2D vss 3D vss nD...

I found out yesterday that trying to make a variant with set of 3D points
(unwittingly) that should have been 2D and then use that variant to create a
LightweightPolyline object didn't have the outcome I quite expected.

"What the...?!" :-@

I tracked down the trouble in a 'Point/Vector addition' subroutine, which
automatically converted and then later output my 2D point list to a 3D point
list - which wasn't what I wanted to do in this case. I fixed it for this
particular problem, but I think it could be better.

So I was wondering what you people do to keep the dimensions correct - and
*especially* if you may be working with 'mixed dimension types', such as
adding a 2D xy vector to a set of 3D xyz points. If you use mapcar '+ for
this, you end up with a 2D pointlist when perhaps you really need those z
values.

When doing this type of point manipulation, have you found it best to
define/use a series of routines, like '1dVectorAdd', '2dVectorAdd',
'3dVectorAdd', etc. - OR - would it be better to create and use one routine
'VectorAdd', along with a desired output 'dimension' argument, perhaps like
(VectorAdd _vec _pts 2) - so that no matter *what* the input vector or point
list dimension happens to be, the output will be as indicated?

I'm leaning toward the second, since it seems to be a more 'all-encompasing'
solution (and fewer routines to keep track of), but I don't know if that
neccessarily results in better - faster/more readable/etc. - or worse code.

If this seems like a silly question or just a rehash of old topics, forgive
me, I've been gone awhile and feel a bit out of practice.

Mind you, I'm not necessarily looking for actual code, just some
opinions/reasons on what you consider to be good coding practice for this
sort of thing.

Many thanks,
David Kozina

 

Here is a thread with some info on handling the points.

http://discussion.autodesk.com/thread.jspa?messageID=2020156

I know you're not looking for code but there is a nice subr from TonyT. that handles them.

Bill

 

I keep my points as lists until I get to the point where I *need* the array.
At that point (groan!) I create the array for the subsequent statement and
then forget about the array.

--
R. Robert Bell, MCSE
www.AcadX.com


I found out yesterday that trying to make a variant with set of 3D points
(unwittingly) that should have been 2D and then use that variant to create a
LightweightPolyline object didn't have the outcome I quite expected.

"What the...?!" :-@

I tracked down the trouble in a 'Point/Vector addition' subroutine, which
automatically converted and then later output my 2D point list to a 3D point
list - which wasn't what I wanted to do in this case. I fixed it for this
particular problem, but I think it could be better.

So I was wondering what you people do to keep the dimensions correct - and
*especially* if you may be working with 'mixed dimension types', such as
adding a 2D xy vector to a set of 3D xyz points. If you use mapcar '+ for
this, you end up with a 2D pointlist when perhaps you really need those z
values.

When doing this type of point manipulation, have you found it best to
define/use a series of routines, like '1dVectorAdd', '2dVectorAdd',
'3dVectorAdd', etc. - OR - would it be better to create and use one routine
'VectorAdd', along with a desired output 'dimension' argument, perhaps like
(VectorAdd _vec _pts 2) - so that no matter *what* the input vector or point
list dimension happens to be, the output will be as indicated?

I'm leaning toward the second, since it seems to be a more 'all-encompasing'
solution (and fewer routines to keep track of), but I don't know if that
neccessarily results in better - faster/more readable/etc. - or worse code.

If this seems like a silly question or just a rehash of old topics, forgive
me, I've been gone awhile and feel a bit out of practice.

Mind you, I'm not necessarily looking for actual code, just some
opinions/reasons on what you consider to be good coding practice for this
sort of thing.

Many thanks,
David Kozina

 

Bill,

Thanks for the reference - much appreciated.

This is what I came up with for 2D<>3D point conversion, which should help
when mapcar'd to point lists and so forth BEFORE making other arithmetic
manipulations, like adding a 2d vector to a 3d point list, or before
creating an array function.

Does anyone/everyone have other suggestions for improvement?

; force a 2D or 3D point to a specific dimension
(defun ForceDPt
(_pt ; xy or xyz point
dims ; number of dimensions (2 or 3) for output
/
)
; begin
; ...determine dimension
(if (eq dims 3)
; then
; ...is point 3D?
(if (caddr _pt)
; then return it as is
_pt
; else add z=0 value
(reverse (cons 0 (reverse _pt)))
);_end if
; else
; ...is point 3D?
(if (caddr _pt)
; then return xy portion
(list (car _pt) (cadr _pt))
; else return it as is
_pt
);_end if
);_end if
; end
);_end defun