ÆþelrædUnræd

Not a guitar, but I did build this six-stringed beauty ❤️: A Germanic lyre, built in a kind-of-Sutton-Hoo-style-but-not-quite. As my nickname pretty much gives away, I'm fascinated with the ancient Germanics (which includes but is not at all limited to the Anglo-Saxons) so I thought it'd be fun to have a lyre 🙂

Not sure what a "Page 2 veteran" is? Google doesn't work here since you just get lots of second pages of whatever about veterans😄 Anyhow, May 29 is not even close to May 5. At least, three full weeks is not something I call close. Is there any further meaning to the date, or is it simply something like "what the heck, why not celebrate it right now"? Also, just to be pro-active in case anyone thinks otherwise, I'm just curious and in no way want to detract from the importance of days like this to remember those who died for your country. I do think it's important to honour those dead. And as a Dutchman, I feel obliged to honour at least those of them who died during WW2 - lots of Americans died while driving the Germans from my country. If not for those brave soldiers, I'd be living in either Greater Germany or the Soviet Union. With world politics as they are, I think there's hardly been a time it was more relevant to stop for a moment and think about what exactly our freedom means, and how much was sacrificed for it. It's very little compared to what those who fell gave, but I do want to give my sincere and heartfelt gratitude to the Allied dead.

Just wondering; it's clear to me from the posts above that this "memorial day" is a day on which the US honours its military dead. But why this date? It seems just so... arbitrary. It's not on the completion date of any of the US wars that sprang to my mind (although the US Civil War is close), and a quick scan of the Wikipedia page also didn't reveal any answers. The equivalent day in the Netherlands would be "dodenherdenking" (lit. "remembrance of the dead"), which is aptly placed on May 4 to pretty much coincide with the German capitulation in the Netherlands on May 5, 1945 while leaving May 5 itself for more festive celebrations ("dodenherdenking" is meant for a solemn remembrance). This doesn't seem to be the case for "memorial day"? Oh, and of course I wish everyone a good memorial day 🙂 ("good" would be fine wouldn't it, since it doesn't imply any specific state of mind?)

Actually, for me land based would be the much preferred option. Taking off and landing on carriers is fun, sure, but flying hundreds of kilometres in a straight line over the ocean with nothing to see but water, water and more water just sounds so insanely boring.... That said, I think Pearl Harbor is a poor choice for a "land battle" as it's just an hour or two of air combat, 99% of it Japanese.

"Better" is subjective, but at least it's a clearer name for everyone but the few history/aviation nerds who happen to know what an "IL2 Sturmovik" is. Nothing against Pink Ladies!

Well, I'm not going to argue with you there. For the first installment it won't be much of a problem anyhow; they can just move the carriers closer to midway for those who prefer it that way. And they'll likely have some time compression algorithm anyhow. As for later installments; time compression clearly favours the single player crowd and scaled maps the multiplayers. Let's hope they can find a solution that caters to both.

And I'd contend that time compression merely changes your perception of the world around you, rather than change the world around you itself, and as such is much preferable to a scaled world. Use it if you want it or leave it; the only aspect it changes is the speed of the simulation (as advertised I would say) instead of having all kinds of nasty side effects in navigation and fuel use etc. I'm with you on the 8 hour Rabaul to Henderson flight that almost nobody is going to fly - that's why I advocate time compression. I don't think anyone really disagrees that few people would enjoy flights like that. I'm curious though why you think a full-size slot map is "just not going to happen". Because of these long flights or because of a perceived technological barrier or development time? How much time would the average multiplayer gamer be willing to fly? Half an hour? Thats 1/16th of the 8 hour flight. Let's round it up to a tenth, or 48 minutes. Guadalcanal is only 40km wide where Henderson field is located; it's runway would cover a quarter of the island! Also the distance from Rabaul to Guadalcanal would change to a hundred kilometres or so - that's within visual range on a clear day. So much for navigation skills. Not to mention engine management. As a matter of curiosity since you advocate scaling down the Solomons, how do you suggest to scale things down this dramatically while still keeping the illusion of scale intact? (Note that any less dramatical scaling down would still make for longer flights than the average Joe is willing to make.) I agree that time compression would be... tricky... to do in multiplayer. I'm sure a solution can be found, although I don't have a perfect one at the ready except airstarts or perhaps placing a few (non-historical) carriers nearby.

