Lisias

i.e.: it's predetermined to be undetermined!

Because they are physics engines. I was talking about physics processors. Having a hardware processor with multiple pipelines crunching numbers in parallel is not the same as having a physics engine that can calculate multiple parts of the same craft at the same time. The inability of calculating physics in parallel on many parts of the same craft is a limitation of KSP (probably due a limitation on how Unity does that). Physics processors knows nothing about any of that.

A rover cabin with the note: "To whom it may concern, take care of this Rover. She saved my life. Watney."

Interesting! AtmosphereAutoPilot overwrites the control-surface symmetry (I have some problems with it sometimes, as I need the symmetry to manually control asymmetric crafts I build now and then), and (IIRC) have better fine tune adjustments. So I don't think this is a part problem, but a physics one. Mainly because you said this happens without an atmosphere too - I think your control surfaces are so massive that they are messing with the CoG when they moves (and not only with the CoL). Can you provide a minimal craft with this problem? This is something interesting to debug!

The nice thing on KSP is that it's all text files. Any Text Editor allows you so see (and change) these files. This is what I got: UrlConfig { parentUrl = Bluedog_DB/Parts/APAS/CXA_APAS_P PART { name = bluedog_CXA_APAS_P module = Part <cut> MODULE { name = TweakScale type = BluedogStack defaultScale = 0.9375 type = free } <cut> } UrlConfig { parentUrl = Bluedog_DB/Parts/APAS/CXA_APAS_A_L04F PART { name = bluedog_CXA_APAS_A_L04F module = Part <cut> MODULE { name = TweakScale type = BluedogStack defaultScale = 0.9375 type = free } <cut> } You see that two "type' thingies? This is the nasty stuff. Bluedog's Author meant to use the BluedogStack, but something else shove a new "type" without deleting the other one, and now things are confused. Until the moment, KSP uses the latest value when a duplicated entry happens (unless you want to read it as an array - but it's not the case here). Problem is - nobody knows from where came that "free", and so nobody knows when that "free" can goes away (by adding, deleting or updating an Add''On , changing the order in which things happens and sometimes not applying the offending patch). And then your crafts goes kaput. This time I didn't got lucky, that "free" thing is pretty common. So I will need to check it the hard way: patch by patch, after downloading everything. At least I already ruled out one or two Add'Ons, as I know they don't use "free" on their patches. I'll get back to you as soon as possible.

