A bunch of questions on aircraft

[gap]

@ vdr1981 Re. "Carriers spawning aircraft when they have a contact detected"

I have finally found the post where TDW describes the working of his patch. I also mentions a testing mission though I am not sure that he included it in the patcher. Hope that helps you in understanding why the patch didn't work during your tests.

Here's the link:
https://www.subsim.com/radioroom/sho...postcount=2193

[propbeanie]

Originally Posted by gap

With IRAI installed, they should. Nonethelesse, like in reality (or even ore), they are not perfect, and for using them each aicraft must be set carefully an tested. IIRC, the nastiest bug with them was they goings backwards once dropped. I don't remember what is the trick that fixes that, but in SH5 and SHIV there should be multiple threads and posts on this topic
Seems to me there were clues in Rubini's SH3 thread...

@ propbeanie,

thank you for your long answer mate. Please find my replies below

In one of your first replies on this thread, you had written that the detection of a submarine "qualifies" for the spawning of up to seven "response aircraft" at any given moment. It seems to, yes I suppose this is the max size of an "ASW response group", though its aircraft won't arrive all contemporarily, but in small groups of 1 or 2 aircraft every few minutes. This is correct From your last reply I also get that, after a longer interval, the game will spawn other "responde groups", even though probably we won't be anymore there for telling it. Correct. Response based upon AirStrike.cfg settings for the "timing" between the 'dice roll', or random number. If you stay in the location, you will see a 2nd wave, and if you are seen, a third wave later, etc.

Is my interpretation of your reports correct? yes

If yes, does the game spawn only one of such a "response groups" at any given time planes will spawn from the same group, until it runs out of planes apparently. I have not tested long enough beyond to see if other bases respond, or more that one group can be spawned if multiple airbases are within player's range? That is difficult to tell for certain, other than all of the planes seem to come in to "play" from the same generally direction. I would say only one AirBase responds, seemingly the closest. One way to tell for certain is to build a scenario and have two AirBases in a "quiet" area of the world, maybe just a few km difference in distance from the "target" area, and both with different AirGroups, roughly 90-130° separation at the location of your "target", and see what happens with the response when the "target" is detected.

Does the number of aircraft assigned to each base (beyond the maximum number of seven aircraft composing each ASW group) affect this scheme and, if yes, how? It does seem to affect the response. If there are only 2 airplanes total for an AirBase, that is all that is sent. AirBases with more airplanes do seem to have a maximum number they will send out at any given "response", based on the AirStrike.cfg counter. One aspect of this "response" is that it is much easier, as a submarine, to "see" something attacking you, than it is to monitor the airplanes attacking something else that might be beyond the horizon.

Within each group, do faster aircraft arrive first, or the order of arrival is totally random? It seems to be the roll of the dice influences the aircraft sent, as well as their spacing. Their flight speed does affect when they arrive, of course, such that it might be you would see all of the planes at about the same time, if the slower planes are spawned first, and the fastest planes last, or it might be that you would see them spaced out over a period of time, if the fastest came first, and then the slowest. They do seem to "fly by" the area, and then "return to base" along similar flight paths, though if another airplane finds something to attack, it does seem to generate a "report", and most of the planes when returning will then also attack. Several runs of a given scenario would all result in different airplanes, different flight paths, different attacks, etc., so that makes "testing" somewhat more difficult.

With the working method I have in mind, we wouldn't have "generic" airbases. Each airbase modelled in game would correspond to a real RAF station, and its air groups will elvolve reflecting its WWII usage by various RAF squadrons. Only squadrons relevant to the game will be considered:

- defensive fighter squadrons located near the coasts of the British Empire, equipped with "bomb-less", "short-ranged" aircraft so that they will only act as coastal defenses against harbor raids, attacking any close-enough enemy with their guns/cannons;

- maritime surveillance/reconnaissance squadrons whose aircraft should be equipped with flares and sea marker bombs rather than explosive bombs.

- ASW squadrons, whose aircraft will be initially equipped with ordinary bombs, gradually replaced by depth bombs and (where they apply) torpedoes and rockets as the war progresses and more specialized weapons become available.

- Anti-shipping squadrons, whose aircraft will be equipped with bombs, torpedoes and rockets.

- Convoy-defense squadrons, whose airccraft will be equipped with a variety of different armaments, depending on base location and on the most likely menace that they would have faced.

