|
In addition to the unit capped cylinder I suggested here
http://news.lugnet.com/cad/dat/parts/?n=6310 Id also like to see some
truncated disc and ndis primitives, like the red ones shown below (the truncated
disc can also be seen as a half chrd primitive).
This would help building many rounded shapes more and more common in LEGO parts.
They would mostly be useful in 48 version to have enough smoothness.
What do you think? how should we name these primitives?
Philo
|
|
|
In lugnet.cad, Philippe Hurbain wrote:
|
In addition to the unit capped cylinder I suggested here
http://news.lugnet.com/cad/dat/parts/?n=6310 Id also like to see some
truncated disc and ndis primitives, like the red ones shown below (the
truncated disc can also be seen as a half chrd primitive).
This would help building many rounded shapes more and more common in LEGO
parts. They would mostly be useful in 48 version to have enough smoothness.
What do you think? how should we name these primitives?
|
First of all, I like the idea of a unit capped cylinder. The stud could then be
changed to a reference to the new primitive. We could use x-ycylc as the
filenames and call the primitive Cylinder Capped. This would be consistent with
the existing 1-4cyls (Cylinder Sloped) primitive.
As for the truncated disc and ndis primitives names, I would recommend x-ytdisc
and x-ytndisc. The tndisc primitives would then not be 8.3 compatible, so
thats only OK if the parts library is now allowed to go by the 1.0 version of
the LDraw file format spec. Additionally, a 1-16tdisc primitive would also not
be 8.3 compatible. Having said that, whats the point of having long filenames
explicitly permitted in the spec if were not going to allow them in the
library?
(Since I dont author parts, I wont comment on their usefulness. Ill take
your word for it that they will be useful.)
--Travis
|
|
|
In lugnet.cad, Travis Cobbs wrote:
|
First of all, I like the idea of a unit capped cylinder. The stud could then
be changed to a reference to the new primitive. We could use x-ycylc as the
filenames and call the primitive Cylinder Capped. This would be consistent
with the existing 1-4cyls (Cylinder Sloped) primitive.
|
OK, so Ill make this primitive, but which way should it face? Should it face
up, like the stud primitive, or should it face down, like the box primitives?
--Travis
|
|
|
In lugnet.cad, Travis Cobbs wrote:
|
OK, so Ill make this primitive, but which way should it face? Should it
face up, like the stud primitive, or should it face down, like the box
primitives?
|
Oh, and should the cap be at Y=0, or should the base be at Y=0? (For reference,
with the stud primitive, the base is at Y=0.)
--Travis
|
|
|
In lugnet.cad, Philippe Hurbain wrote:
|
In addition to the unit capped cylinder I suggested here
http://news.lugnet.com/cad/dat/parts/?n=6310 Id also like to see some
truncated disc and ndis primitives, like the red ones shown below (the
truncated disc can also be seen as a half chrd primitive).
This would help building many rounded shapes more and more common in LEGO
parts. They would mostly be useful in 48 version to have enough smoothness.
What do you think? how should we name these primitives?
Philo
|
Please let me have some example where they are useful. I found during authoring
parts that, just for rounded parts, you would need lots of such truncated
primitives, but they would need to be truncated anywhere and not only the way
this new primitives would be. So I think this can result in much, much more
primitives and I think this is not useful. But I am open for discussion.
cu
mikeheide
|
|
|
In lugnet.cad, Travis Cobbs wrote:
|
In lugnet.cad, Philippe Hurbain wrote:
First of all, I like the idea of a unit capped cylinder. The stud could then
be changed to a reference to the new primitive. We could use x-ycylc as the
filenames and call the primitive Cylinder Capped. This would be consistent
with the existing 1-4cyls (Cylinder Sloped) primitive.
|
One concern about capped cylinders is that they would lead more multiples of
cylinders - single-capped cylinders, double-capped cylinders, 1/4 capped
cylinders, 2/4 capped cylinders, Hi-res 3/8 double-capped cylinders, etc. etc.
And every single capped cylinder will actually be a compound primitive -- i.e.,
a composition of one or more cylinder, disc, and edge primitives.
While capped cylinders would be useful, I dont think they would be worth the
overhead.
Steve
|
|
|
> Please let me have some example where they are useful. I found during
> authoring parts that, just for rounded parts, you would need lots of such
> truncated primitives, but they would need to be truncated anywhere and not
> only the way this new primitives would be. So I think this can result in
> much, much more primitives and I think this is not useful. But I am open for
> discussion.
A few examples where I would use the 48\1-8tdisc (to use Travis naming
suggestion):
58135s01 (rounded sides), 464s01, 54096, 61068
For the 48\1-8tndis (less useful I agree), an example would be 3455.
In many case you can stretch a bit the primitive to match the shape you want, so
only a few usual fractions are needed. And they match the cylinders and edges
primitives so they would make several files more compact and elegant.
