Appears he got it from the link supplied by irneb, from Alanjt's post, (I assume that's the root of your question)
 however he removed the header indicating the author....   

Apparently reading many previous post regarding this subject (removing authors information) is beyond his comprehension level.

Where did you get the code?
This should work.

(defun c:EW  (/ *error* lst f)
 ;; Entity Swap
 (defun *error*  (msg)
   (and f *AcadDoc* (vla-EndUndoMark *AcadDoc*))
   (and msg
(not (wcmatch (strcase msg) "*BREAK*,*CANCEL*,*QUIT*"))
(princ (strcat "\nError: " msg))))
(princ "\nSelect two entities to swap: ")
(if (and (setq ss (ssget))
         (or (= (sslength ss) 2) (print "Must be two objects only"))
         (setq lst (vl-remove-if 'listp (mapcar 'cadr (ssnamex ss)))))
   ((lambda (pts)
      (setq f
     (not (vla-StartUndoMark
    (cond (*AcadDoc*)
  ((setq *AcadDoc* (vla-get-activedocument
     (vlax-get-acad-object))))))))
      (mapcar (function (lambda (o a b)
  (vl-catch-all-apply
    (function vla-move)
    (list o a b))))
      (mapcar (function vlax-ename->vla-object) lst)
      pts
      (reverse pts)))
     (mapcar
(function
 (lambda (x) (vlax-3d-point (cdr (assoc 10 (entget x))))))
lst)))
 (*error* nil)
 (princ))

chlh_jd,

The previous version of the program had function like that.

But in an effort to squeeze every bit of speed I`ve put everything inline
thus gaining a tiny bit of speed by not calling a function.  On a 1000 points
triangulation this is about 50000 calls.

Second suggestion make sense to me.

Setting the current layer also saves me as many  calls to cons  as there are vertices in the contours.

Thanks for the helping hands.

ymg

Topographer,

Sorry for late reply.

In your drawing there is an incorrect triangle, so the sorting routine for the triangle list
crash on it.

I will check the get_tin routine to see if there is a bug there, or add something to prevent this.

Meanwhile use tin on the point and you will get the contour.

The prof routine is for making a profile from a set of 3dface, or a triangulation or yet again from
a 3dpolyline.  However it is not finished, so not ready for production.

OCD is a routine i got from AUGI forum, will be use to trim triangulation.  This is also not ready

ymg

Obvious disclosure: I work for Autodesk. I have a vested interest in seeing this product in your toolbox.

Sticker price is right around $2K in the US, $2500 if you put it on the network. Your typical engineer bills out at $125/hour. If you spend 20 hours a YEAR working on swept-path, parking lots, or reoundabouts, the math is pretty simple. To me, just the parking lot layouts ALONE would have made my money back (stupid offset, array, reoffset, erase, offset, array, reerase, reoffset, erase it all, start over process...) Software is cheap compared to time.

Not to mention, there is a promotion right now for Map 3D and Civil 3D customers. New licenses of Vehicle Tracking 2014 are available at 50% off SRP for Autodesk Subscription customers who already hold a license of AutoCAD 2014, AutoCAD Map 3D 2014, AutoCAD Civil 3D 2014, or Infrastructure Design Suite 2014. This offer ends on April 23, 2014. That's from this page: http://www.autodesk.com/products/vehicle-tracking/overview.

You are right about software is cheap compared to time. And I would not discuss the quality of AVT, I have worked with it and the Vehicle Tracking is not worse or better than other software. The Parking Lit Designer is good and if you do such work a lot, it is a real timesaver. If you don't do this much, a few dynamic blocks will help too.

But your calculation is wrong. You assume the engineer has a 100% profit if you translate $ 2500 into 20 hours.

If he has only $ 5 profit ph, then he need to work 500 hours with the software before he can start making profit again. If he also needs a day of training, then he needs to work almost 700 hourse before he can start making profit. It might be less if the software is such a timesaver that he can send bills for the time it usually cost but he actually worked less with the new software. But in the real world you must be cheaper than a competitor, and you can't keep sending bills for the time it cost in the past, so then there is no extra profit.

And that is why I think the software is quite expensive.