You build a tract on your plat one line at a time. Sometimes you do offsets, sometimes you do perpendiculars, sometimes you extend, etc. On a recent job I had a tract with all sides parallel and perpendicular. It would have been a rectangle except for another rectangle being taken out of one corner. Very simple area to calculate by hand. But, being lazy, I did a PEDIT and checked for area. The resultant area was something like 534,000.06 square feet.
That wouldn't be a problem except that all distances were integer amounts. So, where did the 0.06 part come from. Each line checked out to be like 514.00, as they were intended to be. Every bearing was correct to the nearest second.
Apparently, during my construction, I must have created some minor error somewhere with one or more of the lines. A quick check of each line did not provide the answer. Something, somewhere was off a tiny bit, but, less than 0.01 or less than 1 second, yet I could not find what it was.
This, then, leads one to question the validity of other, much more complex shapes where a PEDIT and AREA process has been used.
Anyone else ever catch an error such as this?
One easy way to have such a problem is to do something like this. You sit on a hill at the side of a road and shoot each direction, once hitting a section corner and once hitting a quarter corner. Later, you enter that information and then LIST the section line. It says it is 1 minute 30 seconds from being north. You rotate everything by that deviation to get your section line running north. Then as you build your plat you add lines that you input as going east, south, west, north, etc. Everything seems to be square. But it isn't. Why, you ask? Because when you did the rotate command it was based on a value that you had read only to the preset precision of your AUTOCAD settings. It was actually 1 minute 30.00998877 instead of exactly 1 minute 30 seconds. If you used offsets and perpendiculars, everything will be fine. But, if you drew lines to specific bearings you will beslightly off.
have you been hanging around with Kent lately?
> You build a tract on your plat one line at a time. Sometimes you do offsets, sometimes you do perpendiculars, sometimes you extend, etc. On a recent job I had a tract with all sides parallel and perpendicular. It would have been a rectangle except for another rectangle being taken out of one corner. Very simple area to calculate by hand. But, being lazy, I did a PEDIT and checked for area. The resultant area was something like 534,000.06 square feet.
>
> That wouldn't be a problem except that all distances were integer amounts. So, where did the 0.06 part come from. Each line checked out to be like 514.00, as they were intended to be. Every bearing was correct to the nearest second.
>
> Apparently, during my construction, I must have created some minor error somewhere with one or more of the lines. A quick check of each line did not provide the answer. Something, somewhere was off a tiny bit, but, less than 0.01 or less than 1 second, yet I could not find what it was.
>
> This, then, leads one to question the validity of other, much more complex shapes where a PEDIT and AREA process has been used.
>
> Anyone else ever catch an error such as this?
>
>
Are all the Z values of the enpoints the same?
Yes. 2-D only.
That is why I calculate points at each corner with coordinates. When I calculate area, I always to Inverse Area and use the points. After I am done with the points, I let the Inverse Area command draw my polyline. I do my plat from there.
Sounds like a significant digits issue
