Dan Sugalski wrote: > At 09:07 AM 10/8/2001 +1000, Damian Conway wrote: > >I see your (and others') point here, but I view NaN as a *marker* indicating > >a non-numeric result. Markers should always compare equal to themselves. > >(Frankly, *everything* should compare equal to itself -- which is where > >IEEE 754 goes horribly wrong.) > > FWIW, I've always viewed NaN as equivalent to SQL's NULL, and treated it as > such. (Of course, proper support for either would require me to build in > three-value logic to Parrot, and I'm not sure I'm willing to go quite that > far... :)
Well, I was hoping that three-value logic will be supported by Parrot (at least I hope it will be still supported by Perl, but without support in Parrot it surely won't be so easy and efficient). It's a superset of Boolean logic so it doesn't collide with any strict Boolean arithmetic. Maybe it would be possible to define what kind of logic the user wants? Boolean, three-value, fuzzy logic - that would be very nice. In a similar way you define meaning of other things in Parrot to match different languages. What do you think about that? - RaFaL Pocztarski, [EMAIL PROTECTED]
