Van Disaster

So ah, on the upside, it's within 500m of the target and nothing broke! which given I forgot RCS completely was not bad. On the other hand I can't open the habitat so nobody can get in it and it won't produce food. Next time, RCS. Now I have to ship something in a hurry to get it upright...

The nearest real life references for spaceplanes would not be normal airliners; try Concorde/Tu-144/SR-71 or any other simple-wing large high-flying heavy aircraft, if there are any. I'd rule out the XB-70 because it used compression lift, and the B-1/Backfire/Blackjack because they have variable geometry. The big issue with high wing loading - or to flip it around, insufficient wing area - is when you're up at 25km trying to hit mach 5 to transit to rockets; not enough wing means you'll be at some huge AoA just when drag is building up quickly.

You have a wing loading that's approaching 1t/m^2 there ( perhaps 0.8ish if you include the horizontal surfaces which might not always be contributing positive lift ), that's a tad high - the Shuttle had what was called "very high" wing loading at 585kg/m^2, and that only flew downwards. If you want the area of any surface piece, stick it on a fuselage & look at the FAR static derivative tab, it will have the total wing area. Mach tuck won't start until the wing airflow reaches mach 1; the wing cross-section FAR uses is I believe a reference supersonic shape so I don't think there should be too much oddness in the near transonic area, certainly not at 0.8. Model planes have a really limited flight regime in general and I suspect some rather crazy power/mass ratio ( and also rather crazy strength/mass ratio ) - certainly use it for rough guidelines but I suspect using that doc as absolute rules is not going to give great results.

That's because -force-d3d11-no-singlethreaded doesn't do anything at all. If I run just that flag the head of my log file looks like this: OS: Windows 7 Service Pack 1 (6.1.7601) 64bit CPU: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz (12) RAM: 12227 GPU: AMD Radeon R9 200 Series (3051MB) SM: 30 (Direct3D 9.0c [aticfx32.dll 8.17.10.1333]) <<-- NOTE THAT. If it's actually in DX11 mode that line would be this: SM: 50 (Direct3D 11.0 [level 11.0]) In the Unity docs it's listed as an option for Windows Store apps which don't have commandline arguments, so I guess it's specific to those builds. There's memory leaks near planet surfaces, changing scene and most definitely in the SPH/VAB. I can run for hours in space and crash out in 10 mins trying to build a craft if I'm in and out of the buildings a lot. I have GCMonitor up all the time, I can watch memory use tick up every update while I'm in the VAB...

I'm going to substitute "detailed" for "complex", I think. Currently it's not *quite* detailed enough; jet engines are a reasonable example, you can just stuff an engine on what would actually be the tailpipe of the entire jet, drop some intakes randomly and off you go. Smaller planes are generally designed around the engine and large ones have to give some thought to how the engines are mounted, it'd be nice if there was at least a little reflection of that. I have some vague design of a modular system that involves seperate intake/core/output devices which between them can make a jet engine - that sort of thing would fit the tech tree quite well, in fact, you can approximate all sorts of jet engines with surprisingly few different components. Doing the same for rocket engines wouldn't be a big step either. Science is the same - planting flags & putting a finger in the ground to take a sample is not really entertaining, especially when you've made the effort to get to somewhere. Having experiments that actually *need* a base because they're complicated/multistage and time consuming would drive so much gameplay it'd be another game on top, almost; you'd need ISRU or at least a good logistics chain, ground structures and connections, the works. If you had to pay more attention to the make-up of the planet involved and change your solution to match - perhaps including having to send robot scout probes - rather than just plopping the same science base down everywhere that'd be even better. I don't want it to be Orbiter though - if I wanted something like that then I'd just play Orbiter.

This managed to light itself really nicely to give some spooky desolation - the unexpected location of the craft ( which did actually get there itself ) adds something else, too. Alternatively it looks like it was faked up on a movie set ( ahem ). Maybe a Gerry Andersen set... Must reinstall some camera tools.

