chaos_forge

This is absolutely a good thing. I would much rather have it delayed so it can be absolutely amazing than have it be rushed and turn out badly.

I think an extremely light (similar to Snacks!) life support system would be ideal. Enough to stop me from launching a ship for a multi-year mission that has a single capsule as its only crew compartment, but not enough to become a game about managing logistics.

Because this is a highly science-based game. It is a game, yes, but it is also realistic. These properties are not in conflict. In fact, KSP is fun in part because it is realistic. KSP is fun because science is fun, and because realism is fun. If people didn't care about the real world when playing KSP, there wouldn't be dozens of recreations of historical rockets made by players, or players motivated to become real-life rocket scientists because of the game, or a huge fanbase of the game among scientists, etc etc etc. Because part of the fun of KSP is its realism, I want the devs to be educated on the science. There are compromises made with realism, yes, and I wouldn't have it any other way. But the thing is, the devs can't properly navigate those compromises between realism and game design if they don't know the science. You have to learn the rules before you can break them, or else you won't be able to foresee and deal with the consequences of your actions.

I find "it's a game" to be a pretty weak excuse, honestly. Yes, no game is 100% realistic. But at the same time, no game is 100% detached from reality either (otherwise you would just have stuff randomly happening for no reason). Realism isn't a yes/no binary, but a spectrum, or a sliding scale. So the question isn't "is it realistic?," but "where do you draw the line?" And as has been stated earlier in this thread, many of us find "speculative engineering, but no speculative science" to be a reasonable place to draw the line. Furthermore, I would claim this is where the KSP and KSP2 devs themselves are attempting to draw the line. The devs are comfortable abstracting away the finnicky engineering concerns that the average player would not want to deal with, such as tank ullage, engine ignition, or throttle depth, but the engines are still based on well-understood scientific principles and concepts. Additionally, the speculative engines shown in the trailer (with the obvious exception of metallic hydrogen, of course), such as MPD thrusters, Orion drives, or ICF drives are all based on well-established scientific principles. These are all engines that we know for sure can be built in real life. They may not operate as efficiently as they do in the game, or they may have engineering challenges that make them not worth actually using, but they can be built, and would operate according to scientific principles that we already understand today. I also think it's worth pointing out that the KSP2 dev who's been doing interviews said in at least one of the interviews something along the lines of "oh, we read a paper saying metallic hydrogen exists, so we feel confident including it in the game." This, combined with other instances of devs failing to understand the science (such as not knowing that binary systems can't be modeled accurately by the SOI approximation), leads me to believe that their decision to include metallic hydrogen engines does not stem from a conscious choice to handwave the science in favor of gameplay, but rather from a misunderstanding of the science itself. And that is, ultimately, the most concerning aspect. I honestly would be much more okay with it if the devs understood the science and were choosing to ignore it than the current situation, in which it seems the devs do not actually understand the science.

We know it exists, but we don't know if it's metastable at low pressures, which is the important part for rocketry. So what I meant is we don't know for sure that metastable metallic hydrogen exists. And given how vanishingly rare it is for states of matter that have a significantly higher binding energy than the ground state to be metastable, the default assumption should be that it is not. Which means until we have conclusive proof that metallic hydrogen is metastable, we should assume it is not.

Another option for an even more advanced alternative to metallic hydrogen is antimatter rockets. At least we know for sure that antimatter exists. So in my opinion, the purpose of life support as a mechanic is twofold: 1) Adding engineering complexity to ships. You can't just slap a crew capsule on top of your 20-year-trip ship, you have to add enough crew modules to support the crew for that amount of time. 2) Adding additional risk/adding urgency to rescue missions. If the Mun lander you designed only has supplies for a few days, that means if you get stranded, you can't just leave the kerbals there and say "I'll go rescue them eventually," you have to rescue them NOW. I do agree that not having self-sufficient colonies would be a management nightmare, but I think life support makes sense as a game mechanic for ships.

How much will mods be able to interface with/modify UI elements? For example, Principia has a lot of trouble interfacing with the stock maneuver node system. We've heard a lot about modding parts and planets, but will UI elements also be moddable?

In the stream, Scott also mentions that the devs said they had no plans of implementing n-body, and that it seemed like they hadn't really thought about how the binary planet would work, so: It seems y'all are right, and I have to eat my words. I'll have to remember to have less faith in game devs in the future. They're in touch with Scott Manley so hopefully he'll be able to stop them from trying to implement some weird hack-job SOI system, but I fear their solution will be to just scrap their plans for binary planets instead of implementing n-body. I still think n-body would be good to include in the stock game, because the main problems players could have with it could be easily done away with by implementing automatic/abstracted-away station-keeping (which is pretty easy to do), and by having a good UI, which is one of the few things I think professional game devs can do much better than modders. Anyways, with regards to metallic hydrogen: This is also where I'm at. As long as it's know to be theoretically possible within our current knowledge of physics, I'm okay with abstracting away the engineering challenges. What I'm less okay with is abstracting away the science.

Here's a twitter thread from someone who got a peek at the game: https://mobile.twitter.com/badnewsbaron/status/1166829617514041344 Not an interview or video, but probably still relevant to the thread

None of those necessarily have to result from n-body. For example, Children of a Dead Earth, another extremely realistic space sim (in many ways even more realistic than KSP), has automated station-keeping that doesn't consume any vessel delta-v. You just check the "station-keeping on" mode, and boom your ship is on a fixed conic section orbit. Such a system would be extremely easy to add to KSP2, and I would be extremely surprised if it wasn't included in KSP2. The SOI model can't handle binary systems, so I don't know how else you think the devs are gonna make Rusk and Rask work.

As far as I know, metallic hydrogen engines wouldn't involve magnetic nozzles. The way they (theoretically) operate is very similar to chemical engines, since the energy does come from a chemical process after all.

The problem is the patched conics approximation, as a theoretical model, simply can't handle binary systems. This isn't a coding problem, it's a physics problem. There's no way to model it that wouldn't result in some highly unphysical behavior.

I mean, I don't think anybody's crying themselves to sleep over it. We'd just rather it not be in the stock game. And it seems to me if the ISP of metallic hydrogen is around the same as LVNs, un-nerfing LVNs is probably a better solution. Then they could fill that same niche of "same thrust as methalox but moderately higher ISP" but with a mild mass penalty so they aren't strictly better in every single situation (just most). And if you want fancier rocket engines, add some liquid-core or nuclear lightbulb style NTRs. At least those are theorically within the bounds of modern-day materials science.

I see you're also a fan of The Expanse

Most of the systems that the devs have added to ksp have been extremely simple and configurable: ISRU uses a single resource type and planets are scanned instantly, the comms network mechanics are extremely simple and can be turned off in difficulty settings, reentry heat can be turned off in difficulty settings, etc. It seems their general approach is to do the simplest/most approachable implementation, and let people who want more complicated mechanics use mods. I'm fairly confident the implementation of life support added by the devs will be a lot more similar to Snacks! than to TAC, and I would be very surprised if anyone found it overcomplicated.