[1.4.1] Kerbal Construction Time 1.4.0.69 (2018-03-24) - Unrapid Planned Assembly

[CobaltWolf]

I'd actually like to see simulations have a large 'up front' cost and then have each individual simulation have a lower runtime cost. This is analogous to buying the hardware for the computer simulations and then paying maintenance and power to run them.

You could even have it scale with runtime, range, etc. Cheap hardware may only support simulations of ~30s and only within the kerbin SOI, with various upgrades to unlock more runtime and range or to reduce running costs.

This actually sounds really cool. I dig it. Pay to unlock the simulation parameters, then a smaller cost for the individual simulations.

[hjrrockies]

I'd actually like to see simulations have a large 'up front' cost and then have each individual simulation have a lower runtime cost. This is analogous to buying the hardware for the computer simulations and then paying maintenance and power to run them.

You could even have it scale with runtime, range, etc. Cheap hardware may only support simulations of ~30s and only within the kerbin SOI, with various upgrades to unlock more runtime and range or to reduce running costs.

This sounds really good.

[forsaken1111]

You could even have simulations take a bit of time to be designed. IRL when they need to simulate a new craft I imagine they have to do more than hit a big red 'simulate' button. Engineers have to design the craft in a computer program and create the simulated mission parameters.

What would you guys think about making simulations take some time upfront as well as a bit of funds cost for an engineer to design.

Say you hit 'simulate' for a vessel, pick a short duration and Kerbin SOI as the area. A simulation is added to the design queue. A simple one may take a few hours and cost some small amount of funds up front. In the background a simulation data is written with 2 components, one being "Kerbin SOI" and one being specific to the craft you designed. You run the simulation and decide to make some changes to your craft, so you click simulate again after making those changes. You again choose a short duration and Kerbin SOI. Since you already have the Kerbin SOI piece designed, the new simulation takes less time to develop based only on your updated vessel design. If you then went back to the first vessel design and did another simulation, it would take no additional time because you already have it designed. It would just cost funds to run.

Hopefully that makes some sense.

Edit1: Another idea, can make it even more interesting by requiring some science data from the SOI you plan to simulate before you can make the simulation. Send out a probe for a munar flyby to unlock a mun SOI simulation for example. Maybe require a certain amount of science data before it unlocks

Edit2: If you guys like these ideas I'll expand on them and see if I can work out something presentable as a real suggestion. Right now I'm just brainstorming

Edited October 15, 2014 by forsaken1111

[Noventta]

Time to build craft model for a simulation would be redundant as it is you directly who is modeling the craft in the VAB in the first place. Time to build an athmospheric model of Duna, for instance: not a problem.

[magico13]

Based on suggestions, here is the version I have tossing around in my head. This is by no means a confirmation of what I'm going to do and things are definitely subject to change.

1) In order to simulate in orbit/landed/maybe even in atmosphere you must first transmit/recover science from the appropriate place (or better, so landed could count for all three).

2) After getting science from a location, you can purchase a Simulation Package for a heap of funds and/or science which then lets you perform simulations at that location. For example, you visit Duna with an orbiting probe and can purchase the Duna Simulation Package which lets you select Duna as a simulation body. This may or may not require time to unlock. Prices will be subject to planetary mass/atmosphere constraints, but independent of distance. Moons will have their own packages that require the parent package, but will cost less than the parent package.

3) After a package is unlocked you can choose it in the editor. You can do orbital simulations if you have any science from the planet. You can do landed sims if you have landed science. If you enter the atmosphere without atmospheric/landed science the sim is terminated/restarts. Same for landed (on touchdown, sim ends). These simulations will cost a fraction of the current costs/the package cost and you'll have better control of time limits (able to set any number of hours/days you want).

I haven't decided if it's worth the effort to make subsequent simulations cheaper, but if I do it will likely be like the system I described in a previous post, since that's actually pretty easy to do.

I don't like the idea of individual simulations taking time to prep (unlocking packages, maybe) since that would suck from a gameplay perspective. Simulations are supposed to let you test a craft without having to waste a bunch of time building it, and quick test-alter-test-repeat sessions would be substantially more tedious. They will certainly take zero time (and even if they didn't, it'd be like 30 minutes).

