[cfp-interest 3664] Re: ISO 60559 6.3 and C2Y F3.3#7

[Damian McGuckin]

On Thu, 16 Oct 2025, Fred J. Tydeman wrote:

>> The former uses "not interpret".
>>
>> So should "generally" be deleted in C2Y or is generally meant to handle 
>> those cases where a NaN is neither input nor result?
>
> I believe that copysign is one function that cares about the sign of the 
> input NaN.  And also an output NaN.

IEEE 754 uses the word interpret to mean that the sign is involved in any
decision. So a copy operation, i.e. negate, absolute value, copysign or
bit-for-bit copy is not aninterpreting.

The difference in language between the two is confusing.

Thanks - Damian