Improve build controls - they stink (relative to RL software)

[Jojomo]

I would say (others may disagree) that building/designing is the core, the base of this entire game.

The devs are to be commended on the building/designing controls they've created: they started from scratch, and achieved a functional result. I'm certain it wasn't easy, and took a huge amount of work. I'm grateful for what's there, and I don't regret buying this game.

I have to say though, that for me, the result just isn't good enough. The reason is speed, or lack of it. Build controls are clumsy and very basic, and as a result very, very slow to build with.
I estimate (and this isn't a random exaggerated number picked out of thin air, it's the most accurate estimate I can make) that building a 3D model of a ship in SM (which is what actually happens) takes me approximately 100 times longer than it would take me to model in other modelling/drawing/engineering software. Honestly, what takes 50 hours in SM I can do in about half an hour with other software.

I model ships with other software at my job, the only significant difference in process/method is that these models aren't played with inside SM afterwards.

What should have happened, and what needs to happen, is that the devs need to look at and be familiar with real life modelling software (or go and find a dev to employ who is) and "steal" the best bits of it for SM. Some well known examples might be AutoCAD, SolidWorks, Rhino3D, etc...
I don't expect or even want SM to be as fully mature and powerful as these other pieces of software, but it should have the main capabilities of actual modelling software.

I don't intend that people should be able to spit out large ships at a ridiculous rate, but I definitely don't think the controls we use should be the limiting factor. (It should be resources, or shipyard capacity, or something else)

Some SM players are time-rich (teenagers between semesters for example) and aren't terribly affected by this. I'm extremely time poor, and although I can think of many exciting things I'd LOVE to do and try in this game, I can't, because building just takes too long (because the building controls are horrible).
Perhaps teenagers with spare time are 99% of the target market for SM, I wouldn't know. But if people like me are a significant part of that target market, then something needs to change. I have many, many extremely positive things to say about this game, but I wouldn't actually recommend that my friends (also time poor) actually buy it.

I'm giving up on my current build, because I haven't had time to touch it for weeks. If anyone is interested in using it to create something, feel free: Osteo Shuttle (It's not actually a shuttle, it's 180m long)

I'm not leaving Starmade - it's an incredible game with amazing potential - but my designs from now on will be relatively simple (few surfaces and details) and will be modelled in external software then imported in with some of the conversion software people have made. It's very disappointing for me to limit myself like this, but it's the only feasible way for me to continue.

tl;dr : Build design controls are functional but clumsy and SLOW. Use real life popular modelling software as guide for what controls should be.

For the record (I want my criticism to be constructive), the basic modelling functions I'd recommend, approximately from most important to least (but everything in this list is important!) are:

• Powerful, fast, flexible copy, paste, select, and move (all with a reference point)
• Keyboard shortcuts for everything
  
• Draw lines, rectangles, circles with the selected block
• Draw planes with the selected block
• Draw spheres/ellipsoids and cuboids with current block
  
• Offset a curve or surface a specified distance
  
• Mirror as a function, not fixed planes
• Rotate selection to any angle
  
• Extrude lines/curves to planes/surfaces
• Loft multiple curves to a surface
• Chamfer/fillet corners

 

[Lukwan]

• Powerful, fast, flexible copy, paste, select, and move (all with a reference point)

• Keyboard shortcuts for everything

• Draw lines, rectangles, circles with the selected block

• Draw planes with the selected block

• Offset a curve or surface a specified distance

• Mirror as a function, not fixed planes

• Rotate selection to any angle

• Extrude lines/curves to planes/surfaces

I'm pretty sure I know what each of these are good for and I would certainly use them. I would also ask that the sliders for XYZ axis be scale-able in the menu of the server-admin (or in the case of SP: the player).