As for purchasing hardware for additional max lengths of simulation time, I kind of like that idea. Maybe I'll just add a new window that lets you upgrade your simulation hardware: installing new planetary packages, upgrading your "processing power" or something similar to allow you to simulate larger vessels, and your "memory" to allow longer simulations (gotta store the data somewhere. Maybe instead of memory it's just "storage" since you'd actually store that data on a hard drive). You could also purchase upgrades like "Low Power Components" to reduce simulation costs by X%. I could make it like strategies where you trade one thing for another (ie 10% more processing power for 5% more money each time, with a commitment value that changes the amounts), or I could make it so there are just general upgrades that only provide benefits for an initial cost (spend 100k funds to reduce costs 5%, costing more each upgrade). I personally like the second one but I'm interested in what you all think.

So yeah, that's my current idea in my head. Like I said, nothing is finalized and I may change all of it if I get some other good suggestions

[forsaken1111]

If we start buying simulation hardware then we could sell processing time on our servers for a small funds income and/or a chance at a science boost?

[Nori]

That sounds great. My only concern with that system is that you can't do a orbital sim and subsequently attempt landing.

My reasoning is related to the fact that I use remotetech which makes it unfeasible (hard?) to send a probe to land on a non-atmo planet. That means in order to sim it I would have to send a manned mission which may not be able to get back. I suppose it could work though as Eve is really the biggest wildcard for landing that that can be landed on easily enough with probes...

[forsaken1111]

That sounds great. My only concern with that system is that you can't do a orbital sim and subsequently attempt landing.

My reasoning is related to the fact that I use remotetech which makes it unfeasible (hard?) to send a probe to land on a non-atmo planet. That means in order to sim it I would have to send a manned mission which may not be able to get back. I suppose it could work though as Eve is really the biggest wildcard for landing that that can be landed on easily enough with probes...

I do unmanned probes all the time in RT2, its just more difficult. You need relays in place prior to landing the rover or probe. If you choose to add a mod like RT2 you have to deal with the extra difficulty it adds.

I suppose you could allow a hires mapping of the planet's surface via scansat to stand in for surface data for simulation purposes though?

[Nori]

I do unmanned probes all the time in RT2, its just more difficult. You need relays in place prior to landing the rover or probe. If you choose to add a mod like RT2 you have to deal with the extra difficulty it adds.

I suppose you could allow a hires mapping of the planet's surface via scansat to stand in for surface data for simulation purposes though?

Yeah I get that. But one would think getting into orbit would allow you enough info to simulate a landing. Maybe require a certain number/type of science experiments first to indicate proper data was gathered.

Do you use KOS or something else to assist in landings? Of course you would need relays setup, I'm talking more about from the time delay perspective and being able to touchdown without exploding.

[magico13]

If we start buying simulation hardware then we could sell processing time on our servers for a small funds income and/or a chance at a science boost?

I see no reason why that couldn't be included. It'd likely be free to setup and have a slight cost to remove (that way you don't do it 100% all the time and just turn it off when you need to do a simulation, it'd be better IMO to contribute a small amount for a long time. I want some penalty so you can't say "here, use our servers. Except in 20 minutes we'll need them, and then again 10 minutes later").

I'm thinking that it could be cool if buying hardware would let you simulate craft for free, up to their limit. Past that you'd have to rent time on someone else's hardware which would have an associated cost. Maybe then there's a static "maintenance fee" each day which could be mitigated by selling unused processing time.

My only concern with that system is that you can't do a orbital sim and subsequently attempt landing.

Valid points. One of the reasons I made simulations in orbit was so you could test lander designs BEFORE trying to actually land. I'll let you land/enter atmosphere, but may restrict SOI only to ones you've purchased the package for. You'd have to have landed science to start a simulation landed on another body still, so that seems fair to me.

