Maße aufeinander abstimmen

[Scholli]

Ich bräuchte eine Erklärung warum folgendes passiert:

Ich habe eine Schulterlänge A46-SH_vorne von 15,5166 cm

Ich habe eine Bogen gezogen um den Punkt A47 mit der Linie_A46_sh_vorne

Dann habe ich mit dem Tool Punkt schneidet Bogen und Achse den Punkt A117 angelegt.

Jetzt habe ich die Linie A117_A47 gezogen.

A46_SH_vorne ist 15,5166 cm

A117_A47 ist 15,5136 cm.

Die müssten doch eigentlich gleich lang sein.

Der Unterschied ist nicht relevant. Ich würde es aber gerne verstehen. Kann mir das jemand erklären?

[Douglas]

@Scholli

The difference isn’t relevant. But I’d like to understand it. Can someone explain it to me?

Rounding error when using real numbers.

Given arc with a 3.5cm radius:

Intersect at 180deg"

Intertsect at 145deg:

Intersect at 120:

Intersect at 90deg:

Notice how the line length at 90 and 180 is 3.5cm… where with the other 2 at different angles - using trig to calculate the length with rounding error each display a different length.

If we were to format the output with 2 decimal places they would all display 3.5 as the other 2 would get rounded up to 3.5. The use of real numbers and rounding are also most likely at the root of some edge case issues related to the SA. For example… If we were to take the above and try and test if Line_A_A1 = 3.5… it will fail in many cases. So we have to use fuzzymath and test if Line_A_A1 = 3.5 ± 0.01 where in all cases we can them say they’re all close enough to be equal.

And yes… .001cm or .01 mm is negligent for our purposes.

“To put that into perspective, .01mm is 5 to 7 times thinner than a standard sheet of standard copy paper (which is about 0.005 cm to 0.007 cm thick) and roughly the thickness of a red blood cell or a strand of spider silk.”

[Scholli]

@Douglas

Notice how the line length at 90 and 180 is 3.5cm… where with the other 2 at different angles - using trig to calculate the length with rounding error each display a different length.

If we were to format the output with 2 decimal places they would all display 3.5 as the other 2 would get rounded up to 3.5. The use of real numbers and rounding are also most likely at the root of some edge case issues related to the SA. For example… If we were to take the above and try and test if Line_A_A1 = 3.5… it will fail in many cases. So we have to use fuzzymath and test if Line_A_A1 = 3.5 ± 0.01 where in all cases we can them say they’re all close enough to be equal.

And yes… .001cm or .01 mm is negligent for our purposes.

“To put that into perspective, .01mm is 5 to 7 times thinner than a standard sheet of standard copy paper (which is about 0.005 cm to 0.007 cm thick) and roughly the thickness of a red blood cell or a strand of spider silk.”

Vielen Dank