hello guys, let me once again say i thank for your help!
and please stay with me a little further …
- ‘rounding’ - i didn’t want to round 0,5 any further, i want to check if i have operands as such and then be sure that a rounding of the result doesn’t harm,
- ‘the computing of A2’ is ‘not so bad’ (one might discuss the rounding for x5 in A1), bad is the result of the calculation in A4, ‘weitz’ can do better, ‘rawsubtract’ can do better (see A5), calc can compute 16 digit resolution outside the range -/+ 22 in A1, precision is not! lost in ‘IEEE’ or ‘doubles’ but in calcs implementation of subtraction,
- ‘It is a matter of primality’ - won’t say it’s trivial but i know about such, about 80% of 1-decimals-digit figures lack enough '5’es in the enumerator to cancel out the '5’es in the enumerator, and already 99,84% of 4-decimals-digits, they stay ‘odd fractions’ for binary representation and become ‘endless’ …
… continued in next comment …