Subject:
|
Re: New Primitive Class
|
Newsgroups:
|
lugnet.cad.dat.parts
|
Date:
|
Sat, 22 Mar 2003 21:18:51 GMT
|
Viewed:
|
918 times
|
| |
|
|
In lugnet.cad.dat.parts, Travis Cobbs writes:
> In lugnet.cad.dat.parts, John Riley writes:
> > My current 33089 (in PT) is a rather bulky file (55KB), so to streamline it,
> > I decided to make a few new primitives. One of these was a torus with major
> > radius 1 and minor radius 1.
> >
> > Under the current naming scheme, this primitive would be named:
> > t04o10000.dat
> > which is not in 8.3 format.
>
> You might be surprised at how close t04o9999.dat would come. I can't see
> that it's any worse than the current t04o3333.dat file. Actually, I think
> I'm going to modify LDView to recognize nnnn in the minor fraction (where n
> is 0-9) to be exactly equal to n / 9. In all likelihood, anyone using nnnn
> really means n / 9 anyway. That would produce exactly 1 for the 9999 case.
> If we modify the spec to say this, you can have your unit torus without a
> new naming convention. Programs that perform primitive substitution can
> have an exact match, and the data in the file itself can be exact.
I'm not terribly surprised. This is a great solution, as then, if for some
reason in the future, inverse ratio tori are needed, they can use the same
convention, just a different starting letter and inverting the ratio as
appropriate in the name (i.e. i04oRRRR.dat)
> I can't imagine an inverse-ratio inner torus ever being used. It would
> overlap itself. So the o is redundant.
I don't see it being used either; at least, not if it isn't cut off at the
+y intercept so that the file remains in the +x+y+z quadrant. The cut off
inner inverse ratio torus makes a curved surface to a point. The surface,
viewed from below, is an point indentation. From above, it's a needle.
Neither surface is common in LEGO (unless someone really wanted to model the
injection point, which tends to be needle like until handled). But for
rather fine detail work, it could be useful. Certainly not a common shape
though.
> I don't think this is a good idea. You're sacrificing one of the precious
> 8.3 digits as an indicator that this is a unit inverse torus. 0 and 1 are
> the only valid choices, and if it's 1, all three of the other digits have to
> be 0 (otherwise it should just use the current naming convention). While I
> realize you're trying to come up with something that solves your current
> problem, I don't think this is a good solution.
I didn't either, that's why I asked. The 9999 solution wasn't obvious to me
at first, but I'm glad you pointed it out.
The best solution, IMO, would be
tFFxRRrr
where RR is the major radius and rr is the minor radius, after being scaled
to the lowest whole number combination.
Thus, the 1 x 0.0625 torus is now a 16 x 1 torus.
This naming convention allows for the inverse ratios. However, this
deviates strongly from most primitive constructions, where 1 of the
dimensional parameters is 1LDU.
All of this is probably moot now, since I'll be using the t04o9999 name for
the unit ratio torus, and nobody will probably use an inverse ratio torus
for a long time, if ever. It's just not a common shape.
Thanks Travis, for the insights on 9999.
John
|
|
Message is in Reply To:
 | | Re: New Primitive Class
|
| (...) You might be surprised at how close t04o9999.dat would come. I can't see that it's any worse than the current t04o3333.dat file. Actually, I think I'm going to modify LDView to recognize nnnn in the minor fraction (where n is 0-9) to be (...) (21 years ago, 22-Mar-03, to lugnet.cad.dat.parts)
|
7 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
