n-fcyls2 / n-fcyls Primitives

[Gerald Lasser]

Till now I used the cyls and cyls2 primitives mainly in their available quarter editions, respectively in their 1/8th size

Now I would need a sub-division that does not start at the smallest part (as it is with the 3-16cyls) but one that starts from the tallest one.

- How should those be called? Basically it would be a 1-8cyl...

- And would LDView do Primitive Substitution for those as well?
   

[Gerald Lasser]

I am bumping this one, as I see the discussion on the tndis prims and their treatment by LDView.

What do you think of those "inverted" cyls/cyls2 primitives?

I cannot recall the part anymore where I wanted to use them...

[N. W. Perry]

I am bumping this one, as I see the discussion on the tndis prims and their treatment by LDView.

What do you think of those "inverted" cyls/cyls2 primitives?

I cannot recall the part anymore where I wanted to use them...

If they're needed, even occasionally, I say go for it. Otherwise you have nice smooth prim-subbed curves next to regular quadrangular faces, no?

[Philippe Hurbain]

Gerald, what's your use case for this? I needed a similar one to create the stud notch in a slope brick. But for me it must be done like on attached image. (actually I need the n-fcyls2 version)

[N. W. Perry]

It seems to me that if you could create primitives for each cylindrical face (i.e., quads or triangles, but understood to represent a curved face for substitution purposes), you could allow for any use case. But you would have to append the name to include which sector of the cylinder it is, and whether it's the cyls or the cyls2 version.

Then you could also build new prims by combining these faces, and naming could reflect the range of sectors that's included. So like 1-8cyls2-3.dat would be 1/8th of a cylinder, made of sectors 2 and 3. Or something like that.

[Gerald Lasser]

Gerald, what's your use case for this? I needed a similar one to create the stud notch in a slope brick. But for me it must be done like on attached image. (actually I need the  n-fcyls2 version)

I do not have a current use case, I think I wanted it for the train coupling hooks.

But I agree, your solution is more flexible, just att the corresponding cyli section and we are good.

[Philippe Hurbain]

More thoughts on this...
   
The proposed new primitives allow to complete a cylinder segment cut by a sloped plane going through a pair of vertices. While this may seem restrictive, this situation often occurs, eg. when a stud meets a cut plane.

• As far as I can tell, only prims covering up to 1/4 cylinder are needed.
  
• But it could make sense for cyls2 to cover up to 1/2 cylinder to simplify usage and limit number of vertices    
  
• 1-4fcyls2 = 1-4cyls2 and 1-4fcyls = 1-4cyls
  

[Philippe Hurbain]

It seems to me that if you could create primitives for each cylindrical face (i.e., quads or triangles, but understood to represent a curved face for substitution purposes), you could allow for any use case. But you would have to append the name to include which sector of the cylinder it is, and whether it's the cyls or the cyls2 version.

Then you could also build new prims by combining these faces, and naming could reflect the range of sectors that's included. So like 1-8cyls2-3.dat would be 1/8th of a cylinder, made of sectors 2 and 3. Or something like that.

This would indeed be the ultimate flexibility but so far I failed to imagine a real life use case?

[N. W. Perry]

This would indeed be the ultimate flexibility but so far I failed to imagine a real life use case?

And I guess that's really the point—if you know what the naming scheme is to be, you can create and combine the primitives as needed, when unforeseen uses arise. (Such as the use case that started this thread: you could create a sector 3-4 primitive without having to propose something new.)

For even more flexibility, you could imagine a type of metadata where there is an underlying curved geometry, such as a cylinder, in a part file. You could then map any tri or quad to this invisible object, so that the curvature is generated by rendering programs. This would be useful for parts with lots of interrupted circles and cylinders (like 4208), or patterns on curved surfaces like the minifig head.

But of course, that goes beyond primitives and into what I imagine as an "LDraw 2.0" specification. :-)

[Travis Cobbs]

I think you confused the "n-f" fraction syntax (used on the primref and in the post title) for the proposed new primitive name (with f as a prefix). If this is not the case, can you tell me what the f is supposed to stand for?

If it is the case, then I would propose either an r prefix (for reversed) or an i prefix (for inverted). We could also use a t prefix (for truncated, like used on tdis), because the proposed top is truncated in a similar fashion to tdis. However, since we start from the opposite end of the existing cyls/cyls2 primitives, perhaps t would just be confusing.

[Travis Cobbs]

I am bumping this one, as I see the discussion on the tndis prims and their treatment by LDView.

Similar to tdis, if parts get posted to the tracker using these new primitives, I will endeavor to support them in LDView. The conversation hashing out how they should work will need to be finalized first, and someone will need to let me know when they get added to the tracker and used in unofficial parts.

(As an aside, I think tndis would have been a better name for the tdis primitives, but that's apparently not what got chosen.)

[Philippe Hurbain]

If it is the case, then I would propose either an r prefix (for reversed) or an i prefix (for inverted). We could also use a t prefix (for truncated, like used on tdis), because the proposed top is truncated in a similar fashion to tdis. However, since we start from the opposite end of the existing cyls/cyls2 primitives, perhaps t would just be confusing.

Clearly yes, I mixed up things with the prefix! tcyls and tcyls2 suit me.

Similar to tdis, if parts get posted to the tracker using these new primitives, I will endeavor to support them in LDView. The conversation hashing out how they should work will need to be finalized first, and someone will need to let me know when they get added to the tracker and used in unofficial parts.

I already have a part on PT that would benefit of a 1-8tcyls2 

(As an aside, I think tndis would have been a better name for the tdis primitives, but that's apparently not what got chosen.)

So I think, but the new prims were already on PT. If we all agree it's not too late to change that!

[Travis Cobbs]

So I think, but the new prims were already on PT. If we all agree it's not too late to change that!

If they do get changed, I'd really like it to be before I release the final LDView 4.4.1.

[Gerald Lasser]

Clearly yes, I mixed up things with the prefix! tcyls and tcyls2 suit me.
I already have a part on PT that would benefit of a 1-8tcyls2 
So I think, but the new prims were already on PT. If we all agree it's not too late to change that!

I agree, the tndis is a better acronym

[Magnus Forsberg]

I agree, the tndis is a better acronym

Me too.

[Travis Cobbs]

Me too.

If they get renamed, I think it needs to happen soon. Some of them are already certified, along with parts that use them.