Stochasty

Was it you? The design I remember was some crazy asymmetrical design using a pair of RT-10s plus a liquid engine. As for your 2x Mun design - hmmm, that's close to what I was thinking as well. My Munlight design was essentially half that: 2 jets, 1 LV-N, ~9 tons of biproellant fuel and 1 ton of Jet fuel (since I was only planning on orbitting once).

If docking/undocking is allowed, this should be pretty easy. It doesn't take much to build a nuclear powered craft with that kind of delta-v; add it as payload to some big jet-powered SSTO, undock, land on the Mun, return, dock, land at KSC, repeat. Without docking, it's a lot harder, since you'll need to carry your jet fuel to the Mun. That craft I used for the Munlight challenge way back when had about 3500 m/s delta-v in orbit (she could make a return trip from Duna) and was only 16 tons, so scaling that up might get you there. Nao had some crazy craft that managed an SSTO Mun return without using jets or ions, so adding jets to that design might also work.

Could we try to keep the bickering out of the thread, please? Take it to pm if you must argue.

Yep. T'was in a post which got eaten by the kraken, but I still have the craft file lying around somewhere. She was 22 tons at takeoff, if I remember correctly. In other news, after six months of effort, I finally have a working stock Eve Plane. She's 258 tons at takeoff, Kerbin SSTO capable (so she launches herself, sans staging, and refuels in orbit - note that I did not refuel her for the K-Prize run), and with staging she has ~12km/s delta-v. I've gotten her into orbit from 2.5km altitude on Eve with ~500m/s delta-v left in the tank and a somewhat inefficient flight plan. I doubt she can make it from sea level, but anywhere above 2km is probably doable. Taking off from Kerbin: Orbital insertion: 75km parking orbit: On approach to KSC: Landed: Reentry was quite tricky; she was never designed to fly empty, to she's actually just slightly negatively stable and wants to reenter tail first. Once speed drops below ~200m/s in the atmosphere, it's relatively easy to get her turned around, and SAS can hold her pointed in the right direction as long as you don't make too sudden a heading change. And now to show her doing what she was designed to do - returning from Eve:

I've done return trips from the Mun, Minmus, Duna, Ike, Gilly (easy), and Laythe. Never done Pol or Bop, but the Laythe craft could have managed those easily. I want to try for Eeloo and Moho; I think I've got a craft with enough delta-V, but I'll need to use gravity assists around Jool or Eve to make it and I haven't taken the time to really work out the trajectory yet.

As long as you aren't using the Nouveau drivers, you're probably okay. With the Nouveau drivers (on CentOS 6.4), the game was so slow it would crash prior to loading more often than not; with the proprietary drivers it runs fine.

That would probably take hours of tedium. That said, I've serously considered it, which should indicate how difficult sea level return really is.

I have a working stock version now, also - well; partially working. She's ten stages, asparagus staged, total of around 12km/s delta-v, and will make it to orbit in stock from any landing site above 4km altitude. The core is essentially the same as my FAR version, but with added drop tanks and a few more engines for thrust. Unfortunately, she's not capable of launching herself from Kerbin without staging, and landing is a pain to say the least, so I haven't yet succeeded in flying the mission for real. I'll post a video of it as soon as I manage it. As an interesting note, the ascent from sea level up to 4km altitude costs over 4km/s delta-v, so I'd need to more than double her size to pull off a sea-level return. Regarding mass and delta-v requirements for spaceplanes versus rockets: much testing on Eve makes it clear - there is /no/ advantage to wings. You are better off with a rocket. Yes, you can get off the ground with a sub-1 TWR, letting you carry twice as much fuel per stage as an optimal rocket, but you are much less fuel efficient during ascent out of the thick atmosphere since you end up flying a shallow trajectory. There may be a slight efficiency gain during the gravity turn, since you start picking up horizontal velocity much sooner than a standard rocket (AKA, you are trading worse efficiency as regards drag losses for better efficiency as regards gravity losses), but it is vastly outweighed by efficiency losses during the early climb. On Kerbin, this is barely noticable (I've mentioned before that I've put spaceplanes into orbit for roughly the same delta-v requirements as rockets); however, on Eve, you spend so much more time in the thick atmosphere, and drag losses are so much higher compared with gravity losses, that the loss in efficiency is extremely noticable. That said, spaceplanes are fun; hence why I've been working on my Eve plane. Just, don't expect building a plane to make the mission easier.

