You missed the point, V... It should always give back exactly 1... no more, no less.
Anything else means it is rounding incorrectly.
mnem
I don't nuderstand how you an get an exact 1 if using floating point numbers, which are by their very nature... only an approximation ?
If a calculator gives spot " 1 " when using floating point number, I guess all it means is it has enough zeros internally to round it to 1 on the display, which shows less digits than the internal representation used internally for the computation.
IOW your " 1 " is not a one, it's just more zeros than other calculators so it's rounded to one when displayed...
Plus, from what I understand not every calculator will give the same result for cos / sin / tan depending what algorithm / approximation they use to calculate it ? A trade-off between computing time and available power on the device ?
Isn't that why accounting people use "BCD" calculator so as to get a true result, 100% reliable, to avoid rounding errors inherent to floating point calculation ?
Seems like a rabbit hole to me.. computing nuttery...
... but it's an interesting subject !