[WIP][1.8.x] SSTULabs - Low Part Count Solutions (Orbiters, Landers, Lifters) - Dev Thread [11-18-18]

[Jimbodiah]

Maybe the idea could serve as inspiration for a new dedicated part for this application. Just like the science bay on the Apollo SM sort of thing? I liked this on the BDB Apollo SM where you can hide stuff under the hatch until you need them. Maybe Mages sampe return idea could implement something similar and be scaled up?

Edited November 26, 2017 by Jimbodiah

[tater]

If it's that much work, then I'm not sure it's worth it to close the petals unless it has other possible utility for you. An alternate idea would be to wait until you need a cargo bay part, then make a cargo bay that opens like a petal (with or without a nosecone) like the BFS cargo version SpaceX has been showing. 

Regarding SMs, a tank model (third upper stage type?) that could have a stock-style "service module" bay door might be a worthwhile compromise for having a sort of SM part that doesn't include an engine. It's a standard SSTU US, and at the base where the fuel would be (the larger tank being oxidizer), it uses some of the difference in diameter as a place to throw science experiments, etc).

Edited November 26, 2017 by tater

[Shadowmage]

11 hours ago, kamikazeamoebe said:

Anyhow a picture of what it looks inside Linux:

Interesting.... that would indeed appear to be broken shaders, but broken in a strange way.  Specifically its the masked/recoloring shader that has gone sideways, and I would bet that it is using too many instructions/registers (uniforms, texture samplers) for the older OpenGL versions.

I've got an issue ticket open on the problem under the TexturesUnlimited repository, and hopefully I'll be able to offer a fallback shader for that scenario.  Might be week or two before I'm able to get to it, but could probably use some help in testing possible fixes if you were interested.

[kamikazeamoebe]

10 hours ago, Shadowmage said:

Interesting.... that would indeed appear to be broken shaders, but broken in a strange way.  Specifically its the masked/recoloring shader that has gone sideways, and I would bet that it is using too many instructions/registers (uniforms, texture samplers) for the older OpenGL versions.

I've got an issue ticket open on the problem under the TexturesUnlimited repository, and hopefully I'll be able to offer a fallback shader for that scenario.  Might be week or two before I'm able to get to it, but could probably use some help in testing possible fixes if you were interested.

Well for now using Windows is an easy ans simpel workaround. And i do understand that gaming with Linux ist still somewhat of a speciality with few enough active users / marginal target group. I myself am pleased to hear there might be a solution but fully understand there are more important things to it. 

Nonetheless, i'd be pleased to help - just let me know what and how!

[vossiewulf]

12 hours ago, Shadowmage said:

You mean there is no picture in the bottom-right?  Yep, because they have no IVA defined.

Yeah finally dawned on me I had several missing scientists too, ergo none of the DOS parts have IVAs. You and Hraben have IVA issues  Yes I know, rather spend time on more/better models and TU and other things. Whenever I stop consistently playing KSP there are lots of things I could do.

[Jimbodiah]

If anyone wanted to make the IVAs, I doubt Mage would object. Maybe even placeholders.

[Shadowmage]

11 hours ago, vossiewulf said:

Yeah finally dawned on me I had several missing scientists too, ergo none of the DOS parts have IVAs. You and Hraben have IVA issues  Yes I know, rather spend time on more/better models and TU and other things. Whenever I stop consistently playing KSP there are lots of things I could do.

 

1 hour ago, Jimbodiah said:

If anyone wanted to make the IVAs, I doubt Mage would object. Maybe even placeholders.

Indeed.

 

To me, IVAs are nothing more than the little picture of the kerb in the bottom-right corner of the screen.  They are not usable, not functional, and an absolute waste of my time to model/texture/work on, not to mention absolutely torturous to actually work on (seriously, modeling tools where never designed for interior spaces/shapes, only exteriors).  I never use IVA view and really have no intentions on doing so in the future.  (IF there was some sort of 'free-look' IVA where you could actually move the kerb's head around, AND they didn't have such a giant forehead making most windows useless, AND the helmets were smaller so that you could reasonably fit the crew inside the capsules, I might have a different view on it)

I'm more than willing to take submissions for IVA models, and/or PRs to add placeholders to the existing parts that lack them, but I'm going to spend zero of my own time on it.

On 11/25/2017 at 2:28 PM, Jimbodiah said:

Yep, it won't all be possible.

For my personal use I would not want engines on the parts so I can add 3rd party engines that I tend to use more, mainly NF Propulsion (I lovz my scifi).

Maybe just have a scalable (change diameter but keep size ratio) apollo and orion SM to start with? That way you can chose fuelcell vs panels by chosing the model and adjust the diameter to match the pod. For my personal use I patched a 2.5m version of the orion capsule and SM for a 3-man setup just because it looks more modern and has panels. Maybe that is an initial option until you work out something better?

I'm really against having any sort of scale-able SM.  Unless you want -very- simple textures, and no external accessories (no handrails, no antenna, no solar... so basically just a fuel tank with some rcs on it...).

