Hi Y'all,

I am working on "Insert Related…" in BrickSmith. The idea is simple: select a door frame, and BrickSmith can tell you which doors fit, and put them into the correct location.*

Insert Related solves two problems:

- If you haven't memorized every single Lego Part, it can be hard to find the right matched part, e.g. the right tire for the wheel, the right door for the frame, the right glass for the window. Some of these are manageable (e.g. windows), some I still don't know after tens of thousands of parts inserted (e.g. tires). Insert related is a database that "knows" which parts fit for certain special roles.

- Some parts have relative origin offsets that aren't on the major brick grid> This typically happens when a part has its origin set for easy rotation instead of easy relative placement. So even once I find the right door for the frame, how does it 'snap' in?

I have two questions:
1. Is anyone else interested in the data for other modeling programs?
2. Does anyone have any feedback on the proposed data format? (This matters mainly if anyone else wants to use the data.)

There's nothing BrickSmith specific going on, hence the inquiry.

Here's what I have right now:

122c01.dat 3641.dat -31 6 0 0 0 1 0 1 0 -1 0 0 Left Tire
122c02.dat 3641.dat -31 6 0 0 0 1 0 1 0 -1 0 0 Left Tire
122c01.dat 4084.dat -31 6 0 0 0 1 0 1 0 -1 0 0 Left Tire
122c02.dat 4084.dat -31 6 0 0 0 1 0 1 0 -1 0 0 Left Tire
122c01.dat 3641.dat 31 6 0 0 0 -1 0 1 0 1 0 0 Right Tire
122c02.dat 3641.dat 31 6 0 0 0 -1 0 1 0 1 0 0 Right Tire
122c01.dat 4084.dat 31 6 0 0 0 -1 0 1 0 1 0 0 Right Tire
122c02.dat 4084.dat 31 6 0 0 0 -1 0 1 0 1 0 0 Right Tire

Basically the format is:

<parent file> <child file> <relative transform> <placement name>

The relative transform is the matrix that a 1 directive would have for the child if the parent is at 0,0,0, unrotated, and the parts are put together according to the relationship described to the right. The relationships names help disambiguate (for UI purposes) multiple attach points, e.g. do we want to put this tire on the left or right wheel, do we want the door to open to the left or right, is this the left or right shutter?

It is intentional that the relationship name is the same for multiple lines with different parts - the idea is that the user can say two things:

"If I want to add part 3641.dat to 122c01.dat, what are my choices?"

(The answer is that 3641.dat can act as a "Left Tire" or "Right Tire", and there are different transforms to do that.)

or the user can ask:

"If I want to add a Left Tire to 122c01.dat, what parts can do this?"

(The answer is that both 3641.dat and 4084.dat fit as tires - by luck the transform is the same for both but this won't be true some of the time.)

The data is 'flat' - that is, relationships are provided only for final parts, not for sub-parts.

To create the data, I created a very simple .mpd file with the parts assembled 'correctly' (using various steps and sub-models to structure the parts) and then ran a Python script over it to get the final flat results. This means that, to create the data, I simply modeled each part once, visually checking to see if the part was positioned correctly.

Anyway, if this is of interest to anyone else working on modeling software, please let me know - I'm happy to share the data.

Cheers
Ben

* My previous post on this:
forums.ldraw.org/showthread.php?tid=8008,8025
got side-tracked into a discussion of 'true connectivity'. This is not that at all - it's just a way to insert parts. I would like to do connectivity some-day, but connectivity is an ocean-liner of a feature, and this is a rubber raft. Or something like that.

All of the current year (2013) city sets should be placed here.

I will be extremely busy until May and I am looking forward to release a new version of my pattern creator (with reference lines!) in June. Furthermore, I will try to release a first alpha version of the part creator in July.

Now with better connection to LDStructure (version 1.2.0.0 is needed).

Website

Now with better connection to LDFind (version 1.3.0.0 is needed).

Website

A number of people have mentioned that minor edits should be fast tracked on the PT. I agree. Here's my proposal:

I'm going to split this off as a new thread since the old one is unwieldy...

When we develop a list of points (via check and snap), there will be (hopefully only a few) remaining points that have not been snapped to each other and remain with only _one_ tri or quad accessing them. We can exploit this as these points must be:

• Part of a t-junction
  
• A point that has not been snapped properly
  
• An edge point (but this should be illegal in an LDraw part)
  

Hi Y'all,

Roland and I have been tossing around part numbers, looking at smoothing results. One thing that surprised me is that we needed a rather large 0.05 LDU error margin to 'snap' the segments of the 6x6 webbed dishes together. (The part is built via 8 sub-sections rotated 45 degrees.) 0.05 seems like a 'big' error because part authors can specify vertices to the 0.001.