[Matrix Aran]

I see no reason why that couldn't be included. It'd likely be free to setup and have a slight cost to remove (that way you don't do it 100% all the time and just turn it off when you need to do a simulation, it'd be better IMO to contribute a small amount for a long time. I want some penalty so you can't say "here, use our servers. Except in 20 minutes we'll need them, and then again 10 minutes later").

I'm thinking that it could be cool if buying hardware would let you simulate craft for free, up to their limit. Past that you'd have to rent time on someone else's hardware which would have an associated cost. Maybe then there's a static "maintenance fee" each day which could be mitigated by selling unused processing time.

Valid points. One of the reasons I made simulations in orbit was so you could test lander designs BEFORE trying to actually land. I'll let you land/enter atmosphere, but may restrict SOI only to ones you've purchased the package for. You'd have to have landed science to start a simulation landed on another body still, so that seems fair to me.

Something about reading this made me start thinking that things like selling server time sound alot like something that would be a good addition to the admin building and strategies.

[magico13]

Something about reading this made me start thinking that things like selling server time sound alot like something that would be a good addition to the admin building and strategies.

It is strategy-esque, but I'm not sure if it would fit perfectly into the strategies system. Is there a limit to the number of strategies you can accept? Or can you accept all of them? If there isn't a limit and I can either create a new category or can easily add what I want to an existing one then I may put it in there. If I'm restricted to only modifying currencies with the strategies system then it definitely won't help me.

I know for a fact that I can easily do what I want by not tying into the stock systems, so I'm really tempted to just do it my way and not mess with strategies, if only because I don't have loads of time to figure all the new things of 0.25 out. I have played for a total of about an hour so far

[hatterson]

Valid points. One of the reasons I made simulations in orbit was so you could test lander designs BEFORE trying to actually land. I'll let you land/enter atmosphere, but may restrict SOI only to ones you've purchased the package for. You'd have to have landed science to start a simulation landed on another body still, so that seems fair to me.

What about allowing simulations to land on planets without atmosphere simply by purchasing the relevant planet/moon package, but for planets *with* an atmosphere you need to have science from within it.

To logic being that if a planet doesn't have atmosphere there's nothing stopping you from simming a landing short of knowing the terrain which is presumably accomplished by the probe/kerbal that returned the science, whereas if it does have an atmosphere you don't really know what it'll do to your craft or what it's made of before you've experience it.

From a gameplay perspective it fits, even with mods like RT2 with signal delay, because it's still fairly easy to land an unmanned probe on a body with an atmosphere so you're not adding significant difficulty.

That way you can still sim your very first minimus lander without having to put a probe down, but if you want to land on Laythe you need to at least send a probe down to get atmospheric readings of what you'll be dealing with.

As far as sim "build" times, I'm not a fan of an individual sim taking time, but I do like the idea of hardware taking time to build. What about breaking it into two categories, hardware and software?

Hardware costs cash and, like ordering something off newegg, it takes a bit of time to ship and then you "install" it. Simulation time on this hardware can be sold for a small amount of cash at first and you can gradually upgrade it to include more processing power/storage. Maybe for simplicity sake just leave it at a simple "upgrade" option without delving into upgrading CPUs, GPUs, RAM, disc space, etc.

Then there's software packages. These would be your base sim suites. You'd have a Kerbin suite, a Mun suite, a Jool suite, etc. Each of these cost science and take a bit of time to develop but once it's complete you own that sim package. Simulations would be priced based on the size of the vessel you're simming and the package you're using. Simming a massive vessel in the Jool package would be quite expensive, but simming a tiny one in the mun package would be cheap. Additionally packages would have certain hardware limits, ie. you need at least 3 "upgrades" to your hardware to use the Jool package, or 2 "upgrades" to sim a vessel that costs over X build points.