Possible gaps between one "operative" air group and the next one, will be filled with "filler air groups". See post #44 in this thread on what I mean for "filler group". Should one of the operative air groups cause any problem (as per your example), I will reduce the number of its aircraft to ridiculous amount, or I will replace it with a filler group.
Sounds like a plan!

Well it all depends on how the game will use those aircraft, but I see your point anyway.

Whatever is the most appropriate number of aircraft per squadron, I want to keep this number as constant as possible. If at some point, in reality, an airbase hosted two squadrons equipped with the same aircraft, in game I want that base to have roughly twice the number of aircraft it would have had if only one squadron was using it. Likewise, if historically a squadron had one or more detachments, in game it should contribute to each of its bases a lesser number of aircraft than it would if all its aircraft where "assigned" to one base.

When we attempted to build "appropriate" bases for a given RL air group, we would usually end up with skies polluted with airplanes. We put the 38th BG B-25 AirGroup at Port Moresby, New Guinea, but 72 planes seemed a bit excessive, so we made it 16 planes. Yet, a person could not operate their submarine between New Guinea and the Admiralities north of the Bismarck Sea, and Guadalcanal south of the Solomon Sea, for all of the airplanes. The B-25s were constantly in the air, and the Japanese of course, had to have an air response to the B-25s... lol - mass hysteria! Plus, they sank everything in sight! Nothing left for the subs. So we cut the B-25 group back to 6 airplanes and shortened their operating radius quite a bit. It is much better now.

[Demon777]

it looks like in SH4/SH5 all aircaft have "two sides of one coin"

on one side, we have almost stupid AI with no immersion to air warfare like in IL-2 etc. On the other side, the simple increase of aircart numbers may result in ending up the Battle of Atlantic 1-2 years earlier

[propbeanie]

You sir, are correct! The planes were not modeled to be a major player in the game, and the player was not supposed to want to look at the beautiful scenery and wonderfully modeled airplanes and ships of the game, but to rather stay inside his little steel coffin, water dripping from every surface, and grin and bear it looking through a periscope instead... lol - That way, you never see the airplane crash after it dive bombs you, or you never see that what took your boat to the bottom was the plane itself, and not its bomb, which had fallen harmlessly 30 meters away... lol - It is a balancing act to get a decent, appropriate airplane response. Plus, the planes do not "dog fight", so their only true natural enemy are surface ships and submarines... - unless, of course, they crash into each other, which they do from time to time... or crash into hills that are over 500m in height... "AI" need not apply for a job here. Only AD (Artificial Dum-dum) is accepted by the game.

[gap]

Originally Posted by propbeanie

Seems to me there were clues in Rubini's SH3 thread...

Weren't air torpedoes a feature introduced with SHIV?

Indeed, there is an Air Torpedo mod for SHIII by Rubini, but I think it exploited the same workaround as the ship torpedo mod (where torpedoes are spawned in game as the muzzle flash effect of an invisible dummy gun); I doubt anything related with their working and with their setup to apply to later SH games

Originally Posted by propbeanie

planes will spawn from the same group, until it runs out of planes apparently. [...] All of the planes seem to come in to "play" from the same generally direction. I would say only one AirBase responds, seemingly the closest.

Resuming:

• Once an unit is detected, there is a chance that response aircraft will spawn from the nearest enemy base within detected unit's range. If, considering max radii of the assigned aircraft, no air base is within range, no response aircraft will spawn.
  
• The chance that the selected airbase will actually spawn response aircraft depends on a number of factors like airbase veterancy level and day/night time as well as a number of related parameters in AirStrike.cfg (i.e: Default Air Strike Probability, Probability Increase factors, Atenuation Factor, Airbase Modifiers and Night Modifier). If someone could provide the formula actually used by the game for calculating this chance, that would be cool.
  
• The chance of a response "session" is calculated at fixed time intervals equal to Logic Steps Between Air Sessions (AirStrike.cfg) x 90 seconds.
  