These may require more explanation:

• Loft multiple curves to a surface

• Chamfer/fillet corners

 

[IG-1000]

Idk I'm pretty happy with the current build mode, it beats the ridiculousness that you get with Minecraft anyways and that's just fine with me lol

 

[jontyfreack]

sorry to slightly burst the bubble here but starmade is a voxel based building game, so many of those things you would want I don't think would be possible. HOWEVER I think everyone wants a better copy paste, even though I don't have a problem with it currently I can see why people want it to be easier to use and be able to set what block it will paste from (if that is even how one would say that, English is not my strong point)

so in short, I only agree with the better selection/copy and paste bit of this, because well a lot of people already want it to be improved and I don't see why not and this thread might give a few more ideas into how it could be improved.

 

[nightrune]

All of these are possible I think. Maybe not feasible though. The one that pops out as infeasible is the rotation at any angle. The reason they haven't been working on stuff like this is it can taken an incredible amount of time fiddling to get it right.

EDIT: Spelling

 

[Valiant70]

I would have to agree with a lot of this. I spend a large amount of time just making the base hull for a lot of my builds (and I'm talking 5k-20k mass, not Death Star stuff). I'd rather be able to create interesting shapes and wedge them with little effort, then spend my time decorating them. It would make the building process more enjoyable and allow builders to try out different things more easily.

 

[sayerulz]

Just give me a "fill" function similar to the one in From The Depths, and the ability to auto-fill build helper shapes, and it will easily take off a third of my build time. That much should be doable I expect.

 

[Mr.Steam]

Comparing a 15 dollar game to 1000 dollar professional software suites seems a bit unfair. Plus I think adding a bunch of additional functions will just scare people away. Most players probably don't want to take the time to learn what professionals spend months or years perfecting. It seems like a ton of work would be needed for a small segment of the player base.

 

[Jojomo]

I'm pretty sure I know what each of these are good for and I would certainly use them. I would also ask that the sliders for XYZ axis be scale-able in the menu of the server-admin (or in the case of SP: the player).

These may require more explanation:

Here's lofting: You select a series of curves, and a surface is created that flows from one to the next, all the way along the series.


and filleting: Filleting is just curving the edge between two lines, or surfaces, into a smooth radius.


[doublepost=1479598215,1479598092][/doublepost]

sorry to slightly burst the bubble here but starmade is a voxel based building game, so many of those things you would want I don't think would be possible.

Some would have limitations, but all of them are very much possible. All these things existed in older software that worked in terms of pixels, and voxels function in an identical manner.
[doublepost=1479598425][/doublepost]

All of these are possible I think. Maybe not feasible though. The one that pops out as infeasible is the rotation at any angle. The readon they haven't been working on stuff like this is it can taken an incredible amount of time fiddling to get it right.

It's true that rotation would be more useful for larger objects - smaller objects turned to small angles would look messy. But everything would benefit from 45 and 90 deg rotations, and large objects would benefit from rotations to any angle.
It may be worth pointing out that the latest MCEdit has a rotate function like this.
[doublepost=1479599167][/doublepost]

Comparing a 15 dollar game to 1000 dollar professional software suites seems a bit unfair. Plus I think adding a bunch of additional functions will just scare people away. Most players probably don't want to take the time to learn what professionals spend months or years perfecting. It seems like a ton of work would be needed for a small segment of the player base.

This isn't a comparison. The building controls are (for me) inadequate (too slow) looked at purely in isolation: I don't have enough time to build the things I want, in the time I have, but I would if the controls were better.

Other software only comes into it when a dev asks "How can I make the controls better?". The answer is: "Look at existing software that already does it better."

Additional functions can be ignored (a general truth). The only replacements of existing functions I suggested are the first point, and mirroring. Anyone who wouldn't want to know anything about using the rest wouldn't have to. (And although they may seem a little daunting to someone who hasn't used them, they're actually easy to get to know and use)

I honestly have no idea whether players like me are a large segment of the (potential) player base, or a small segment. (I suspect you probaby don't either). If we're a small segment, then yes, it makes financial sense not to bother. But I hope that's not the case. As for expense, I imagine they could hire a dev for a while to work only on this and nothing else, and that the cost of that would be covered by future players like me even if we're a minority (which is an unknown to us players).