Philo
|
|
|
In lugnet.cad, Travis Cobbs wrote:
|
In lugnet.cad, Travis Cobbs wrote:
|
OK, so Ill make this primitive, but which way should it face? Should it
face up, like the stud primitive, or should it face down, like the box
primitives?
|
Oh, and should the cap be at Y=0, or should the base be at Y=0? (For
reference, with the stud primitive, the base is at Y=0.)
--Travis
|
I would use the cylinder as model, with the cap at Y=0 on top and an edge at
bottom.
But I somehow agree with Steve Bliss that it may be overkill if we decline
capped cylinder in all angles...
Philo
|
|
|
In lugnet.cad, Philippe Hurbain wrote:
> > Please let me have some example where they are useful. I found during
> > authoring parts that, just for rounded parts, you would need lots of such
> > truncated primitives, but they would need to be truncated anywhere and not
> > only the way this new primitives would be. So I think this can result in
> > much, much more primitives and I think this is not useful. But I am open for
> > discussion.
>
> A few examples where I would use the 48\1-8tdisc (to use Travis naming
> suggestion):
> 58135s01 (rounded sides), 464s01, 54096, 61068
>
> For the 48\1-8tndis (less useful I agree), an example would be 3455.
>
> In many case you can stretch a bit the primitive to match the shape you want, so
> only a few usual fractions are needed. And they match the cylinders and edges
> primitives so they would make several files more compact and elegant.
>
> Philo
I think I can follow your thoughts. Have you tried that (48/1-8tdisc) already in
some items with different curvatures? If yes and you have found out that this
can also be used without dramatically mathematics (I prefer to use triangles so
I can see that the points are correct) then I will vote on that part.
If we decide yes for this, then I think we also should create the opposite part.
cu
mikeheide
|
|
|
In lugnet.cad, Steve Bliss wrote:
|
In lugnet.cad, Travis Cobbs wrote:
|
In lugnet.cad, Philippe Hurbain wrote:
First of all, I like the idea of a unit capped cylinder. The stud could
then be changed to a reference to the new primitive. We could use x-ycylc
as the filenames and call the primitive Cylinder Capped. This would be
consistent with the existing 1-4cyls (Cylinder Sloped) primitive.
|
One concern about capped cylinders is that they would lead more multiples of
cylinders - single-capped cylinders, double-capped cylinders, 1/4 capped
cylinders, 2/4 capped cylinders, Hi-res 3/8 double-capped cylinders, etc.
etc. And every single capped cylinder will actually be a compound primitive
-- i.e., a composition of one or more cylinder, disc, and edge primitives.
While capped cylinders would be useful, I dont think they would be worth the
overhead.
Steve
|
I agree with you in all points.
But if I see that also you are thinking that capped cylinders are useful maybe
there is somebody out who can write a prog for that. I have tried something
similar in the past and was not very happy with the handling. You can found that
on my webside (under LDraw / Programme / LDraw - Tools) as excel file
cylindercut.xls and circlecut.xls.
cu
mikeheide
|
|
|
> I think I can follow your thoughts. Have you tried that (48/1-8tdisc) already in
> some items with different curvatures? If yes and you have found out that this
> can also be used without dramatically mathematics (I prefer to use triangles so
> I can see that the points are correct) then I will vote on that part.
>
> If we decide yes for this, then I think we also should create the opposite part.
Mathematics to use 1-8tdisk is easy, divide the width by 0.707 and height by
0.293 (1-(sqrt(2)/2) to get size of the tdisk. Move the center accordingly.
Yes I have used that trick several times, using regular disc/ndis primitives,
inlining them and editing out the "spike". I'll create them next time I need
them...
Philo
|
|
|
In lugnet.cad, Andrew Westrate wrote:
I wasnt aware of the tang primitive, and I had forgotten about the chrd
primitive. Going from the chrd primitive to the proposed truncated disc
requires the addition of one triangle, and the total number of triangles after
adding that one is still the minimum needed to code the primitive. So Id
definitely say that the chrd primitive is the way to go here.
The not-truncated-disc is harder to construct from the tang primitive. For the
1/8th case, it requires three extra triangles, and this yields 5 total
triangles, vs. the minimum of 2 that are required. Im not sure theres any
point to versions other than 1/8th. Having said all that, the tang primitive is
more versatile, since the geometry in it is so slim that it will fit any almost
any conceivable place where theres stuff outside a circular hole.
--Travis
|
|
|
> There are already circle chord and tangent primitives on the parts tracker.
> You would just need to add a couple of triangles to get the ones you're
> proposing.
>
> See:
>
> <http://www.ldraw.org/cgi-bin/ptdetail.cgi?f=p/48/1-4chrd.dat>
>
> <http://www.ldraw.org/cgi-bin/ptdetail.cgi?f=p/1-4tang.dat>
Andrew you are completely right. I was kind of obsessed by the missing triangle
so I didn't thought well... I'm a little bit less convinced by the tang
primitive that need several adaptation triangles (is it really better than to
match directly the triangles to the circle primitives?).
Philo
|
|
|