• The maximum number of response aircraft spawned per session is variable, depending on the type of unit (ship, submarine, port, coastal defense, etc, Axis/Allied threat zones count as if they were a refular unit) whose detection has "triggered" the response. In SHIV, the typical response against a submarine is composed of a maximum of seven aircraft.
• The minimum number number of response aircraft spawned per session, depends on the number of aircraft with the appropriate range assigned to the selected airbase and "available" for the mission.
• Resuming, if we are detected and the nearest airbase has more than 7 aircraft assigned with enough range to perform a response raid, only 7 aircraft will spawn (i.e. the maximum hardcoded limit for ASW attacks), but if that base has only 4 aircracft available, then only those 4 will spawn.
  
• All the aircraft composing a response group will come from the general direction of the airbase they spawned from. If that airbase has more than one aircraft type assigned and available, the composition of the response group will be random. Aircraft won't get to the operation area all at once, but at short intervals and in small groups of 1, 2 or 3 aircraft.
  
• No matter if the aircraft composing a response raid are killed in action or they return to base after accomplishing their mission, they become "unavailable" for the next raids, yet, on they way back to base they are still "active"; they will report any enemy unit if they detect one (thus triggering a potential cascade of "resposne raids"), and they will attack it if they still have any weapon available.
  
• If the unit which had triggered the air response was not destroyed during the first raid, more raids will be considered by the game if the selected airbase has still some aircraft available. Indeed, that will involve waiting for the next dice roll.

Originally Posted by propbeanie

When we attempted to build "appropriate" bases for a given RL air group, we would usually end up with skies polluted with airplanes. We put the 38th BG B-25 AirGroup at Port Moresby, New Guinea, but 72 planes seemed a bit excessive, so we made it 16 planes. Yet, a person could not operate their submarine between New Guinea and the Admiralities north of the Bismarck Sea, and Guadalcanal south of the Solomon Sea, for all of the airplanes. The B-25s were constantly in the air, and the Japanese of course, had to have an air response to the B-25s... lol - mass hysteria! Plus, they sank everything in sight! Nothing left for the subs. So we cut the B-25 group back to 6 airplanes and shortened their operating radius quite a bit. It is much better now.

LOL

Did you try playing with AirStrike.cfg settings in order to reduce aircraft activity?

[gap]

I have just found a couple of books containing information on USAF Squadrons similar to the information we already had on RAF ones:

Air Force Combat Units of World War II


Combat Squadrons of the Air Force WWII

SHIV modders as well as SH5 ones might be interested in having a close look into those documents

[propbeanie]

Originally Posted by gap

Weren't air torpedoes a feature introduced with SHIV? Yes, but I can't remember if that was the initial release, or one of the "updates"... The problem of the torpedo not launching properly was there also though - or at least, for a period of time. Part of the problem was plane speed, as well as plane height when launching, which may have been "fixed" in v1.5. I do not recall where it was, but someone mentions working on fixing the backwards torpedo launch, and it may well have been a TDW thread instead of Rubini...

Indeed, there is an Air Torpedo mod for SHIII by Rubini, but I think it exploited the same workaround as the ship torpedo mod (where torpedoes are spawned in game as the muzzle flash effect of an invisible dummy gun); I doubt anything related with their working and with their setup to apply to later SH games Exactly on the "spawn" method. This does bring to mind though, the fix, and that was the actual spawning of the torpedo being initiated from the "splash" of the "fake" drop... The drop did not matter, since the "splash" is what spawned the actual torpredo. Then, it is just a matter of "aiming", of which there is none... lol - well, actually, there is, it's just that the aiming of the torpedo for the game AI utilizes its base platform, which is a gun, and a shell shot from a gun travels much faster than a torpedo...

Resuming:

• Once an unit is detected, there is a chance that response aircraft will spawn from the nearest enemy base within detected unit's range. If, considering max radii of the assigned aircraft, no air base is within range, no response aircraft will spawn. yes
  
• The chance that the selected airbase will actually spawn response aircraft depends on a number of factors like airbase veterancy level and day/night time as well as a number of related parameters in AirStrike.cfg (i.e: Default Air Strike Probability, Probability Increase factors, Atenuation Factor, Airbase Modifiers and Night Modifier). If someone could provide the formula actually used by the game for calculating this chance, that would be cool. There is a write-up on that - in the AirStrike.cfg itself?? Anyway, the game decides if it will respond (is usually does on the first run-through). It then decides which airplanes are available that can reach out at the detected range, and pulls randomly from that pool. The game will usually not use one plane type exclusively, but rather, spread the response, as can be seen in that video, where the first plane is a P-40, the next three are Wildcats, then a PBY (probably not seen in that particular video), etc. The other factors are then applied to the response equation.
  