The existing controls are an improvement over vanilla minecraft building controls (obviously), but structures in this game are often at a massively different scale to minecraft, and better tools are needed than the existing ones.

 

[Madman198237]

I really would like to see more advanced build options, and these fit the bill. Drawing off of existing profession design software is a good idea. Those programs provide layers of functionality that this game can use. Also, they're often newbie-friendly, mostly. Its not hard to pick up some CAD skills. I've done it myself....on accident.

 

[psteiner]

Here's lofting: You select a series of curves, and a surface is created that flows from one to the next, all the way along the series.


and filleting: Filleting is just curving the edge between two lines, or surfaces, into a smooth radius.


[doublepost=1479598215,1479598092][/doublepost]

Some would have limitations, but all of them are very much possible. All these things existed in older software that worked in terms of pixels, and voxels function in an identical manner.
[doublepost=1479598425][/doublepost]

It's true that rotation would be more useful for larger objects - smaller objects turned to small angles would look messy. But everything would benefit from 45 and 90 deg rotations, and large objects would benefit from rotations to any angle.
It may be worth pointing out that the latest MCEdit has a rotate function like this.
[doublepost=1479599167][/doublepost]

This isn't a comparison. The building controls are (for me) inadequate (too slow) looked at purely in isolation: I don't have enough time to build the things I want, in the time I have, but I would if the controls were better.

Other software only comes into it when a dev asks "How can I make the controls better?". The answer is: "Look at existing software that already does it better."

Additional functions can be ignored (a general truth). The only replacements of existing functions I suggested are the first point, and mirroring. Anyone who wouldn't want to know anything about using the rest wouldn't have to. (And although they may seem a little daunting to someone who hasn't used them, they're actually easy to get to know and use)

I honestly have no idea whether players like me are a large segment of the (potential) player base, or a small segment. (I suspect you probaby don't either). If we're a small segment, then yes, it makes financial sense not to bother. But I hope that's not the case. As for expense, I imagine they could hire a dev for a while to work only on this and nothing else, and that the cost of that would be covered by future players like me even if we're a minority (which is an unknown to us players).

The existing controls are an improvement over vanilla minecraft building controls (obviously), but structures in this game are often at a massively different scale to minecraft, and better tools are needed than the existing ones.

Although these building tools would be great in Starmade, Starmade hasn't been designed for this level of building and consequently is missing several of the key features that 3d modeling software uses to achieve results. The software that I use (Autodesk Inventor and Sketchup) both rely on the definition of planes and surfaces between lines and other 2d and 3d objects. For example, the fillet tool looks at two planes and defines a curve at a set edge between the two. However, this would be significantly harder to do in Starmade as the planes are not already predefined.

Of course these ideas could be implemented by using a select tool much like the copy tool to define planes. However, with the use of voxels, a second step would be required to interpolate blocks around the created mathematical surface. In a sense, you and the game would have to (In context to your loft tool assuming we leave out the end point):

1. Define a mathematical edge of best fit around all block shapes used as intermediaries for the tool
2. Select which edges you want to use
3. Define the order order of shapes (Most likely done automatically)
4. Calculate the mathematical projection in 3d space for all edges, interpolating between the 2d edges
5. Calculate the best fit for voxels for the 3d shape
6. Either project a build guide or fill in the voxel result with a selected type of block

Of course this feasibly could be integrated into the game. It would just require some new tools (Suggestions follow):

