Comparisons of incompatible types

[Steven D'Aprano]

On Wed, 08 Dec 2010 13:01:28 -0800, John Nagle wrote:

> On 12/7/2010 3:59 PM, Mark Wooding wrote:
>>> Exactly one of
>>> >
>>> >  	a>  b
>>> >  	a = b
>>> >  	a<  b
>>> >
>>> >  is true, or an type exception must be raised.
>> This will get the numerical people screaming.  Non-signalling NaNs are
>> useful, and they don't obey these axioms.
> 
>     As a sometime numerical person, I've been screaming at this from
> the other side.   The problem with comparing non-signalling NaNs is that
> eventually, the program has to make a control flow decision, and it may
> not make it correctly.

Then use signalling NANs. Nobody is suggesting that quiet NANs should be 
compulsory, or are the solution for all problems. But they're a solution 
for some problems, which is why people use them.

[...]
>     I personally think that comparing NaN with numbers or other
> NaNs should raise an exception.  There's no valid result for such
> comparisons.

If NAN and 1 are unordered, then NAN is not less or equal to 1, nor is it 
larger than 1. Hence both NAN <= 1 and NAN >= 1 are false. The problem 
only comes when the caller mistakenly thinks that floats are real 
numbers, and tries to reason about floats like they would reason about 
real numbers.



-- 
Steven