In addition, adding new sim suites would increase the value of the sim time you're selling. Perhaps you could tie this into the contract system so that a contract may appear for some other company wanting to purchase sim time for the Jool system. You'd need to provide them with X number of hours of uninterrupted sim time within Y days in return for 10x the cash you'd normally get from selling time. Failing to do so costs you reputation and a little bit of cash (a cancellation fee). Perhaps with each completed contract you have a random chance of receiving science based on its complexity (not sure if the contract system supports random rewards)

Perhaps make it so that the only way you can make money selling sim time is with the contract system? Other than that it's assumed you're selling the excess processing power you have to a random bidder and basically receiving back what it costs to run the server (ie, no money change if you're sitting idle). The cost for sims is justified based on the fact that you can't sell that processing power while you're using it so you have to actually pay to upkeep the server. That also justifies the larger cost for simming in more advanced packages. If you have a massive server farm and are only simming on the Mun you can still sell most of your power, but if you're simming for Jool you need it all for yourself so you can't sell any.

[Nori]

What about allowing simulations to land on planets without atmosphere simply by purchasing the relevant planet/moon package, but for planets *with* an atmosphere you need to have science from within it.

To logic being that if a planet doesn't have atmosphere there's nothing stopping you from simming a landing short of knowing the terrain which is presumably accomplished by the probe/kerbal that returned the science, whereas if it does have an atmosphere you don't really know what it'll do to your craft or what it's made of before you've experience it.

From a gameplay perspective it fits, even with mods like RT2 with signal delay, because it's still fairly easy to land an unmanned probe on a body with an atmosphere so you're not adding significant difficulty.

That way you can still sim your very first minimus lander without having to put a probe down, but if you want to land on Laythe you need to at least send a probe down to get atmospheric readings of what you'll be dealing with.

I like this idea.

Magico - Are you shooting for a simple comparability release soonish, or are you going for a more feature release including the changes you've talked about?

[Noventta]

What about allowing simulations to land on planets without atmosphere simply by purchasing the relevant planet/moon package, but for planets *with* an atmosphere you need to have science from within it.

I agree on that. Maybe from a gameplay perpective it adds something to need to visit it first but it would be nice to have an option to disable it separately.

Edited October 16, 2014 by Noventta

[magico13]

I like this idea.

Magico - Are you shooting for a simple comparability release soonish, or are you going for a more feature release including the changes you've talked about?

I'm liking the ideas that hatterson provided as well, especially with regards to atmospheres.

I'm planning on making some finishing changes to the current Beta tonight and hopefully doing a full release tomorrow. The bigger changes will be in a future update. Since we're talking about it and its exciting I might do it for next update, but I've got other features I need to put in first (vessel rollout). And the whole switch to the other inventory backend is still like halfway done and not anywhere near functional.

I agree on that. Maybe from a gameplay perpective it adds something to need to visit it first but it would be nice to have an option to disable it separately.

Don't worry, when I redo the simulations everything will be configurable. Don't want to pay for simulations or upgrades or deal with upgrades at all? No problem. Don't want to have to visit atmospheres or even planets before simulating there? Done. And also likely configs for increasing or decreasing prices for things. I like making things configurable

[magico13]

Just uploaded another beta version, 1.0.2.3. Current Beta users should be notified through the KCT update checker if it is enabled. Only change is that the default mass for reconditioning is double what it used to be and simulation costs have been cut substantially. Actual rework of simulations will occur in a later update (likely to some degree, if not entirely, in the next one). If this beta is without severe issue, then I will promote it to the official release after a recompile in release mode and a version change. I don't foresee any issues with this beta though.

[hatterson]

And the whole switch to the other inventory backend is still like halfway done and not anywhere near functional.

I'm forgetting if this was mentioned earlier, but will we be able to sell inventory items for cash with the new system? Specifically I'm thinking along the lines of using a mod like Karbundium or Interstellar to mine/collect rare fuels and sell them for cash. Basically making a mining company to raise funds to keep exploring

[magico13]

I'm forgetting if this was mentioned earlier, but will we be able to sell inventory items for cash with the new system? Specifically I'm thinking along the lines of using a mod like Karbundium or Interstellar to mine/collect rare fuels and sell them for cash. Basically making a mining company to raise funds to keep exploring

Yes, I plan on adding a way to sell parts/resources. I'd like it to be a part of the backend system, but from discussions it sounds like each mod will have to implement their own method. We might end up doing a mixed system where the backend has a simple system in place. It's absolutely necessary in my opinion that you be able to sell parts and especially resources, for the very reason you state. There are some mods that add resources that are only designed to be sold for funds, and I don't want to make those not functional anymore. That would be sad

[mRn]

Only change is that the default mass for reconditioning is double what it used to be and simulation costs have been cut substantially.