Put some stuff around Mun: I'll land the habitat stack & the lander tomorrow I think. Also started work on an obstacle course station for training a friend up in IVA docking & orbital rendesvous:

Nop. Think it was just the Firespitter DLL that needed updating, but that's common to a bunch of things.

I got to the point of having a script that'd pick a package of mods to link into GameData, process it and then start the game, but I fiddle with my install too much to rely on that. It's handy for things like shared games or running DMP though. The only problem with my approach is it'll also end up with shared PluginData dirs but *usually* that's beneficial. Only having to manage one set of mods is handy given how much I fiddle with those too...

~15x more economic, I thought? even as they are the stock ones will need rebalancing for new aero. FAR and B9 have a MM config to cut the stock engines in half, more or less, and they're still good for their size. The parts really need to be bigger, especially the Rapier - currently they're the end of the engine with no design considerations necessary for the rest of the engine. Jet engines are big! Rapiers aren't jet engines but the cooler part which the stock one doesn't even have would make them as big as an axial flow turbine anyway. What would fit KSP quite well is something of a modular tech tree compatible system of intakes/cores/outputs, which I have some thoughts on but haven't finalized yet ( I think I'll post when I have ).

FAR is mandatory if you're serious. Either of the procedural wing mods makes a ridiculous improvement - B9's can take fuel but is a little less mature than DYJs. After that the popular parts mods seem to be B9 ( the proc wings are a standalone part ) and Nertea's Mk4 - not to be confused with TT's Mk4 - Retrofuture and possibly Firespitter, but not sure how compatible the parts are with 0.90. Taverius' pizza pack has a few bits and pieces which fill some gaps too. You don't need all the parts from all the mods, especially if you're using procedural wings. There's also AJE which is an attempt at realistic engines - I don't know what it's like currently, it had some odd interactions with other mod installs when I last tried it, but that was at least one major version ago.

Might want to be a bit wary of adding something so close to the pad - people love to make wide launchers! looking good so far though. I don't have time or much motivation to do texture work at the moment, but I might suggest - based on your other point about Kerbinside - that making the Kerbinside textures a little more grimy & shaded would be good texturing practice. If it's possible anyway, I've not looked how they've mapped those.

For the cargo bay you can just use a bunch of box colliders for the walls, which unfortunately doesn't directly help for a round bucket although I suspect that'd probably be the way to go anyway - would be nice if there was an option to generate multiple box colliders instead of a mesh collider. Learning from mistakes I've made, concave meshes appear to fall through terrain ( which is not going to be a convex collider! ).

I get them too, sometimes - usually after using the offset tool. In my case saving & reloading the craft seems to cure it.

From looking at drawings it looks like a few degrees of anhedral, if anything - or the wing underside is nominally flat and the wing is so thick that the upper surface has a good amount, anyway. In flight I guess they'd have bent up to be neutral-ish. Jumping on the lack-of-vertical-stabiliser bandwagon, the real one's fin runs all the way up to the front area of the wings. Whatever the layout it needs to convert sideslip into a turning moment into the sideslip, so just fiddle with the vertical stabs with that in mind.

No, I have an escape shuttle that only carries a little jet fuel, RCS is plenty to get it to re-enter.

Just going to hang on for Unity 5 I think - I have enough exposure to unix as it is. I removed ATM completely & saved about 400MB - TextureReplacer dumps about 1GB of textures out of main memory on load, presumably once they're offloaded to VRAM; ATM wasn't actually doing anything at all ( I wasn't running any size reduction modes ). Been trying to fix Minmus but it doesn't appear to be a simple issue - I've actually managed to get it visible about 30km from surface but only at a narrow range of camera angles. I'm thinking that maybe just saving the surface info off & creating an entirely new planet might be the answer...

Differences: * No analysis tools * No supersonic effects - this includes no supersonic levels of drag, which makes a considerable difference aerobraking or re-entering. Aside from that it's just a little CoL shift and as you speed up flight surfaces start behaving more like boards than shaped wings, which means you can fly at absurd AoAs if you can keep the craft balanced. * Not sure right now because I don't track NEAR, but I think it's using an older version of FAR; Ferram mentioned he wasn't updating the NEAR aero model a while ago * No failures, but they're optional under FAR anyway. So really the difference is no supersonic effects, no way of checking your plane out before flying it, and possibly an older version of FAR as a flight model ( and perhaps it doesn't have the engine nerf MM patches). No supersonic drag would probably make getting a spaceplane to orbit easier, actually - but coming back, not so much.

