From: paulNO@... (Paul J. Sanchez)
Date: 2003-08-20T05:45:15+09:00
Subject: Re: NaN and Infinity

>>>>> "MS" == Mike Stok <mike@ratdog.stok.co.uk> writes:

    MS> In article <86r83r5v8l.fsf@daemon.argelfraster.org>,
    MS> Paul J. Sanchez <paulNO@SPAMargelfraster.org> wrote:
    >> 
    >> [Code Snipped]
    >>
    >> Note that by initializing min and max to +/- Infinity,
    >> respectively, the update stage for them doesn't need a special
    >> case for the first time.  If only the same were true of the
    >> mean and variance updates...
    >> 
    >> If anyone has any suggestions for tightening this up, by the
    >> way, they'd be appreciated!

    MS> Does it make sense that

    MS>   s = SimpleStats.new s.max => Infinity

Actually, it yields -Infinity.  s.min would be Infinity.

    MS> nil seems more reasonable for a "no value" result.

Since infinity is a bound, not a value, it can be interpreted as
saying "the smallest thing I've seen so far is unbounded above, the
biggest thing is unbounded below."

To tell you the truth, though, it's a translation from Java of an
assignment I give the intro programming students, and I switch the
initialization between nil and infinity in alternate quarters.  It
makes a great filter to catch those who are copying solutions from the
previous class.

    MS> If you let @variance, @mean, @max and @min start out life as
    MS> nil then

    MS> def newObs(x) @max = [@max, x].max @min = [@min, x].min @n +=
    MS> 1

    MS>   unless @mean # if @mean.nil?  @mean, @variance = x, 0.0 else
    MS> x -= @mean @mean += x / @n @variance = (@n - 2.0) / (@n - 1.0)
    MS> * @variance + x * x / @n end end

    MS> might be worth trying; Enumerable#max and Enumerable#min do
    MS> the "right thing" with nil.

    MS> Hope this helps

Thanks for the input!

--paul