[WIP] Orbital Manufacturing

[mrrpamplemousse]

Nice sketchs, looking forward for this.

[Yorik]

Here's a quick video on how I'd do the drill, very early progress.

[Patupi]

Very cool, and fast work Yorik! Thanks. I like the guide on the door to stabilize the drill.

One thing, I think we should decide on a sort of look for the parts, just to keep things consistent. I'm meaning details like if we go for soft, rounded edges, sharp corners, cowls with rivets etc. I was actually thinking of something like the last one when I drew the drill originally, but I kind of like the way the cowl you did works.

As a quick glance at it, I'd say make the main drum for the drills a little less rounded... I think. Just a first glance opinion. that came to me. Mainly just seems it would fit better with the rest. Not sure why given that the cowl you have there is smooth... *shrugs*

EDIT: I tried to do a detailed pic of what I was saying, sort of half way between my original idea and the rough you had in the video. (And yes, I got the number of drills wrong below. Somehow it ended up with seven! *lol*)

The hinge thing was to try to emulate the slide, popup, then swivel you had in the video. I think this would work if we can get it to look as if the mechanics do the motion and it's smooth. Cool idea in that.

Edited November 8, 2013 by Patupi

[AdmiralTigerclaw]

A suggestion from me.

The biggest thing about orbital construction and manufacturing is that it allows you to start building bigger.

But with bigger you run into the limitation of the game engine.

I would like to suggest two or three tiers of station parts for orbital construction and manufacturing. The first being the normal station parts, which are used to build larger construction components like orbital scaffold pieces. (You COULD build scaffolding parts on the ground, but you need huge rockets to get them into orbit. Just build in orbit.) Then you construct a new station elsewhere with these so that your station part count doesn't go crazy. And a third tier for a full on orbital shipyard. (Single Parts, hundreds of meters long/wide.) That way, the building of bigger and better stations doesn't run into the problem of KSP trying to do a physics count on a 1500 part structure.

And of course, with bigger parts, comes bigger modules. Ideally, Scaffolding and stations would have 'core' modules of some kind that have plug-in points for other huge components that have their resource functions scaled up accordingly. I tried to build a large station a week ago and quickly ran into the problem of parts scaling very, VERY quickly. (Didn't even have the power module subsection attached before I hit the framerate wall.)

[123nick]

this mod is awesome. i cannot WAIT for the full release! hows development? like, % finished?

[johnsonwax]

That way, the building of bigger and better stations doesn't run into the problem of KSP trying to do a physics count on a 1500 part structure.

Welding can handle an awful lot of that problem.

[AdmiralTigerclaw]

Welding can handle an awful lot of that problem.

A cumbersom solution at best. I've already used the welding tool, and while it is indeed rather cool, it's also clunky and ill-suited to what I envision. Especially if you're trying to work in career mode. You have to create your part (which includes coming up with a name and description for it), and then wait for the component file to generate. Then you have to shut down the game, find the new part config in the files, assign a tech tree point to it, and restart the entire game.

Now build 37 large superparts in this manner and see how fast you get tired of the routine. I was pretty much 'done' with it after four.

[attosecond]

this mod is awesome. i cannot WAIT for the full release! hows development? like, % finished?

About 35%. There's a fundamental flaw with the way ExLaunchpads does vessel construction in orbit, and unfortunately the alternative is much less clean. Working on it, now.

[Delwin]

You have to create your part (which includes coming up with a name and description for it), and then wait for the component file to generate. Then you have to shut down the game, find the new part config in the files, assign a tech tree point to it, and restart the entire game.

That sounds like a limitation of the welding tool rather than the concept. The tool should be able to assign tech tree location at a minimum.

Edit: Just looked at the latest state of the mod. It's currently disabled in career mode because access to the tech tree is still limited. It does however auto reload the database so you don't have to shut down the game anymore.

Edited November 11, 2013 by Delwin

[attosecond]

I don't suppose there's any way to get KSP to reload a subset of the parts without reloading the whole engine?

There is: from the SpaceCenter scene (not in the VAB, SPH, etc, but just looking at all the buildings), hit <Alt-F12> to bring up the debug toolbar. Click the "Database" button, the click the "Reload All button." This only works for parts, not plugins.