Truth. It's actually harder than FAR with the aero failures turned off... Most info about FAR can be applied to NEAR, aside from anything relating to supersonic aeronautics which is a fairly rare topic, tbh - so hunting for FAR discussions will probably answer any questions. To strengthen Wanderfound's points: forget anything you knew about KSP atmospheric flight whether it's aircraft or rockets - just pretend you don't know anything at all other than how to attach a part to another part. Now just look at a few real life planes & try and make something similar ( or even rockets ) and you should be ok.

BTW do you think you could throwing in similar functionality to http://forum.kerbalspaceprogram.com/threads/101632-0-25-Part-Icon-Fixer-Tweaks-%28updated-12-2%29 is within scope? would save extra dll overhead at least, does seem reasonably within this one's area.

Plasmeus: I think the best way to solve that would be to make sure there's as little oscillation as possible in 1x time, to be honest. You could try using an alternate control damper like Pilot Assistant also. Silentdeth: I wish...

I have my 0.90 main install, and then outside the entire folder tree I have what is currently called 0.90_Modstore, which is where I install mods. After that I just symlink mods into GameData ( use mklink on windows NT-based OS ), which means all I have to do to swap entire modsets around is to list mod folders in a text file somewhere & run a simple windows script to swap entire installs without actually copying files anywhere at all. Also handy for checking stuff out in a bare bones base install ( which itself is mostly just a new folder & lots of symbolic links too ). If you load a game with a mod missing, just alt-F4 before it's had a chance to do anything & it should be ok. Actually I've loaded & saved games with ships not loaded because of missing parts, exited properly, put the missing parts back in the install & had the ship reappear, but I don't know exactly what the limits of that are. I run Nereid's S.A.V.E to take snapshots but wow does that eat disk space fast...

Exactly the same in SW England, high & low cloud with occasional breaks. The seagulls went noisily crazy though...

Well here's a new one - this is all it's giving me, but it's blocking all functionality unsurprisingly. [EXC 04:00:03.682] KeyNotFoundException: The given key was not present in the dictionary. System.Collections.Generic.Dictionary`2[System.String,System.String].get_Item (System.String key) FilterExtensions.Utility.PartType.checkFolder (.AvailablePart part, System.String[] values) FilterExtensions.Utility.PartType.checkFolder (.AvailablePart part, System.String value) FilterExtensions.ConfigNodes.Check.checkPart (.AvailablePart part) FilterExtensions.ConfigNodes.Filter.checkFilter (.AvailablePart part) FilterExtensions.ConfigNodes.customSubCategory.checkFilters (.AvailablePart part) FilterExtensions.ConfigNodes.customSubCategory.<initialise>b__0 (.AvailablePart p) System.Linq.Enumerable+<CreateWhereIterator>c__Iterator1D`1[AvailablePart].MoveNext () System.Collections.Generic.List`1[AvailablePart].AddEnumerable (IEnumerable`1 enumerable) System.Collections.Generic.List`1[AvailablePart]..ctor (IEnumerable`1 collection) System.Linq.Enumerable.ToList[AvailablePart] (IEnumerable`1 source) EditorPartListFilterList`1[AvailablePart].GetFilteredList (System.Collections.Generic.List`1 list) EditorPartList.RefreshPartList () EditorPartList.Refresh () EditorPartList.Refresh (State state) PartCategorizer+.MoveNext ()

Except there's other considerations like mass ratios of tanks and fuel densities to consider, which is what the other mods fix while they're making engines which perform accurately. You can't just pick one thing. You can make everything perform relative correctly if you pick a reference point in the stock KSP universe, but until your universe is an absolute match for real life then nothing else in it is going to perform absolutely correctly. Engines don't even work right in the current version, how are you going to make them "feel like the real thing"?