Full scale is much more than just the time it takes to fly from one end of the map to the other: There's the visual aspect; I want islands (and the islands between them) to be as large as they are in reality to give a good overview of the scope of the battles back then. There's navigation. Shorter distances make navigation much less of a challenge. If your target is enemy ships, finding them is even more challenging as they'll have moved quite a bit in the time it takes to reach them. Fuel consumption is another thing. Or would you suggest to increase fuel consumption as well to match the smaller size of the maps?

Correct; performance issues are rarely caused by map size, if a proper terrain loading algorithm is implemented. Case in point: MSFS. Unparallelled detail across the whole globe. Where map size does matter is development time and disk space requirements. One solution I can think of is where they'd in fact have a single "globular" map with just ocean, that then gets filled in with islands (or land in general) based on which modules you bought. I respectfully disagree here. When done properly, the size of the map doesn't really cost much more computing power and if you'd be cutting the distance between islands, not much of development time either since sea is pretty much free in that regard. This leaves lengthy flights as the only real problem, and here I think good time compression is a much better answer. I think proper distances are important to keep the general perception of size intact.

A moderator actually looking forward to dealing with people. You're either much too kind for the job, or a sadist 😛 I like your star wars theme by the way

Not really, AFAIK the only "confirmation" exists in that in their Career page, they have a job listing requiring familiarity with UE5 blueprints and C++. Even if they indeed use UE5, that doesn't mean they'll also use nanites. Also note that they'd probably heavily modify the engine code to make UE5 more suitable for a flight sim. Agreed here too, for screenshots or videos you need at least some detail. But yes, it should come second to more important things 🙂

Not really, that's why LODs (Levels Of Detail) are created. If you move far from an object, it'll load a different model with much less detail in order to save resources. Unreal's super-hyped Nanite stuff basically takes this to the extreme. It's insanely detailed, but performs quite well. Less detail will save a lot of development time however. For a flight sim, you're mostly not going to be very close to a fuel truck, so why spend much time to model every screw in 3D?

Actually, shaders like this create their own geometry 🙂 They're 3D just fine, but instead of using wave geometry that was created in the normal update loop on the CPU, they themselves create this geometry in the vertex shader on the GPU. Which makes sense, since GPUs are pretty good at performing the type of calculations needed. Getting this data back into the CPU memory is a hassle (and most probably in the wrong format anyhow for what you'd want to do with it in the update loop) so that's not usually done; therefore for all purposes other than display it's indeed a 2D surface. The display part is purely 3D though and happens before the fragment shader which actually paints the surface, therefore the fragment shader "sees" a surface with thousands of vertices where there were only a couple before 🙂 There are ways to create 3D effects while keeping a 2D surface even in the shader, such as normal, bump or parallax mapping. That's not what you see in the screenshot I posted above though. Volume is quite hard to do, really; it's rare to see actual volume shaders where they aren't absolutely needed (e.g. volumetric clouds). In the case of water, you can get pretty good results by just taking a "pseudo volume": you know the height of the vertex under the water surface, and you know the world space angle from which you look at the vertex. You can then approximate the length of the under water part of the view ray by simple trigonometry, and adjust the colour accordingly. Transparency itself is rather simple, but water is also refractive which is a major problem, actually. Unless you're using raytracing (which for almost all other purposes is extremely inefficient), you're likely much better off doing an approximation in screen space. Anyhow, my money's on them doing a healthy mix of "real" 3D and "shader" 3D. Real 3D created in the update loop for larger waves that ships (and aircraft, should you be so unlucky) should react to, and shader effects for smaller waves and ripples.

Nope, it isn't: Very clearly 3D. This may depend on graphics settings, and it's probably done in the shader so purely a graphical effect, but 3D nonetheless.

I don't think that's quite correct. IL2 uses 3D water. It may be that the vertices are only generated in the shader, and I think it does depend on graphics settings, but it's 3D water nonetheless. Anyhow, I don't think IL2 uses 3D water *physics* and that's where the real challenge is. Making a surface look like water is one thing, making objects actually interact with it in a realistic manner is quite another.