[boberts314]

A cumbersom solution at best. I've already used the welding tool, and while it is indeed rather cool, it's also clunky and ill-suited to what I envision.

Problematic though it may be, it does at least allow for some creative freedom on the part of the player. I tend to delete pre-designed components like you suggest simply because I see designing your own ships and equipment as fundamental.

[johnsonwax]

Right. The welding concept is not flawed, but it's new to the game and there's exactly one effort underway to utilize it. I'm sure that will improve, either from Squad or from more competition from mod makers.

But there's nothing wrong with welding modular components for large craft. It's faster than modding a new part in and performs better than not welding them, and with a bit of forethought and planning, you can reuse that part over and over. I have an orbital tug whose only job is to grab ship and parts and move them into position. It's got a nuke drive and a bit of fuel and a ton of RCS and thrusters. It has lights and a few other odds and ends to make it a viable ship, and I welded all but the engine and docking port into one part, so it's literally a 3 part ship that started as probably 20-30. I always have two on the station (6 parts vs 40-60) and I often send one up with a large station module because I use the tug to achieve orbit and handle rendezvous, and then I'll deorbit one. I've probably launched 20 of them, and I can still attach a new sensor, put a different engine on it, etc.

The spine of the station is made of welded parts as well. The whole external truss is a single welded piece for each module. To that I can attach anything I want, but it's the part that connects to other station truss elements and has the little aesthetic flourishes that make it look like a station - panels, lights, etc and it's just one part that I use over and over. I probably have 10 of them on the last station I built.

[123nick]

Right. The welding concept is not flawed, but it's new to the game and there's exactly one effort underway to utilize it. I'm sure that will improve, either from Squad or from more competition from mod makers.

But there's nothing wrong with welding modular components for large craft. It's faster than modding a new part in and performs better than not welding them, and with a bit of forethought and planning, you can reuse that part over and over. I have an orbital tug whose only job is to grab ship and parts and move them into position. It's got a nuke drive and a bit of fuel and a ton of RCS and thrusters. It has lights and a few other odds and ends to make it a viable ship, and I welded all but the engine and docking port into one part, so it's literally a 3 part ship that started as probably 20-30. I always have two on the station (6 parts vs 40-60) and I often send one up with a large station module because I use the tug to achieve orbit and handle rendezvous, and then I'll deorbit one. I've probably launched 20 of them, and I can still attach a new sensor, put a different engine on it, etc.

The spine of the station is made of welded parts as well. The whole external truss is a single welded piece for each module. To that I can attach anything I want, but it's the part that connects to other station truss elements and has the little aesthetic flourishes that make it look like a station - panels, lights, etc and it's just one part that I use over and over. I probably have 10 of them on the last station I built.

the thing is with welding is that it reloads ALL the parts. if you have lots of mods, you cant afford to reload each and every part without crashing because the immense ammounts of ram required to do that. its like having 2x all the mods you have now.

[Delwin]

the thing is with welding is that it reloads ALL the parts. if you have lots of mods, you cant afford to reload each and every part without crashing because the immense ammounts of ram required to do that. its like having 2x all the mods you have now.

64-bit KSP would solve this but we still have no ETA on Win64 KSP

[evilC]

Hey all, been AWOL a bit for a while, back playing a bit of career mode casually and stumbled across this thread.

Glad to hear someone is finally trying to solve the big problem I had with OC that was stopping me doing the overhaul I so wanted to do - pretty much the exact same list of goals you listed...

Obviously my code was open source, so feel free to nick any parts of OC or EL that I wrote that interest you, but I doubt it, it was a bit of a mess

Wishing you all the best

[boberts314]

Hey all, been AWOL a bit for a while, back playing a bit of career mode casually and stumbled across this thread.

Glad to hear someone is finally trying to solve the big problem I had with OC that was stopping me doing the overhaul I so wanted to do - pretty much the exact same list of goals you listed...

Obviously my code was open source, so feel free to nick any parts of OC or EL that I wrote that interest you, but I doubt it, it was a bit of a mess

Wishing you all the best

Hey evilC! I'd wondered where you'd gotten off to. Glad to see you're still around, even if you never managed to solve the limitations in skykooler's code. Having fun with .22?