(No, I have absolutely no intentions of making ETS 'short' Apollo SM's, at least not with that texturing style; if they exist, it will be as a standard fuel tank; want the ETS stuff?  Go play with BDB...).

 

This is why I asked how you would propose making a modular 'Orion-SM' part?  I really don't see it as being possible.  It is a specific shape and visual style... any change to it and it is no longer an 'Orion-SM', but something completely different.  And I don't see the appeal in doing a bunch of work for just-more-fuel-tank-variant parts.

[RedParadize]

Copy pasting stock IVA is enough for my need. The only I have spent time in is the ASET pod, but just to turn on lights and LSS, and open windows.I have seen Hardcore gamer use mods to control everything from the IVA seat. Might be interesting but I would find it boring over time. 

Edit: I did a IVA docking once. That was cool... never did it again.

Edited November 27, 2017 by RedParadize

[Jimbodiah]

Modular went out the window when you said making so many length versions and diameters is a no go. I meant just rescale the entire thing to fixed sizes, ie 1.875, 2.5, 3.75, 5m as I did in my 2.5m rescale (same aspect ratio, so basically an Orion mini-me version). Don't know if that is an option somehow?

[RedParadize]

I would take a tank with RCS mount on both ends. In fact I did make one  in the past. I would just have to update it. But I think it a modular SM would basically be that plus maybe solar panel.

Looks like I found myself something to do for tonight.

[tater]

Off topic a fair bit, but something I occasionally think about, and KIS talk on github got me thinking about base building again.

I have suggested in a couple places that a relatively simple way to make pretty realistic bases would be to have a base part that either deploys to a state buried in regolith, or in a world with a better sort of EPL, the player might place a stake where the base is to be constructed, and the already buried base is established.

Using terrain scatter I show a sort of thing it might look like---though less boulder looking, and more like a pile of loose regolith, obviously:

The final look would be an airlock protruding. Using the ability to recolor parts, it might have a set of switchable textures that match the default terrain textures on the various worlds. This being closer to the desired look than my poor attempt above:

The parts might in fact look like 1/2 versions of the expandable parts, resulting in a quonset hut looking structure, that then gets the dirt model added on top upon expansion. Obviously the collider has to be the regolith bit, so that kerbals can walk on top. 

Mars/Duna version:

 

Edited November 27, 2017 by tater

[RedParadize]

@tater @Shadowmage Well that could make great use of Textures Unlimited! A texture/color switch to match planets and moon color. Might even be possible to use Squad texture for that if permitted.

I like the idea. Always have dreamed of a truly massive base of +200 kerbal. It might be the best way to create a massive base without a complex leg/corridor/attachment etc. One can simply assume that everything is connected trough tunnel.

[tater]

6 minutes ago, RedParadize said:

@tater @Shadowmage Well that could make great use of Textures Unlimited! A texture/color switch to match planets and moon color. Might even be possible to use Squad texture for that if permitted.

I like the idea. Always have dreamed of a truly massive base of +200 kerbal. It might be the best way to create a massive base without a complex leg/corridor/attachment etc. One can simply assume that everything is connected trough tunnel.

Yes, exactly. The assumption would be that they are connected underground---assuming the player first lands the boring part (2.5m cylinder, pretty massive, with substantial power requirements), perhaps. Else they are stand-alone.

The idea is that you land infrastructure to build, then rocket parts/material kits. The size can potentially be pretty large. Land a few hundred tons of rocket parts within 1km of the workshop, and perhaps build a dome that rivals the large centrifuge part in diameter. Perhaps the first EPL-like part is a workshop that looks like a hanger. Personally, with the idea towards lower part count... I wonder if you could make a hanger with an overhead door. Any craft placed inside then gets saved when the door is closed as a craft file that is respawned as needed, but otherwise unrendered.

[RedParadize]

1 minute ago, tater said:

Yes, exactly. The assumption would be that they are connected underground---assuming the player first lands the boring part (2.5m cylinder, pretty massive, with substantial power requirements), perhaps. Else they are stand-alone.

The idea is that you land infrastructure to build, then rocket parts/material kits. The size can potentially be pretty large. Land a few hundred tons of rocket parts within 1km of the workshop, and perhaps build a dome that rivals the large centrifuge part in diameter. Perhaps the first EPL-like part is a workshop that looks like a hanger. Personally, with the idea towards lower part count... I wonder if you could make a hanger with an overhead door. Any craft placed inside then gets saved when the door is closed as a craft file that is respawned as needed, but otherwise unrendered.

Yes, basically a second KSC. Past initial base construction, I would like to fly rocket between colony without having to worry too much about how base them self work. Then contract like transport 10 kerbal from KSC Kerbin to KSC Duna or transport rare material for here to there would be awesome.