Mmmm... for this challenge, ion engines /will/ set speed records. You can get a lot more delta-v out of an ion engine than anything else, and nobody is going to by flying brachistocrones out past Duna orbit so thrust doesn't matter. As far as jet engines, I agree with you; they don't make enough difference to be worth banning.

For a vertical ascent, if the density of air were constant, flying at terminal velocity is the most fuel efficient method. For air density exponentially decreasing with altitude (such as for KSP), it also happens to be the case that terminal velocity is the most efficient for vertical ascent (although in this case you will need a TWR somewhat higher than 2 to maintain the correct speed). For non-vertical ascent, or for air density profiles that differ from exponential, the problem is much more difficult to solve, and it is not obvious that terminal velocity is the correct speed for maximum efficiency; however, typically, during the non-vertical portion of your ascent you should already be clear of the thick part of the atmosphere and so drag losses should be less important. Thus, "always terminal velocity" is a good rule of thumb to follow.

Lower ISP just means more stages. 17 of them, to be exact; total delta-v for that rocket was in the neighborhood of 14km/s, weighing in at 508 tons of RT-10 goodness. Hmmm... 14km/s, and I know that I was a bit inefficient with the staging since I built her by trial and error without diagnostics. I could probably manage a redesign and squeeze out another few km/s; that'd put her in the right range for an Eve return trip.

You should retry this challenge using FAR. Realistic drag makes a huge difference.

This is nearly the least efficient descent profile that you can manage. About the only way to make it less efficient is to decide to hover at 1m altitude above the Mun rather than actually landing. It is also the hard way to land, because your velocity is always vertical, so if you wait too late for your burn you're out of luck. The reverse gravity turn method is easier, more efficient, and much more forgiving to mistakes (if you wait too late, you just wind up burning a bit more fuel rather than testing the viscosity of your spacecraft).

How about a Kerbolar impact? Edit: Hrmm; rereading your challenge post, I'm confused. Is the challenge to build a craft using solid rockets and see how much you can do with it, or is the challenge to see what you can do with the specific craft file you posted?

Cubic octagonal struts are your friend. You may have to abandon some of your sense of aesthetics (or work really hard to hide the intakes) but you can put intakes just about anywhere if you are creative. They do generally need to be pointing forward, though. Yes; less drag.

There are some counter-intuitive rules of thumb for building effective spaceplanes, but once you learn them getting a plane into orbit is rather easy. I'll try to lay them out: 1) When considering the jet engines, thrust is your enemy, not your friend. The reason for this is that the max speed obtainable by your jet engines is almost strictly determined by their operational ceiling, which is determined by the ratio of intakes to engines. Adding more engines doesn't help, since you have to just add that many more intakes to feed them. You always want to shoot for the fewest amount of engines necessary to do the job. (The hardest stage of a spaceplanes ascent is the climb from 10km to 15km altitude. Add engines until you can manage that climb, then stop adding engines and start adding intakes.) 2) Corollary to 1), carry only the jet fuel that you need. One of the easiest ways for your design to fail is for you to accidentally add too much jet fuel. Any fuel unspent during your climb is dead weight in orbit and will limit your operational range. 3) In contrast to jet fuel, you want to carry as much rocket fuel as you can lift. Every extra unit increases your operational range. 4) You should not need much in the way of rocket engines to finish off your orbit. A proper ascent trajectory has you level out near the operational ceiling of your turbojets so you have time to accelerate as much as possible before you ignite the rockets. Done right, the ascent on jets should put you at near-orbital velocity, and you should not need much thrust at all to circularize. It is entirely possibly to design a spaceplane capable of orbit using only turbojets and ion engines, although I typically use nukes because I can't stand 15 minute long burns. 5) Use only the ram intakes. Everything else is added drag for paltry intake air. The radials might seem like they add a lot, but once you are at altitude they are far less effective than the ram intakes. 6) If you still can't make it work, try building a spaceplane using aerospikes first. It's quite possible to build an SSTO spaceplane without using jets at all (although you won't have much in the way of payload); one aerospike, about six of the long 1m tanks, and enough wing surface to get you airborne should suffice for a one-man pod. Once you know what it takes to make orbit without jet engines, you can start adding jets to the design. You'll be amazed at how much easier it is to make orbit, and how much extra fuel you have left.

