Am 20.02.2012 um 17:23 schrieb Jordi Gutiérrez Hermoso:
> It sounds to me like we have chosen a canonical NaN value, NA, so we
> should be preserving it across operations?
If that means that NaN + NA will always result in NA then it might pose a
problem too, since - literally speaking - mapping Garbage + NotAvailable ->
NotAvailable might make sense for special applications, but will generally
obscure the fact that garbage has crept up in the calculation.
So maybe
( NaN, but not NA ) OP anything else -> NaN, and
( NA, but not NaN ) OP anything else -> agreed upon result
might give somewhat better semantics, but it's still not optimal.
Maybe Knuth has some hints on this? In the third edition of TAoCP he promised
to deal with these topics in the next edition, but I don't have anything newer
handy.
> I'm not sure what the correct solution is here, but perhas we should
> implement NA-skipping behaviour in the functions that Alois overwrote.
Hmmm ... these are pretty many. How would it affect performance on in-order
architectures like Intel's Atom if for example sum had to check every summand
for being NA?
Later,
Alex
--
Dr. Alexander Klein, Diplom-Mathematiker
Physiologisches Institut | TransMIT-Bereich
Raum 543 | für Mathematische Analysen
| und Feld-Simulationen
Aulweg 129 | Heinrich-Buff-Ring 44
35392 Giessen | 35392 Giessen
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
Octave-dev mailing list
Octave-dev@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/octave-dev