• The chance of a response "session" is calculated at fixed time intervals equal to Logic Steps Between Air Sessions (AirStrike.cfg) x 90 seconds.
  
• The maximum number of response aircraft spawned per session is variable, depending on the type of unit (ship, submarine, port, coastal defense, etc, Axis/Allied threat zones count as if they were a refular unit) whose detection has "triggered" the response. In SHIV, the typical response against a submarine is composed of a maximum of seven aircraft. This is what it appears to be.
• The minimum number number of response aircraft spawned per session, depends on the number of aircraft with the appropriate range assigned to the selected airbase and "available" for the mission. Yes, but does include all types of planes, dependent upon range.
• Resuming, if we are detected and the nearest airbase has more than 7 aircraft assigned with enough range to perform a response raid, only 7 aircraft will spawn (i.e. the maximum hardcoded limit for ASW attacks), but if that base has only 4 aircracft available, then only those 4 will spawn. yes - total aircraft though, not just one type
  
• All the aircraft composing a response group will come from the general direction of the airbase they spawned from. If that airbase has more than one aircraft type assigned and available, the composition of the response group will be random. Aircraft won't get to the operation area all at once, but at short intervals and in small groups of 1, 2 or 3 aircraft. yes
  
• No matter if the aircraft composing a response raid are killed in action or they return to base after accomplishing their mission, they become "unavailable" for the next raids, yet, on they way back to base they are still "active"; they will report any enemy unit if they detect one (thus triggering a potential cascade of "resposne raids"), and they will attack it if they still have any weapon available. I am not certain of "availability". It seems that an AirBase seemingly generates more airplanes than it should, sometimes apparently generating its new response off of the airplanes that should still be enroute back to base, but are outside the player's "spawn range".
  
• If the unit which had triggered the air response was not destroyed during the first raid, more raids will be considered by the game if the selected airbase has still some aircraft available. Indeed, that will involve waiting for the next dice roll. yes

LOL

Did you try playing with AirStrike.cfg settings in order to reduce aircraft activity?

Yes, we "tweak" those every so often in attempts to get a more predictable response. Probably one of the main problems found in FotRSU, and this applies to Sea assets as well, is the disparate sources of a lot of the assets imported to the mod. Each author of a ship or plane has their own idea of how to do things, and they build them for their own version of the game, which that "environment" may well be significantly different from what they are currently being used in...

Originally Posted by gap

I have just found a couple of books containing information on USAF Squadrons similar to the information we already had on RAF ones:

Air Force Combat Units of World War II


Combat Squadrons of the Air Force WWII

SHIV modders as well as SH5 ones might be interested in having a close look into those documents

Very nice! Thank you.

[gap]

Originally Posted by propbeanie

The problem of the torpedo not launching properly was there also though - or at least, for a period of time. Part of the problem was plane speed, as well as plane height when launching, which may have been "fixed" in v1.5. I do not recall where it was, but someone mentions working on fixing the backwards torpedo launch, and it may well have been a TDW thread instead of Rubini...

I have made a quick search on this topic on subsim forums, and all the mentions I have found are about SHIV and SH5. I myself experienced the issue while importing in game a beautiful B-25 model by JU_88.

I am not 100% sure about that, but I think the fix resides in eqp node naming or in its positioning relative to the 3D model of the plane

Originally Posted by propbeanie

There is a write-up on that - in the AirStrike.cfg itself??

Those comments are not entirely clear to me, for one thing they don't say how the 'attenuation factor' is applied, and the part relative to 'coverage factor' calculation is a bit nebulous. I would have preferred a set of formulas, but at least those notes shed some light on the general meaning of the main settings. Most of them are multipliers, so if we double one we should half the Default Air Strike Probability (or viceversa), if we want to keep about the same air raid probability.

Originally Posted by propbeanie

I am not certain of "availability". It seems that an AirBase seemingly generates more airplanes than it should, sometimes apparently generating its new response off of the airplanes that should still be enroute back to base, but are outside the player's "spawn range".

I am sure I took this "availability" concept from one of your previous replies, but I probably misunderstood you. Thank tou for clarifying that!

