XNumbers can help with demonstrating the approximation of Pi by succesively
computing the perimeters of regular inscribed polygons with 2^n sides. But
for anything requiring fractional powers or fractional roots, you'll need
something else, because XNumbers truncates fractional powers and fractional
roots to integers. Compare XNumbers to Excel:

Excel:
=1.98^1.98
(Returns 3.86720395054666)
=1.98^(1/1.98)
(Returns 1.41198766954688)

XNumbers:
=xpow(1.98,1.98)
(Returns 1.98)
=xroot(1.98,1.98)
(Returns 1.98)

My add-in xlPrecision 2.0 returns fractional powers and fractional roots
with up to 32,767 significant digits of precision:

=xlpPOWER(1.98,1.98)
(Returns
3.867203950546664475197024334694561094821782762326 45703981220472
990158197209281613382250690200215 etc., up to 32,767 digits)
=xlpROOT(1.98,1.98)
(Returns
1.411987669546878795740148157203790542076237995834 60566043557515
9344161656315787429344740444142505 etc., up to 32,767 digits)


You can download the free edition of xlPrecision 2.0 here, and use it as
long as you wish:

http://PrecisionCalc.com


Thanks,

Greg Lovern
http://PrecisionCalc.com
Get Your Numbers Right




"Niek Otten" wrote in message
...

Excel's precision is 15 significant digits.
If you need more, you might take a look at the XNUMBERS add-in:

http://digilander.libero.it/foxes/MultiPrecision.htm


"Richard Lionheart" wrote in message
...

I demonstrating the approximation of Pi by successively computing the
perimeters of regular inscribed polygons with 2^n sides. Excel's
precision was overwhelmed at about 4000 sides.

Is there any way to get "double precision"? If worse comes to worse,
I'll use something like BigDecimal in Ruby or Perl to get greater
precision.