Re: Strange result with exponents...

[Simon Coulter <shc@xxxxxxxxxxxxxxxxx>]

On Friday, February 20, 2004, at 11:32  AM, Scott Klement wrote:

> What's odd, to me, is that the exponent operator doesn't seem to work 
> the
> same way that the others do.  I can do addition or multiplication to a 
> 30p
> 0 field, and not run into problems when it gets above 17 digits -- but
> with exponents, the same field runs into problems.  That's not 
> intuitive,
> even if it's documented.

I agree the inconsistency is a problem. That's why I think always using 
float for exponentiation was a bad design choice.

> Furthermore, as you mentioned, it used to work.  And it does work
> everywhere else.   Maybe that's luck, but heck, it's the same number!
> Unless you're intimately familiar with how the microcode works, you're
> likely to expect it to work the same way.

That's why I suggested earlier that the PMR answer be rejected by 
pointing out the change in previous behaviour and that using a 20i0 
field gets the correct answer while a 30p0 field is rounded up to the 
wrong answer. Rejecting the PMR will have to be done by the person who 
opened it. Rob has already said he doesn't care enough to pursue it. So 
if you do care, open your own PMR, point out the problem in light of 
what has been discussed here, and see what happens.

Regards,
Simon Coulter.
--------------------------------------------------------------------
   FlyByNight Software         AS/400 Technical Specialists

   http://www.flybynight.com.au/
   Phone: +61 3 9419 0175   Mobile: +61 0411 091 400        /"\
   Fax:   +61 3 9419 0175                                   \ /
                                                             X
                 ASCII Ribbon campaign against HTML E-Mail  / \
--------------------------------------------------------------------