Originally Posted by propbeanie

Yes, we "tweak" those every so often in attempts to get a more predictable response. Probably one of the main problems found in FotRSU, and this applies to Sea assets as well, is the disparate sources of a lot of the assets imported to the mod. Each author of a ship or plane has their own idea of how to do things, and they build them for their own version of the game, which that "environment" may well be significantly different from what they are currently being used in...

That's a general proble. In theory, before being merged into a mega-mod, any mod should be tuned to work well with the rest, but indeed that's not always an easy task.

Anyway, talking more specifically about air strikes, the only aircraft property with a direct effect on their probability, should be max radius and maybe max speed.

Originally Posted by propbeanie

You sir, are correct! The planes were not modeled to be a major player in the game, and the player was not supposed to want to look at the beautiful scenery and wonderfully modeled airplanes and ships of the game, but to rather stay inside his little steel coffin, water dripping from every surface, and grin and bear it looking through a periscope instead... lol - That way, you never see the airplane crash after it dive bombs you, or you never see that what took your boat to the bottom was the plane itself, and not its bomb, which had fallen harmlessly 30 meters away... lol - It is a balancing act to get a decent, appropriate airplane response. Plus, the planes do not "dog fight", so their only true natural enemy are surface ships and submarines... - unless, of course, they crash into each other, which they do from time to time... or crash into hills that are over 500m in height... "AI" need not apply for a job here. Only AD (Artificial Dum-dum) is accepted by the game.

You make it sound like a joke whereas the dumb aircraft could be described as modding drama.

Just one note: TDW developed a patch for making SH5 aircraft to "dogfight". Well, their air duels don't come even near to an actual dogfighting, but seen from a periscopic perspective...

[propbeanie]

Originally Posted by gap

I have made a quick search on this topic on subsim forums, and all the mentions I have found are about SHIV and SH5. I myself experienced the issue while importing in game a beautiful B-25 model by JU_88.

Some of the "discussions" are in unrelated threads. TDW was famous for divulging details on another of his mods, while discussing a newer one. But I thought someone in the SH3 threads, while discussing Rubini's torpedoes had done something to "fix" them before SH4 had "fixed" them... dad-blame worn-out RAM I've got in my noggin...

Originally Posted by gap

I am not 100% sure about that, but I think the fix resides in eqp node naming or in its positioning relative to the 3D model of the plane

wish I could remember...

Originally Posted by gap

Those comments are not entirely clear to me, for one thing they don't say how the 'attenuation factor' is applied, and the part relative to 'coverage factor' calculation is a bit nebulous. I would have preferred a set of formulas, but at least those notes shed some light on the general meaning of the main settings. Most of them are multipliers, so if we double one we should half the Default Air Strike Probability (or viceversa), if we want to keep about the same air raid probability.

Unfortunately, they don't say what type of multiplier, such as one seems to be as it is written, while the next seems to be a percentage, but I am never certain if I'm dealing with 1 x 0.0001, or a 1 x 0.01 when they say '0.01' for some of the settings. In other words, is it really a one hundredth, or is it one one hundredth of a percent, which is a "one ten thousandth"... lol

Originally Posted by gap

I am sure I took this "availability" concept from one of your previous replies, but I probably misunderstood you. Thank tou for clarifying that!

Again though, I am not certain where that "line" is that the game considers a platform as "available" to go back out once again, but I have seen it where a "fast" airplane, such as a Wildcat, will seem either like it's late to the response, or that maybe the game ran an early AirStrike probability, because I have already seen 4 of 4 Wildcats around my sub, and yet, just after they have dropped off of radar, there are a couple coming toward me again. No other way to explain, unless they did happen to come from a 2nd airbase, but followed the same routing as the first base... ?? The only way to find out, is to do a series of tests with only one airbase with only one air group with a very limited supply of airplanes, and see what happens when the enemy is detected...

Originally Posted by gap

That's a general proble. In theory, before being merged into a mega-mod, any mod should be tuned to work well with the rest, but indeed that's not always an easy task.