Today I cracked open the part and took a look at the numbers, and the large error margin now makes sense.

The fundamental problem is absolute error amplification: if the error on a matrix ratio is 1/1000 (0.001 precision) then the actual error in the position of a rotated part is 0.001 * R where R is the approximate radius of rotation. In other words, a part that extends 60 LDU from its center will have an absolute error of 60/1000 (0.06) after rotation. (In fact, the real results are up to 3x worse because error accumulation is additive and we add the X, Y, and Z components of the matrix).

The actual numbers of the 6x6 dish show that the worst case error could have been much worse than this actual part. The bottom line is: 0.001 just isn't _that_ precise for a rotation component that's going to be used for a moderately big part. :-(

This matters because the rounding tolerance for welding rotated part segments for smoothing defines the smallest 3-d detail an author can make without having their triangles collapsed.

I realize that we have a gajillion parts and my going "we should do X" isn't going to actually change the library, but I wonder if this perspective should result in some kind of updated practices for authoring.

Naively, I would suggest two possible changes...

1. We could allow higher precision for matrices. For example, the part is using 0.707 for a 45-degree rotation. If I go into the text file and replace it with 0.707106781186548 I can cut my error margin down by 4x. (Clearly BrickSmith isn't using all of 0.707106781186548 - I just typed a big long number into the text file...BrickSmith is using 32-bit float.)

2. Authors could avoid angle rotations that aren't 90 degree increments. At 90 degrees, the rotation matrix can be _exactly_ represented - all rotational coefficients are 0, 1, or -1. :-) So the 'lock-up' of vertices around 90 degree rotations should be pretty good - certainly in the same ballpark as the geometry itself.

This would mean 'bigger' parts - 90 degree swathes of dish instead of 45 degrees, but the precision would be a lot better.

Cheers
Ben

Version 0.2.1 of LDForge is now out, fixing a bug causing LDForge to skip every 300th line of a given file. See release document and downloads here

• BFC red/green view, albeit present, isn't always right; it assumes everything is CCW and it fails to catch INVERTNEXT in subfiles. Also it has to draw everything twice so there's a bit of a performance hit.
  
• The camera suffers of gimbal lock. I'm going to try address this for future versions.
  

• List view ala MLCAD with multi-selection. One object per line, one line per object. Items not colored main or edge color (16/24) have their color reflected in the list view for identifying.
  
• Parse error recovery, if a line/object cannot be parsed properly it will be displayed as an errorneous object. This object can be selected and its contents edited and have it reparsed, so you can fix these errors within LDForge.
  
• 6 camera modes plus a free-angle one.
  
• Drawing mode that allows you to literally draw polygons and lines into the screen.
  
• Object hiding
  
• Select by color or type
  
• Quick edge-lining, takes any number of polygons and creates edgelines around them
  
• Ability to edit object's LDraw code directly
  
• Inlining, plus deep inlining which grinds down to polygons only
  
• Auto-coloring (sets color to the first found unused color), uncoloring (sets colors to main/edge color based on type)
  
• Coordinate rounding, inverting, coordinate replacing, flipping, quad splitting
  
• Screenshotting
  
• Vertex object, generic radial primitive object
  
• LDConfig.ldr parsing for color information
  
• Ability to launch Philo's utilities and automatically merge in output
  
• BFC red/green view (incomplete)
  
• Wireframe mode, axis drawing
  
• Image overlays for getting part data from pictures
  

Hi,
first thank you for your work.

After a lot of years of utilization, I developed the dream of to do as many as possible of official sets with the correct parts.
I know that this dream is quite impossible and that there is similar part that can be used, but I want to try to achieve my dream.

With humility, I ask if there is someone with spare time to create some of the following parts to help me (I have tried with a conversion from LDD but the result is visually good, but in reality I do not succeed to optimize the parts so to use them properly).

Best regard and still many thanks for the part made in this year
CL

The list:
6014b Wheel 11.5mm D. x 12mm, Hole Notched for Wheels Holder Pin (the actual version is very different)
58088 Wheel Cover 7 Spoke with Axle Hole (for the set 8145 Ferrari 599 GTB Fiorano)
60032 Window 1 x 2 x 2 Plane, Single Hole Top and Bottom for Glass (for set 4997 Transport Ferry)
30625 Hinge Panel 1 x 4 x 3 2/3 with 6 holes, 2 studs on top, Locking 2 Fingers
3245c Brick 1 x 2 x 2 with Inside Stud Holder