Actually, horses before a cart is the one vehicle I'd like to see most in IL2 and FC; in both wars it was the most common vehicle for many of the belligerents. 😛

No matter who supervised/greenlit the "detachment vehicle", I'm a huge fan of it. I honestly think its main defect is that it currently only exists for two countries, both of which are on the same side and both of which are on the Western Front. Should this sim ever move towards ground battles (Guadalcanal, New Guinea, Marianas, ....) I think anything similar would be an important bonus. That said, the only time we're going to see carriers is either when you're taking off from or landing on them, or when bombing them. In none of those cases should there be more than a couple of people on deck.

Yes, it may. Or the AI might not comprehend just exactly what you're trying to accomplish, and come up with a valid solution to the wrong problem. Or the result may be severely unoptimised. However, if you only ask it to generate relatively common functions, there's a pretty high chance it'll do a good job, and this will still tremendously speed up coding.

That's not at all what I have noticed. Based on all the threads and comments I've seen, it's the IL2 standalone installer that's notorious for suddenly ceasing to work, or to keep downloading without an end. I haven't seen much problems of the kind on the Steam platform. Anyhow, as long as there's a Steam version as well, I'm happy. I'll gladly buy any expansions from the site itself, like in IL2, but I like to have the base game within Steam. But if it's possible and doesn't cost too much time/money to have a standalone installer as well for those who prefer it that way, then why not? 🙂

Absolutely! Naming and grouping things is a must-have. Which leads me to another point you should add to your list: the ability to group objects as well as define inputs and outputs for messages to/from an object inside the group. Groups are a great way to keep things clear, but linking to an object inside a group (or the other way around) is a major hassle that keeps me from using groups more often. A bit stupid they didn't implement it really, since the basic functionality already exists and is used inside the mission templates for campaigns (the various Helper MCUs). I have no interest in further prolonging this useless discussion, but I do ask you to please refrain from spreading blatant lies. None of us have ever said we want to keep other people from using the editor. In fact I've stated the opposite multiple times already! Either you didn't read my posts well, or you don't comprehend them. If you don't comprehend them, I'll gladly explain them to you in simpler English, but stop putting words in my mouth I never said! Less importantly, neither of us ever used Arma or its editor as any argument; it was you who brought the game up in the first place.

If I sound a bit touchy about it, it's not because I don't respect dissenting opinions, but because those dissenting opinions address none of the arguments I bring forward to support my viewpoint. It feels like I have to keep repeating myself - if you want to know my viewpoint about a "better organized UI" you can read it in many of my posts above. If you want to have an in-depth and mature discussion in which you actually address the issues Gambit and I bring up like budget and inherent complexity, as well as support your own viewpoint with actual arguments yourself, I'm happy to continue. If instead you prefer to debate without addressing any of my arguments but rather throw around accusations of narrow viewpoints and touchiness, then I think this discussion has outlived its usefulness.

Yes, I've tried many mission editors in many different games across various genres. I haven't used Arma3 or its editor, so I won't make any comments on it as it isn't relevant anyhow. It's a wildly different game/genre, and most importantly, based on SteamDB estimates, it's got more than 20 times the sales figure. And that right there is one of the main reasons it's unlikely we'll see "all in one" or "brand new improved design": budget. Creating a complex editor can take as much resources as developing a small game. My thinking may be narrow, but at least it's realistic. That said, I do think some solution will be found eventually 🙂

Phrased like this, I can agree with it a lot more 🙂 Of course, deliberately making an editor difficult to use to keep people out would be wrong. But I don't think anyone's advocating that. What I'm saying is that the converse, deliberately dumbing down a mission editor to keep everyone in, is equally wrong. I hope whatever mission editor they come up with will be streamlined as much as possible, but I'm sceptical about how well a powerful editor can be combined with one that's easy to use. That's why I'm thinking along the lines of two separate editors. The thought of being able to write simple custom scripts in-editor and well integrated with its graphical level is making me drool. 🤤 There, see?