[vossiewulf]

4 minutes ago, tater said:

Yes, exactly. The assumption would be that they are connected underground---assuming the player first lands the boring part (2.5m cylinder, pretty massive, with substantial power requirements), perhaps. Else they are stand-alone.

I like the idea too, but I don't think it's very practical considering the engine design. Starting with the fact that during development they had to decide how to handle situations where some part of a vessel's/object's collider extended into that of a celestial body. After some thought, during which they abandoned the ideas of 1) simply never allow it to occur, 2) ignore it if relative velocity < 2ms, 3) move the vessel so it's not intersecting, and settled on 4) calculate the volume/mass of the intersection and convert it to pure energy as if the planet was matter and it was antimatter, and apply all resultant acceleration to the vessel. AND to any Kerbals who happen to be standing around.

So anything approaching real underground construction seems highly unlikely. The best you could do is plant a couple doors on the surface and have the whole base virtualized with those doors being the only real objects.

[tater]

5 hours ago, vossiewulf said:

I like the idea too, but I don't think it's very practical considering the engine design. Starting with the fact that during development they had to decide how to handle situations where some part of a vessel's/object's collider extended into that of a celestial body. After some thought, during which they abandoned the ideas of 1) simply never allow it to occur, 2) ignore it if relative velocity < 2ms, 3) move the vessel so it's not intersecting, and settled on 4) calculate the volume/mass of the intersection and convert it to pure energy as if the planet was matter and it was antimatter, and apply all resultant acceleration to the vessel. AND to any Kerbals who happen to be standing around.

So anything approaching real underground construction seems highly unlikely. The best you could do is plant a couple doors on the surface and have the whole base virtualized with those doors being the only real objects.

One, there is actually already a mod that puts habitats underground with IVAs.

Two, there need not be anything underground at all. We are merely assuming that they are connected underground. If you land parts within X km of a boring part you have landed, then all future facilities are considered connected. They don;t have to BE connected, just considered as such.

Using KIS/KAS, if you attach 2 crew parts with a "pipe" crew can move between them. Crew cannot walk through the small pipe, it's an abstraction.

Edited November 28, 2017 by tater

[RedParadize]

Indeed. That and not all mesh have to be covered by colider. So you can have a intersection that blend the ground quite well if its relatively flat.

Edit: I feel inspired. I will post a mockup after work. What I have in mind is quite a massive amount of work... So @Shadowmage don't take it as a request. I know for sure that what I will post is beyond the scope of what you have planed. If it inspire you then that's great. If not that's fine, I will enjoy doing it, didn't do serious modeling in years...

Edited November 27, 2017 by RedParadize

[Jimbodiah]

RoverDude's ART does as much by mining an asteroid and hollowing it out, allowing you to convert it into fuel tanks or even habs.

[tater]

The IVA has zero physics, and the IVA is offset in unity. In this case, it even renders the IVA in cutaway view, but the actual part is just the bit on the surface. 

 

Note that I'm merely suggesting that the underground bits be entirely abstracted. IVAs get visited a few times, then never again.

Edited November 27, 2017 by tater

[Jimbodiah]

Re ART: Not sure the hab part works, but he has a crew hatch that attaches to the asteroid

 

Edited November 27, 2017 by Jimbodiah

[RedParadize]

@tater Since Iva is mostly useless. The IVA view could actually show just the structure of the buried crew habitat. Just to give a feeling of the size of it.

[tater]

8 minutes ago, RedParadize said:

@tater Since Iva is mostly useless. The IVA view could actually show just the structure of the buried crew habitat. Just to give a feeling of the size of it.

That's true. 

Regardless of the underground part, the bit I was suggesting was merely an actual part, sitting on the surface, that happens to look like a pile of dirt. Ideally it would have at least a flat "hard point" on the top, so that an antenna, cupola, etc might be placed there. 

[Shadowmage]

Spent pretty much all of my working / dev time yesterday to make this possible:

Note the important bit, the 'retract panels' button...

And because  @RedParadize gave me an idea....

(which doesn't actually work like you would think because stock aero / drag cubes really suck; it flies heat shield-end backwards, sadly, even with the giant draggy wings sticking out)

Still a very buggy mess, and probably another 10-15 hours work to straighten it all out.  Was not kidding when I said it would be painful to implement because the module was never designed for the feature.  Almost would be easier to start from scratch on the module with the new feature-set in mind from the beginning.

(yes, it is Monday.... so take what I write with a grain of salt)

Edited November 28, 2017 by Shadowmage

[RedParadize]

@Shadowmage Oh... That's fantastic! Thank you so much!  I might have a idea to make it behave with stock aero, I imagine the drag cube is calculated around part colider? If yes, it probably only see the base, so increasing the angularDrag and maximum_drag might do it. If not, I will try make a part with the heat shield included, that or a base plate. Since I now I feel guilty of wasting your time, let me do these test myself to redeem myself!

Edited November 28, 2017 by RedParadize

[tater]

Will this at least be possibly useful to you later? Ie: to make similar parts for different uses you might imagine?