[Helix935]

i know you are making your own original parts for this but if you want an idea of want more you can do for this you may want to consider seeing this

[Patupi]

Interesting, totally Kerbalized Seems like the tube they're in is a bit small, but I love the way they are attempting to implement the 'stupidity' factor and even training them! It'll be something to watch. I'd still like to see some automation, even if it's always more effective to use Kerbals. Sometimes You might want a fabricator out in the deep dark where there are no Kerbals to man things.

[evilC]

Hey evilC! I'd wondered where you'd gotten off to. Glad to see you're still around, even if you never managed to solve the limitations in skykooler's code. Having fun with .22?

Skykooler does EL. AFAIK that has no such issues - it was OC that had the horrific teleport hack. Not sure who wrote that, but seeing as a better technique still has not been found, I guess we can't criticize too much eh? I seem to remember that a couple of the more knowledgeable coders had ideas of how to do it differently, but I never really understood the mechanisms (rails etc) to appreciate what they were saying. If you haven't asked around in the dev channel on IRC, I recommend doing so...

I love the idea of science, but the implementation leaves me a little cold - too much working out what experiments are available where, and then too much clicking while you are trying to fly.

I am hoping that something like the R&D overview mod can solve these issues though, but I think the dev doesn't know how to work out the available science.

I think if KSP were more accessible as a casual game then I would play more - I normally have an hour or two to spare, so I would like to be able to fire up KSP, use a tool like R&D to plan experiments, build ship, fly mission, experiments triggered automatically, land, return trip, done.

As it stands I spend way too much time trying to work out what experiments have been done in what biomes on what planet etc.

Also, holding off on diving in more until mission cost is implemented. I spent so much of my previous playtime on spaceplanes that I would like to be able to put it into practice to save costs.

[boberts314]

Skykooler does EL. AFAIK that has no such issues - it was OC that had the horrific teleport hack.

I was referring to the issues that kept his launchpads from being used in orbit. I'm sure I recall you being around when it sounded like some of the modders might actually manage to overcome that and apply the launch method to OC, although to be fair that was ages ago now. Could just be my memory going.

I'm running into some of the same complaints about research. I think maybe the biomes are intended to allow people a choice between doing repeated missions on one solar body or pushing out to visit different ones? To me though, it feels kind of like I'm leaving a job undone. Completionist mindset, maybe.

It's the diminishing returns I like least. I wish you could just set an experiment to start transmitting and have it continue transmitting until no more science is to be gained, or the conditions for the experiment are no longer met.

[Yorik]

I had some time today to work on the drill. I improved the model according to the latest sketch, but I had to come up with a different mounting for it.

[Patupi]

I like it. Straight forward hinging mechanism. If we wanted to go with excessive detail you might put a ram or something on the bottom to drive the motion as it tips, but graphically I suspect it would clutter it too much. Right now it looks good. Fluid, believable movement. Good work!

[attosecond]

I was referring to the issues that kept his launchpads from being used in orbit. I'm sure I recall you being around when it sounded like some of the modders might actually manage to overcome that and apply the launch method to OC, although to be fair that was ages ago now. Could just be my memory going.

I've tried to do exactly this, and failed... the technique is simple and slick, but it doesn't work without the player performing a save/load after construction is complete. Since this is unacceptable, I'm moving on to piecemeal construction, which incidentally dovetails nicely with a delayed construction feature.

[Yorik]

Here's the smelter. As I'm going to start work on the textures next week: any ideas for a color palette to use? Edited November 15, 2013 by Yorik

[Patupi]

I'd suggest expanding the top and bottom plates. After drawing it I realized it should probably have mounting plates that are a similar diameter to the body of the smelter. Might help the look of it too as I'm not actually happy with my drawing of that, though your rendering has improved it.

As to color, I was thinking cream, off white (maybe dirty) for the insulating tiles, and a dark, maybe oil-black metal look at the top and bottom. Might even be a tarnished or sandblasted dark metal like some kind of wrought iron or something. I guess the actual disks that mount to the rest of the ship would look more smooth and similar to Stock parts.