So basically the game is easier now? Damn you, freaking casuals!

No, seriously, biggest fan of your mod here. I play my KSP with nearly 80 mods and I would call yours the most essential one because it is the very basis to make the career mode playable. I have set up a quite challenging difficulty with hardmode (no revert + your simulation feature = best thing ever) and more individual sciencegain-nerf through scienceparams.cfg as also probably every known mod which adds further challenges like FAR, DRE, RemoteTech, Mission Controller Extended, TAC-LS, DangIt, actually no MechJeb but kOS and VOID and a ....load of Partaddons. The only thing I didn't yet figure out is how to make the part about funds more interesting. I had my first satellites now at Mun and Minmus and I am still at 1,25m rockets techtreewise, but I already earn money on a level that I stopped caring about it at all and also stopped using Parachutes for your StageRecovery because it doesn't matter at all. So In my opinion it would be nice if there will be a way to force higher fundspending. I read here that simulation at other planets are absurdly costful, I didn't made it to that point yet, but in my opinion it would be nice if the simulations at kerbinlaunch are higher (actually a magnitude higher). Either way, keep up the good work!

Oh by the way, just out of curiosity: Do you know if that incompatibility with the Tarsier Mod is something that went wrong on your side and can be fixed by you at some time or is it something the Tarsierguy has to patch to make both mods compatible? It wasn't hard for me to decide which mod I dump but nevertheless I kinda miss the additional ways of science from his mod in my game.

Edited October 17, 2014 by mRn

[shand]

@mRn, change your difficulty level to give less funds and higher penalties. personally i've got it set on hardmode and do sandboxey things in addition to contracts, thus making money a bit tighter - but shimmy the funds down a tad if you want more challange (can be done on an active save)

[Galenmacil]

Don't worry, when I redo the simulations everything will be configurable. Don't want to pay for simulations or upgrades or deal with upgrades at all? No problem. Don't want to have to visit atmospheres or even planets before simulating there? Done. And also likely configs for increasing or decreasing prices for things. I like making things configurable

How about no simulation at all? I mean, as an option of course... Would make the game even more challenging: Build it or don't. Want to test, still? Then by all means test... but with consequences!

I am a fan of your mod as it is. Just adding my thoughts on the simulation subject.

[Reisdal]

Trying the beta version. For some reason the build time isn't updating so warping until completion just continue to warp without the counter going down. Any one who knows if there's any mod conflicts?

[magico13]

Trying the beta version. For some reason the build time isn't updating so warping until completion just continue to warp without the counter going down. Any one who knows if there's any mod conflicts?

Have you made sure the mouse cursor isn't hovering over any of the KCT windows? Unfortunately there's a bug with stock KSP (since 0.24) that causes time to stop advancing when the KSC locks are activated (which shouldn't be related at all and I'm very frustrated about it). I have to enable those locks because of another bug in 0.24 that causes windows to get clicked through to the buildings behind them. I'm very unhappy with the GUI changes Squad made in 0.24 and then subsequently ignored my bug report about, only to later realize were a big issue and (possibly) haven't realized that they weren't issues prior to 0.24.

I get unhappy about that particular bug's existence, I'm sorry.

So yeah, check for that first, if it still doesn't change and you're sure the cursor is off the window then it's definitely a bug.

[severedsolo]

Don't know if you saw this the first time I posted but this bug is still in the new DLL:

When you exit a simulation, any parts that have been recovered during the simulation are taken into account in the build time estimate (ie they are put into the inventory).

This is cosmetic only, if you go ahead and build the parts are removed and the true build time is used (ie, in my case the craft was going to take 5 days with the simulated parts, but the build time went to the proper time of 10 days when I actually hit build). I'm also using stage recovery so might be a bug with that instead. Output log (sorry it was a long session): https://www.dropbox.com/s/ivrtklna104p7ty/output_log.txt?dl=0