I'm not trying to gatekeep mission building. My point is that by limiting the options and possibilities of the Mission Editor in order to make it accessible, you're setting an upper limit on the quality of all future missions. Many of the popular campaigns in IL2 GB, both official and free, would not have been possible with the editor of 1946. The possibilities with the IL2 GB editor are *almost* endless, if you know what you're doing and are willing to spend some time (even though there are some critical flaws and "missing features" as IckyAtlas rightly points out). And please, cut the dramatics. The IL2 mission editor has a steep learning curve, absolutely. And its UI can get clunky with target/object links all over the place. I'll also gladly admit that it sometimes unnecessarily complicates relatively simple actions and relatedly, as IckyAtlas also mentioned already, it could do with a couple of pre-defined logic blocks (IF-ELSE, RANDOM, etc.). But its basic event-based operation should be immediately obvious to anyone with a bit of programming experience. Whatever way you slice it, creating missions *is* a form of programming/scripting since you're pre-determining which actions an AI should take under which circumstances. The fact that you need some slight knowledge of programming is therefore not at all an unreasonable requirement. Representing IL2s editor as a "deliberately badly designed piece of software that only the developers understand" is hence a blatant misrepresentation of the truth. Similarly, I can get where you're coming from 🙂 Indeed, a bad mission writer won't create a good campaign on the best mission editor in the world, while a good mission writer will be able to chunk out something fun even if using some pretty spartan tools. I can agree there. But only a good mission writer using a powerful tool will be able to create a true marvel 😉 Anyhow, it seems the clause in my previous post wasn't as clear as it could have been. I'll re-iterate it more explicitly: Ideally, we get a mission editor that's at least as powerful as IL2's while at the same time easy to use. Although I personally think a separate "full" mission editor à la IL2 GB and an in-game "light" mission editor à la IL2 1946 may be a better option as it caters to both user bases. In practice I doubt this is realistically achievable. If a choice needs to be made between the two, I think choosing the easier "IL2 1946" editor will hurt the quality of future produced content as it effectively puts a ceiling on the complexity of a mission.

The technology is there; there's just two factors that I'm not 100% sure could be solved at this point in time (I've got experience with running, training and even designing neural nets): Neural nets (I hate the "deep learning" buzzword) are hard to train; they'd need human opponent to fly against the AI for hours and hours on end, training many different situations. Just to give a ballpark estimate: to train the average neural net reasonably well, you'd need at least some 40,000 separate examples. Given that it takes a while for an action to give any result (e.g. the AI makes an Immelmann -> Gets shot down 1min later -> the decision 1min ago was apparently not a good one), it'll take ages to learn. Kinda related: training combat situations as described above will result in good AI, not necessarily realistic AI. Unrealistic behaviour can be penalized in training, but you first have to realise it even exists. Harder than one might think 😉 Neural nets are relatively hardware-intensive. To do one iteration, you need to do many millions or even billions of calculations. Granted, they're simple calculations and modern hardware (especially GPUs) are pretty good at it, but you're already taxing your PC with graphics, physics, etc. Now, there are some possible solutions. Once you've trained an AI to an acceptable level, you can actually train it against itself, requiring a human opponent only to keep its behaviour within realistic bounds. Or you could pit it against "traditional" AI, but then you'd need to program that first of course. For the second point, you don't need to run an iteration (make a decision) every single frame. Running it once every few seconds is good enough. Similarly, it doesn't really matter if it takes a second to compute when your GPU is already doing other stuff. I mean, it takes a real person some time to make decisions too, and if we make one we don't suddenly decide something entirely different half a second later. Another obvious solution is to make the neural net relatively small and lightweight (which will also tremendously speed up its training!!!) but this may have implications for how complex its decisions can be. For pilot animations, I don't think Neural Nets would be the best option as there's relatively little animations you'd need to create. I think the IL2 pilot has an idle animation, a handwave, look left, look right, wiping off sweat, bailout and sitting on ground. I could do that in half an afternoon if I'd need to (although at an inferior quality - I'm a poor animator I admit). On the other hand, training a Neural Net to do it for me will take months, if not years. For big studios like Ubisoft, the equation may be different. Big games tend to have hundreds or even thousands of animations, including some very complex ones (e.g. facial expressions). Also they need to be of a high quality/resolution since you tend to get up close to characters as opposed to a pilot you'll only see from a few metres distance at the closest.