AngouriMath
Arcsin(System.Numerics.Complex)
Method (no overloads)
Summary
The arcsine the library evaluates to.
Remarks
Asin(System.Numerics.Complex) agrees with
Evaled everywhere except on the two branch cuts, the real
arguments outside [-1, 1], where it takes the upper side and the library takes the
lower one: the library reads arcsin(3) as pi/2 - 1.7627i, and
Complex as pi/2 + 1.7627i.
points spanning both cuts, both sides of each and the interval between them, this
agrees with Evaled at every one.
cut -- the only place anything tested it -- and wrong everywhere else. It made
compiled arcsine the conjugate of the arcsine, sosin(arcsin(z)) came back
as the conjugate of z andarcsin(z) + arccos(z) was not pi/2 off the real
axis. Newton's method felt it as a divergence away from a root it was started on:
see https://github.com/asc-community/AngouriMath/issues/115.
Evaled everywhere except on the two branch cuts, the real
arguments outside [-1, 1], where it takes the upper side and the library takes the
lower one: the library reads arcsin(3) as pi/2 - 1.7627i, and
Complex as pi/2 + 1.7627i.
points spanning both cuts, both sides of each and the interval between them, this
agrees with Evaled at every one.
cut -- the only place anything tested it -- and wrong everywhere else. It made
compiled arcsine the conjugate of the arcsine, so
as the conjugate of z and
axis. Newton's method felt it as a divergence away from a root it was started on:
see https://github.com/asc-community/AngouriMath/issues/115.
Angouri © 2019-2023 · Project's repo · Site's repo · Octicons · Transparency · 1953 pages online