It's closer to the RL performances of various engines, although in the real world the engines have much better TWR. Unfortunately, because of scaling issues in KSP, it breaks the game. It takes around 4.5km/s delta-v to achieve LKO in KSP (with FAR, it drops further to ~3.5km/s); it takes closer to 10km/s delta-v to achieve LEO in real life. Delta-v costs for maneuvering around the solar system are also higher by about the same factor of 2.5 for RL versus KSP. Thus, a 10km/s delta-v rocket in KSP is the rough equivalent of a 25km/s rocket in RL, meaning that in KSP with modular fuel tanks it's quite easy to build high-payload SSTOs (without even utilizing air-breathing engines). Adding FAR to the mix, I managed to build a ship theoretically capable of single-stage-to-Eve-landing-and-return (total delta-v expenditures are on the order of 12km/s for such a trip). Thus, sadly, while this mod adds a lot of interesting realism, once you understand the system it makes the game too easy. I would recommend rescaling the fuel tank mass ratios to put them back closer into line with those in the stock game, even if fuel tanks in the stock game are unrealistically heavy.

But this is the crux of the issue, isn't it? Please tell me where, in the climate science literature, I can find studies of these sorts for the GCM models which are used to make the climate predictions which appear in the IPCC reports and other politically relevant literature? I've looked; I haven't found them. If you happen to know of such, by all means point me in that direction! Because, as far as I can tell, this type of assessment has never been made, which means that the primary point of those essays to which I linked is right on the money. This really isn't controversial stuff I'm asking for here. If you've worked on quantum modelling, as you've claimed, then I'm sure you've been forced to do similar assessments of your own models to pass peer review (or used models for which similar assessments had been made). Failure to make these types of assessments means that you aren't yet doing legitimate science. Why is asking about this type of stuff as regards climate science so taboo? Why does it provoke such an argument from those on your side of the debate when I point it out? Why don't you, rather, just post relevant links to the literature showing that my questions have been addressed or are, for whatever reason, off base?

