like to sneak up on my prey, human or animal undetected when possible.

 

Please excuse me if this has been answered previously:
Im considering purchasing this asset for my game. I have read what I can on the Wiki, and would point to page http://emerald-animal-ai.wikia.com/wiki/Tutorials
On this page is describes setting up player damage for UFPS and the Unity FPS controller. As my game is 3rd person and I am using my own custom controller, can the Unity FPS instructions be applied here. I am not using the character controller or FPS controller, but it is a relatively simple system, although multi player. I would need to synch the transform and animations of each animal, and also synch the damage report to the player.
Although I dont expect a complete run-down of the process for doing this last, I am prepared to experiment a bit with that, using your damage system may be nice as I havent yet implemented that feature with my game, and if I know it works with Emerald Id be half way there...

Mark

 

Yes, that's correct. To breed a player must activate breed mode for that particular animal. However, this is done with an external script so it can be modified to suit user's needs until we release dynamically breeding animals.

We are looking into other options for the tag system. What would be ideal is being able to use a string rather than an actual tag. This way users can have an array of options rather than just one. We have a few ideas of how to do this and we're going to try them to day. We will post our progress later with how it goes.

This could be a simple script that we could add or that can be created. It would just need to watch the EmerladAI script and when isDead is true, disable the extra components you want.

We're looking into support for ranged attacks for 1.3. We also don't want to pack too much into one update, like we did with 1.2. We would rather have every couple of week updates that add various features, rather than having 50+ new features that users have to wait months for. We will see if we have time to add it in the 1.3 update. If not, we will add it in the 1.4 update.

Random attack animations are already finished with 1.2. We have also added random damage amounts based on the Min and Max amounts set within the editor. They aren't based on individual attacks though. It's something we can look into adding in the feature though. Defending is another thigh we would like to add in feature updates.

Thanks for the suggestions.

Damage is handled though the Damage() function. This allows you to call the damage amount however you'd like. For the demo FPS damage system, we use a RayCast from the center of the camera. You could write your own system that does the same, but instead cast the RayCast from your player's position.