Got it! [LOG 18:46:30.095] [TweakScale] INFO: TweakScale::WriteDryCost: Concluded : 0 checks failed ; 0 parts with issues overruled ; 2 Show Stoppers found; 21 Sanity Check failed; I checked that 21 "insane" parts. By some reason, someone tried to shove TweakScake on the kerbalEVA* parts and on a flag (5 parts) due the Issue #30. There're a lot of Add'Ons mangling the kerbalEVA* parts, this will take some time to sort out - but I will do it on the weekend. Well, these are not scalable anyway. I don't think Valentina would do well as a "50 foot Kerbalette" The other ones are parts from Making History (MODULEPARTVARIANT with mass) and Firespitter (FSBuyoancy) that are not supported (yet) by TweakScale. You can use them too, but they don't scale (yet). The other two are a new! (or I having memory issues again? ) [LOG 18:46:29.056] [TweakScale] WARNING: **FATAL** Found a showstopper problem on bluedog.CXA.APAS.A.L04F (CADS 0.9375m Docking Port (Active)). [LOG 18:46:29.056] [TweakScale] ERROR: **FATAL** Part bluedog.CXA.APAS.A.L04F (CADS 0.9375m Docking Port (Active)) has a fatal problem due having duplicated properties - see issue [#34]( https://github.com/net-lisias-ksp/TweakScale/issues/34 ). [LOG 18:46:29.056] [TweakScale] WARNING: **FATAL** Found a showstopper problem on bluedog.CXA.APAS.P (CADS 0.9375m Docking System (Passive)). [LOG 18:46:29.057] [TweakScale] ERROR: **FATAL** Part bluedog.CXA.APAS.P (CADS 0.9375m Docking System (Passive)) has a fatal problem due having duplicated properties - see issue [#34]( https://github.com/net-lisias-ksp/TweakScale/issues/34 ). Let's check one of them: [LOG 18:37:31.864] Config(@PART[bluedog_CXA_APAS_P]:HAS[!MODULE[ModuleConnectedLivingSpace]]:NEEDS[ConnectedLivingSpace]) Bluedog_DB/Compatibility/ConnectedLivingSpace/CLSBluedogDB/@PART[bluedog_CXA_APAS_P]:HAS[!MODULE[ModuleConnectedLivingSpace]]:NEEDS[ConnectedLivingSpace] [LOG 18:37:31.872] Config(@PART[bluedog_CXA_APAS_P]:NEEDS[TweakScale]) Bluedog_DB/Compatibility/Tweakscale/tweakscale_APAS/@PART[bluedog_CXA_APAS_P]:NEEDS[TweakScale] [LOG 18:37:31.880] Config(PART) Bluedog_DB/Parts/APAS/CXA_APAS_P/bluedog_CXA_APAS_P [LOG 2019-08-07 18:33:54.994] Deleting root node in file Bluedog_DB/Compatibility/ConnectedLivingSpace/CLSBluedogDB node: @PART[bluedog_CXA_APAS_P]:HAS[!MODULE[ModuleConnectedLivingSpace]]:NEEDS[ConnectedLivingSpace] as it can't satisfy its NEEDS [LOG 2019-08-07 18:34:02.289] Applying update Bluedog_DB/Compatibility/Tweakscale/tweakscale_APAS/@PART[bluedog_CXA_APAS_P]:NEEDS[TweakScale] to Bluedog_DB/Parts/APAS/CXA_APAS_P.cfg/PART Well, I got the part source on this file, Bluedog-Design-Bureau/Gamedata/Bluedog_DB/Compatibility/Tweakscale/tweakscale_APAS.cfg: @PART[bluedog_CXA_APAS_P]:NEEDS[TweakScale] // CADS 0.9375m Docking System (Passive) { %MODULE[TweakScale] { type = BluedogStack defaultScale = 0.9375 } } Since this is a bluedog part, I would add ":FOR[Bluedog_DB]" on the PART thingy, but this is not the source of the problem! (but perhaps it would help to prevent it?) It's something else mangling with this part, together this patch. Problem is…. NO ONE ELSE did it (explicitly) on the log! Can you please post the "ModuleMaager.ConfigCache"? By looking on the part there, I can have a hint about who is the another guy patching it!

Yep. Got it: [LOG 18:17:15.639] [TweakScale] INFO: TweakScale::WriteDryCost: Concluded : 0 checks failed ; 0 parts with issues overruled ; 1 Show Stoppers found; 21 Sanity Check failed; [LOG 18:17:15.604] [TweakScale] WARNING: **FATAL** Found a showstopper problem on batteryBankMini (Z-200 Rechargeable Battery Bank). [LOG 18:17:15.604] [TweakScale] ERROR: **FATAL** Part batteryBankMini (Z-200 Rechargeable Battery Bank) has a fatal problem due having duplicated properties - see issue [#34]( https://github.com/net-lisias-ksp/TweakScale/issues/34 ). On the bright side, it's only one serious issue. The batteryBankMini is borking due a patch on ROs, see this post about how to fix. There's a lot of potential problems on the RO's patch, and unfortunately I lack the time to proper support TweakScale and also propose fixes for RO's patches, so I kindly ask you to reach RO's maintainers about the issue, pinpointing this post as a source of information. [yup. copy&paste ! ] On a not that bright side, you have about 12 "non RO - *** Cryogenic Fuel Tank" parts that had TweakScale withdrawn due some incompatibility between Third Parties Add'Ons. Not a TweakScale issue, but on that situation trying to scale that parts leads to disasters. I can give you a proper report so you can reach the Add'On maintainers for a solution if you can post the ModuleManager.ConfigCache file (on your GameData) by crossing the log file to the cache and checking exactly the Add'Ons involved on the mess. They are not a problem right now, you just can't scale them. —— addendum — — TL;DR : you need to edit RO patches to make sure the look like this: @PART<yadda yadda yadda> @MODULE[TweakScale]:NEEDS[TweakScale] { %type = RealismOverhaulStackSolid } — — — — — — — — — The 9 remaining parts (basically engine plates) are parts that TweskScale doesn't know (yet) how to scale (but you can use them alright). This will be fixed on the 2.4.4 series.