This is a good point, but one that I vehemently disagree with. However, I don't know how to convince you or anyone else of the reasons behind this disagreement. I could start by pointing out that certain well known names in climate science - Dr. James Hansen, for instance - are also climate activists with well-known biases. I could point out that certain fields of climate science - paleoclimate proxy reconstructions, for instance - are plagued by bad statistical methods, and that the most well-known result of one of the biggest names in the field - Dr. Michael Mann - has been thoroughly discredited and were he in any other field he would have been forced to retract his most well-known paper. I can point out that there is evidence (granted, evidence based on stolen emails) that a small group of influential climate scientists have intentionally manipulated the peer review system. I can point out that the singularly most influential political climate change organization - the IPCC - is controlled by people (Rajendra Pachauri being the best example) who have a vested, monetary interest in seeing carbon control measures put into place. I can point out that grants for funding climate change research are invariably given to those who favor the anthropocentric climate change side of the argument (try to get funding to disprove anthropocentric climate change if you don't believe me on this one). However, none of those (except for the bit about manipulation of peer review, which is the one that sticks in my craw the most) is proof that there is bad climate science going on, or that (as you put it) climate scientists aren't doing legitimate science. You are probably right that, in general, they aren't manipulating results fraudulently (although I do happen to know of one case that should probably be considered out-right fraud, committed by one of the actors I mentioned above). However, you are wrong to suggest that they aren't manipulating results to improve their funding prospects. I highly suggest that you go back and reread the part of the second essay where he talks about the "fishbowl ecology" of the modern scientific community. I can tell you, from first hand experience, that he is absolutely correct on this. That is how modern science funding works, both in America and abroad. Science, except in specific cases with direct application to the real world, is not funded by corporations or private individuals. It is funded by governments, at their politically motivated whims, and we scientists have strikingly little say in what actually gets funded. Right now, in the US and in Europe, the pro-anthropogenic climate change side has won the political debate and is in control of the distribution of funds; in China, because the Chinese have a vested interest in coal, the opposite is the case. This is why there is no funding for climate change skeptics in the US, but the Chinese Academy of Science has begun to fund them. How this relates to manipulation of data: if you happen to find a result that you know contradicts the going orthodoxy and is likely to negatively impact your future funding chances, whatever you wind up doing with that data you do not publish it. So far as I can tell, this is the reason why the full Law Dome ice core data set was not published for more than a decade after it was taken (only portions of it were published), despite the fact that it offers the highest resolution data yet collected. This is the biggest problem with the intersection between politics and science; the modern science establishment only really works well when the people who control the purse strings do not have a vested interest in the results. Hell, this suppression of negative results happens even when there's no money at all riding on the outcome. Search the literature look for papers which report "no statistically significant result was found." You'll find that these papers, when they occur, occur only in response to previously published papers which had found a statistically significant result for a similar investigation. It's not sexy to publish "no result" papers, even when "no result" is the most common outcome, and if you happen to publish a "no result" paper that steps on the toes of someone who cared you can be in trouble down the line. With a few exceptions, I don't really fault those involved. I happen to believe the flaw is systemic, and has nothing to do with the individual scientists. I also happen to believe that this systemic flaw is worse for the overall health of the science industry than some nefarious conspiracy could ever be. Academic politics sucks balls. It is no secret that, once you have managed to obtain a PhD, actual ability as a scientist is not the primary contributor to a successful career. What matters more than anything else is who you know and who knows you; in any given field there are far fewer research positions available than there are qualified applicants, and the pool of potential bosses is extremely limited. Also, everyone already established in the field already knows everyone else, so prospective scientists can't really afford to piss off any of them or they risk a short end to their careers. This, as expected, makes for a disturbing amount of groupthink and nepotism. (Hell, it's no real secret that my own research job is entirely the product of nepotism. I once calculated that there are fewer than ten jobs in the entire world that make use of my particular specialty; the only reason that I happen to have landed one of them was that, as a grad. student, I got to know and work with several influential people in the field, one of whom is now my boss.) What does all of this add up to? Nothing conclusive. But, if you want the reason why I feel so jaded, there it is. I can't really convince you that feeling jaded about the state of science - especially climate science - is the right way to feel, but I can assure you that the easiest way to gain that feeling is to become a scientist yourself.

The main point he's arguing against - averaging results of multiple models to produce a "projection" - is used by the IPCC, and had you done any more than reading the essay looking for ways to scoff at it you would know that. Nevertheless, that's not the interesting part of his post. The interesting part is his discussion of the history of modelling the electronic structure of the Carbon atom and how, even today, we are forced to resort to semi-empirical models because we simply don't know how to solve the equations of the real physical system (despite the fact that all of the physics - at least at this scale - is known). These models have known, unphysical, fudge factors designed to make the output of the model conform with data. Molecular dynamics simulations - even for simulations of large proteins - are vastly simpler than trying to simulate the full climate of the Earth, and yet even they are fraught with peril. The equations of the dynamics of inter-atom electron binding may not be the same as the Navier-Stokes equations, but the difficulties of modelling complex, non-linear systems are the same. Ignore his arguments if you wish, but it is not his ignorance you are displaying when you do so.

My plan: pick some guy I can't stand (and I mean truly detest) and land him on Venus.

Ah. I see. I guess I should have used the term "numerical general relativity" since that's what's actually being simulated, but everyone in the field just uses "numerical relativity."

Errr... I mean numerical relativity? Numerical simulation/evolution of the Einstein field equations for gravity (with hydrodynamical simulation thrown in for good measure for those classes of problems which involve matter). I don't understand what possible other meaning the term numerical relativity could have, so I admit to being confused by your confusion.

Pak, in case you're interested, I ran into two very interesting essays recently written by a computational physicist at Duke discussing some of the issues you raised earlier in the thread about the nature of computer modelling. Links here and here.

Cool movie, but BH accretion disks aren't really the state of the art in numerical relativity. Why simulate dust falling into a black hole when you can simulate two black holes falling into each other, or neutron stars falling into black holes, or white dwarves colliding and possibly forming black holes? Harder to find good movies of these, though, since most of the visualizations are created with research (rather than publicity) purposes in mind.