AI visual, and other sensors also come into play, and not just for the airplanes. The ships or shore installations that might report a sighting of the enemy, or a radar detection, or a huff-duff, etc., all come into play for the "balance". We have seen seemingly innocuous changes affect more than we would like, and oftentimes, it isn't noticed until a large group of people get their hands on the mod... "What happened to the DD? Are they deaf dumb & blind?", only to have the next round have a "What happened to the DD? I died within 2 minutes of shooting my torpedoes! They were on my like stink, and sank me on the first pass!"... lol

[QUOTE=gap;2714991]Anyway, talking more specifically about air strikes, the only aircraft property with a direct effect on their probability, should be max radius and maybe max speed.[quote]
yes, and I'm not certain the game considers the plane's speed. I would imagine the game considers the plane type also, especially if the detection is a long way off.

Originally Posted by gap

You make it sound like a joke whereas the dumb aircraft could be described as modding drama.

Just one note: TDW developed a patch for making SH5 aircraft to "dogfight". Well, their air duels don't come even near to an actual dogfighting, but seen from a periscopic perspective...

I do remember having fun watching that quite a few years ago... lol

[gap]

Originally Posted by propbeanie

Unfortunately, they don't say what type of multiplier, such as one seems to be as it is written, while the next seems to be a percentage, but I am never certain if I'm dealing with 1 x 0.0001, or a 1 x 0.01 when they say '0.01' for some of the settings. In other words, is it really a one hundredth, or is it one one hundredth of a percent, which is a "one ten thousandth"... lol

Hi Beanie, your doubts illustrate pretty well why I prefer formulas to textual notes. Airfield and night modifiers for sure are simple multipliers (1=100%).'Probability Increase' factors are not that clear though. They might be in percent (1=1%) and they might be actually added to the calculated probability. If that was true, supposing that the calculated probability of an airfield spawning planes is, le'ts say, 20% and a 'Probability Increase on Player Detection' of 40 applies, the final probability would be 20% x (1 + 0.4) = 28%

Originally Posted by propbeanie

The only way to find out, is to do a series of tests with only one airbase with only one air group with a very limited supply of airplanes, and see what happens when the enemy is detected...

I agree

Originally Posted by propbeanie

AI visual, and other sensors also come into play, and not just for the airplanes.

You are righty. In the worst possible scenario, if two enemy airbases were placed within each other's range, aircraft from one of the bases would trigger a response from the other base and viceversa. Unless aircraft "availability" is considered, I suppose that would trigger a endless loop of violence lol

That would be a nice testing situation btw.

Originally Posted by gap

I'm not certain the game considers the plane's speed. I would imagine the game considers the plane type also, especially if the detection is a long way off.

If memory serves, once TDW said that aircraft had to be within n minutes (travelling time) from an enemy unit for them to be considered for a "response" strike against that unit, "n" being a function of the loigical steps setting. Not 100% sure about that, but if that was true, faster aircraft would have an higher chance to spawn since the would cover the same distance in lesser time than slower ones...

[U-190]

Originally Posted by gap

As the title says, here are some long-standing doubts of mine that I never managed resolving:

1. Are aircraft (or any other AI units) able to use their non-visual sensors for spotting other AI units, or they are only effective against player's boat?
   
   
2. For airbase-spawned aircraft, do they need to be in the same base and/or airgroup to act coordinately (i.e. a group of bombers with their fighter escort)?
   
   
3. For airbase/carrier-spawned aircraft, is there any way to make some classes not to spawn at night or - conversely - to only spawn in night time?
   
   
4. For airbase/carrier-spawned aircraft, do they inherit their veterancy level from the unit they spawn from, or how else is their veterancy level determined?
   
   
5. For airbase/carrier-spawned aircraft with multiple armament loadouts, is there any way to predict which loadout will be actually used in different circumstances? Is that random, or how else the game is assigning them?
   
   
6. For airbases and carriers with multiple airgroups, do their airgroups need to be sorted sequentially according to their start date? Can an airgroup start before the previous airgroup is ended? In other words, can an airgroup have a start date which is earlier than previous group's end date?
   
   
7. Has anyone tried adding one or more airgroup(s) to any non-carrier sea unit? That would be useful for simulating CAM/MAC ships and scout planes aboard cruisers and battleships.

Any hint by some knowledgeable subsim mate would be much appreciated

The behavior of aircraft and ships is random. The AI behavior is recorded in one common configuration file. I was trying to solve the problem you are talking about in IRAI v0.0.45. For more advanced behavior, each individual AI should have its own behavior configuration file. A similar expansion is observed only in later versions of modern games.
I think that I will still make attempts to work on this problem.