Revision history [back]
 | 1 | initial version |
I can not reproduce your computation (the definition of s seems uncomplete), but you should notice that -8e-202 is extremly small (in particular, the first 200 digits of s and N(sin(50/180*pi)-s,digits=200) are equal).
If you want a result with an explicit bound on the error, you should probably use RealBallField or RealIntervalField.
