Sorting floating-point values, and NaN

[tn]

On Tuesday, 12 November 2013 at 21:09:34 UTC, Vladimir Panteleev 
wrote:
> On Tuesday, 12 November 2013 at 21:03:03 UTC, tn wrote:
>>> assert((equiv(a, b) && equiv(b, c)) <= equiv(a, c));
>>>
>>> ("<=" on Booleans is actually implication.)
>>
>> Shouldn't the implication be to the other direction? Then it 
>> becomes
>>
>> assert((equiv(a, b) && equiv(b, c)) => equiv(a, c));
>
> <= as in ≤ (less or equal), not ⇐ (reverse implies). <= can be 
> used on booleans as "implies" due to the way it treats booleans 
> as integers (true as 1, false as 0).

Thanks for the explanation. I thought it was pseudocode. That is 
a horribly confusing trick.