• Defining 2d shapes (2d sketches in Inventor)
  • Could be accomplished by using a select tool to define adjacent blocks as part of a shape and displaying that group as a "shape" with a visible interpolated mathematically defined edge (example)
  • If shape terminates before loop is completed, loop is closed with a straight line
  • These shapes would be selectable
    • Would most likely need to have a "select shape" option in adv build mode which, toggles on/off (When selected you select shapes instead of placing or removing blocks
    • If selected, could be deleted with backspace or delete
• Extrude tool (Push/Pull tool in Sketchup
  • When used, you would be prompted to select a shape (automatically open select shape tool if one is not already selected), then a dialog box would open asking for the distance of extrusion (Up/ down for +- x blocks) and if it was to be filled or not on the top, inside, and lateral surfaces (checkboxes)
  • This tool would then extend the 2d mathematical model by x blocks length, interpolate voxels from this projection, and place the currently selected block where the voxels are projected to be placed. As long as voxels -> edge and edge -> voxels is forwards/ backwards compatible and does not lose data between switches, there should be no error.
• Loft Tool
  1. Prompt select shapes. Once done select "done" to end select and move to step 2
  2. Prompt fill on inside, ends, and lateral surfaces (checkboxes)
  3. Calculate 3d mathematical projection from selected 2d shape mathematical edges, project voxels and fill with currently selected block
• Fillet Tool
  • Prompt select shapes (Must be two perpendicular 2d planes with a common edge) and moves on after two are selected and "done" is pressed
  • Program detects commonly shared blocks (intersection) and uses that as the axis of fillet rotation
  • Prompts for fillet radius
  • Calculates curved lateral surface
  • Example:
    
  • Projects voxels from mathematical shape and filled with currently selected blocks

In all essentials, this method could be used to introduce a wealth of build tools to Starmade. Overall, my biggest concern is with 2d shape selection. The method I proposed would be slow and heavy of resources. It may be more effective to use this method to select 2d shapes:

• 2d Shape Tool
  • Calculates curves only from blocks adjacent to each other or diagonal to each other. Any block farther away would add a straight "jumper" line to the shape between the previously selected block and the selected block.
  • Blocks would be ordered in the order they were selected in to prevent shape edges to jump between any isolated block groups.
  • Jumper lines would only occur between the last adjacent/diagonal block (or the last block if the last block was not adjacent/diagonal) and the currently selected block
  • If the tool icon was selected again, the shape is completed
    
  • If shape terminates before full loop is completed, loop is closed with a straight line

 

[jayman38]

I disagree with the OP. Starmade should not get more building helpers. Starmade should get an OBJ importer. It can be imperfect, that's OK. The bottom line: Use the right tool for the job. So if CAD software can already do it, yes, create a 3D obj file and then import it into Starmade at a user-defined scale. (I don't think SMEdit counts, because it cannot adequately keep up with changes to the game.)

It would be really, really nice if the importer were to implement wedges and other angles, but I can put those in manually after the import if necessary.

I'm trying to remember if obj files can have different "material" settings. That way, you can model systems and different blocks, and then set those different blocks to certain block types on import. It would be a shame to import a model full of nothing but grey hull.

 

[Lone_Puppy]

I to am plagued with being time poor to build in this game and wish for a lot of features mentioned. One thing I don't see mentioned a lot is touch device or gesture controls, or at the very least mouse centric controls. I wish for more of these, because sometimes I just don't want to sit at a keyboard for several hours on end.

Effective and up to date conversion tools like SMedit would be a simple way to speed up build time. It would have to be intuitive enough to use wedges and slabs. That way, you would only require the built in SM editing tools for systems, modification and fleshing out the over-all design.

I'm giving up on my current build, because I haven't had time to touch it for weeks. If anyone is interested in using it to create something, feel free: Osteo Shuttle

I like where you were heading and those spikes look cool!
Using that plug-in is a great idea. Being able to rotate and zoom in/out to see more detail inside/out is so much cooler than looking at snaps or videos.

 

[Jojomo]

I disagree with the OP. Starmade should not get more building helpers. Starmade should get an OBJ importer. It can be imperfect, that's OK. The bottom line: Use the right tool for the job. So if CAD software can already do it, yes, create a 3D obj file and then import it into Starmade at a user-defined scale. (I don't think SMEdit counts, because it cannot adequately keep up with changes to the game.)

It would be really, really nice if the importer were to implement wedges and other angles, but I can put those in manually after the import if necessary.

I'm trying to remember if obj files can have different "material" settings. That way, you can model systems and different blocks, and then set those different blocks to certain block types on import. It would be a shame to import a model full of nothing but grey hull.

That's a workaround, not a solution. I will be modelling externally and using this to import in future.