Follows a brute-force change log This is what I got from the github (inverted order than jenkins). 3d85ddc v4.0.3 e2027f1 Merge pull request #148 from sarbian/FixNewlinesInCache f2fba86 Ensure tabs and newlines don't break cache 96201a7 Fix ModifyNode handling of escaped characters 0feef88 Fix DeepCopy handling of escaped values 8e5c29a Clean up test a bit 3e4a9f9 Fix TestConfigNode's handling of escaped chars ee5995f Remove redundant parent class 9311cfb Improve TestConfigNodeTest 9795e36 make StreamLoggerTest.TestLog work on \n platforms eb3db60 unnecessary using 3d8594d Set FileType based on common types a27cb3d Convert UrlBuilder cfg test to theory e1bed94 improve file sha generator 69ab085 Alphabetize .csproj files ae0d7a7 Put extension on cache URLs 5e5314e Use custom code to lookup file by url 28bd925 Another unnecessary using 5f7071d Mark disposables as sealed 3650d69 Use better assertion 5711994 Unnecessary using 1c93735 Merge pull request #146 from sarbian/RedoLogging a8bcad2 ModLogger -> PrefixLogger 0a2c842 Initialize timestamp with log message 07afe29 Don't put date on every log message fc74f1f Initialize LogMessage from old LogMessage d4d0fb7 Redo logging interface cfaac6a Fix StreamLogger ignoring messages 7271725 Add name value to applying patch messages a0c1dfc add test for ProtoUrlConfig 6e6a465 use new KeyValueCache class as regex cache 918b0a2 add thread-safe KeyValueCache Other than release dates (and RO being yet in the 4.0.2 version), they appear to be identical: https://ksp.sarbian.com/jenkins/job/ModuleManager-RO/changes https://ksp.sarbian.com/jenkins/job/ModuleManager/changes (see the change #149)

There's something absolutely weird on your installment! TweakScale is being loaded TWICE! [LOG 00:07:39.922] AssemblyLoader: Loading assembly at D:\SteamLibrary\steamapps\common\Kerbal Space Program\GameData\GameData\TweakScale\Plugins\Scale.dll [LOG 00:07:39.923] AssemblyLoader: KSPAssembly 'Scale' V2.4.0 ... [LOG 00:07:40.073] AssemblyLoader: Loading assembly at D:\SteamLibrary\steamapps\common\Kerbal Space Program\GameData\TweakScale\Plugins\Scale.dll [LOG 00:07:40.098] AssemblyLoader: KSPAssembly 'Scale' V2.4.0 Delete "D:\SteamLibrary\steamapps\common\Kerbal Space Program\GameData\GameData\" from your rig. Are you using CKAN? In a way or another, that "[LOG 00:15:34.199] [TweakScale] INFO: TweakScale::WriteDryCost: Concluded : 0 checks failed ; 0 parts with issues overruled ; 575 Show Stoppers found; 7 Sanity Check failed;" message fits the diagnose. With two sets of TweakScale patches on your installment, TweakScale patches will be applied twice.

And you being born is consequence of your parents had choosen to engage in non-forum-compliant activities when your mother were fertile. Even if they didn't knew your mom were fertile at that night. Of course, not all events are consequences of a choice - that meteor that extinguished the dinosaurs are a perfect example of that. But that didn't means that the dinosaurs were destined to be extinct by a meteor. We are predestined to die. But this doesn't means that we were predestined to born! You need to breath to be alive. Have you a choice to stop breathing? You need to eat to be alive. Have you a choice about what to eat? you don't have choices about everything. But this doesn't means that you never have a choice at all!

If we were, how would we know it? Being predestined means we have no choice, but by knowing the predestination in advance we would have the choice to avoid it. So your questions would be unanswerable by definition. by logical inference, if you expect an answer, the only possible would be "no".

You can use them without worries. TweakScale is telling you that these parts cannot be scaled by lack of proper support, but they are alright to be used. Warnings and advises are just informative. These are yellow or white, and means that things are not perfect (yet), but are not bad either. The nasty things are informed in red, and these gets in the way preventing you to proceed until you dismiss them. About that 9 parts, version 2.4.4 will have them supported.

Or you can apply a OVERRULE patch to get rid of the "extra" Fuel Switch on the parts that have more than one. You can have more than one Fuel Switch installed on the GameData, things goes through the tubes when people install more than one Fuel Switch on a part. The easiest way of avoid the mess is to do no install more than one Fuel Switch on your KSP - so patches just can't shove more than one on a part even if they try (and some will). But easiest is not always best. For your case, the OVERRULE patch is the solution.

Nope. Only the events due the consequences of the choices we do.

Nops. These things are happening since 1.4.1 at least (I got my first serious crashes and savegames corruptions on 1.4.3, last year). You should use the latest TweakScale on any KSP >= 1.4.1, as all of them suffer from the problem TS detects. You may have some reluctrance to get older Add'On versions fixed for older KSP versions, but that is the main reason I cooked the OVERRULES patches. The worst happens, we handcraft a patch for you to keep your game running on a safe, besides far from ideal, way.

Mötorhead - acoustic!

I don't see, really, a real necessity of getting that mods "killed". As you said, 1.3 was the closest they ever get into a stable, good featured and low bug count as they ever got. There're players on 1.3 yet. And players using pre DLC KSP versions. And players that don't wanna the bells and whistles of the DLCs. They are still there, they didn't ceased to exist due the DLCs. I don't see this as a mutually exclusive situation. With that, I agree 100%. I don't mind cut down DLCs. I even buy them to get i touch with what's happening (the new robotics aesthetics looks awesome with some designs of mine, besides IR/N still being my workhorse on robotics - and there's KAL). But some technical debits are starting to get way more than annoying. There're 90 MILLION active players/month on Steam alone - says the last survey (2019), being that 96.28% of Windows players. Guess what? There're 3.72% of Macs and Linux players there. Or about 3.402.000 potential customers, looking for a stable game on their platforms. There're only a few games for them, while that 96.28% of Windows players have their attention and interest disputed by 30.000 titles. As I said before, I think these guys are suffering from Survivorship Bias. There's life out of the Windows gaming, and for someone that it's fighting to get 1 or 2K more players by month (approximately what Steam Charts are saying me since Breaking Ground gone live), trying to get the attention of 3 millions dedicated but disdained users appears to be a good bet to me. Mainly because you can get them without risking losing any of the Windows ones.

It's pointless to play a game at first place. We do play games for fun. We do speculate about the game for more fun.

Nope. It means that part has no influence on the physics significance, but it's still handled by the physics engine: "humm… this part should not be applied, I will jump it". There's no free lunch! Sometimes, it costs more to do not calculate an exception than just calculate it together the basket of numbers and just don't use it later. It's an idiosyncrasy related at how the physics processors work (highly parallelized pipelines).

The best place to start looking for trouble is the KSP.log. Read this, publish the logs and then let's see what we find. Most problems are hard and cumbersome to diagnose, but simple to fix - let's see how lucky you are.

Excuse sir, but he must be worried. A lot. We are not normal people. Wee see poofed Kerbals… "And Now for Something Completely Different" — Monty Python style…. It's interesting how our brain shuts down gradually as we get tired. I was working nearly non-stop for 12 hours on a very intricate middleware (yeah, I had to wrote my own service-bus to get that extra bit of performance. ), thought it would be a good idea to cool down around here, got into a discussion and ended up mixing the mean density of the whole damned rocky planet with its atmospheric density. Made a whole argument on it without realizing the complete absurdity of the numbers. Well… Enough internet for today. Time to get some sleep, as it appears.

I already know that. Thanks. Because nobody wrote the patches. It's an interesting feature for the 2.4.4.x series.

UUUGH!!! 5240 is the mean density of the planet!!! #facepalm I copied the wrong value and applied it mechanically without really thinking on the thing. Ok. Let's try again. Water has a density of 997 kg/m³. Our atmosphere is 1.2kg/m³ . Venus' atmosphere has a density of 67 kg/m³. ~14.8 times less denser than water. But yet 55.8 times denser than our atmosphere. Not feasible at all for that monster or anything remotely near that, but Venus would withhold floating structures way heavier than our atmosphere. (thinking about what I'm writing is a nice habit I should exercise more regularly.)

Venus atmospheric pressure is up to 90 times the Earth's. Or 90 * 1013.25 millibars = 91.192,5 millibars. The Pressure on Earth's oceans at 100m deep are 9.81 bar, or ~9.810,00 millibars. 10 times less than Venus' atmospheric pressure. But wait, it's the density that matters, right? Water has a density of 997 kg/m³. Our atmosphere is 1.2kg/m³ . Venus' atmosphere has a density of 5240 kg/m³. ~5.2 times denser than water. Or ~4.320,33 more denser than our atmosphere. Do you know what? Pretty feasible, if we manage to shove all that steel on that planet in one piece. (Copyeur und Pasteur is a dangerous technique on scientific essays)

No, sir. Science Fiction - no fantasy.