RE: [Re: RPGILE V4.3 Gotcha]

[Joel Fritz <JFritz@xxxxxxxxxxxxxxxx>]

Have you looked at the manual on precision for intermediate results for
eval?  The default precision will often produce downright odd results.  You
can use EXPROPTS(*RESDECPOS) in an H spec to make all  intermediate results
in expression evaluation have the same precision as the result field, or use
the "r" operation extender with eval.  The manual gives some interesting
cautions including that the result of the expression may only be a packed
field.  

The default precision rule is designed to guard against overflow.  Of
course, we always do that in programming and data base design anyway. <g>
I've fooled around with the precision stuff some, and have had pretty good
results with the "r".  Personally, I'd rather that "r" were the default, and
try to guard against overflow myself. 
> -----Original Message-----
> From: Jim Langston [mailto:jlangston@conexfreight.com]
> Sent: Monday, October 04, 1999 7:30 AM
> To: RPG400-L@midrange.com
> Subject: Re: [Re: RPGILE V4.3 Gotcha]
> 
> 
> Okay, let me just say it this way then, it is not safe to do ANY type
> of math on any number that contains a fraction.
> 
> If I have to worry about loosing accuracy after ONE significant digit,
> something is definitely wrong with the math engine.
> 
> Regards,
> 
> Jim Langston
> 
> boldt@ca.ibm.com wrote:
> 
> > someone wrote:
> > >Well, dang, that's no good at all!  This being the case 
> you can't do any
> > >type of floating point math with anything resembling any accuracy.
> > >
> > >Other languages use a fixed size for a floating point 
> value.  Why not
> > >just bump the size to 30,20 (or however your largest floating point
> > >value is) and forget about it?  Or are you saying your Mantissa can
> > >only store 30 significant digits?
> > >
> > >This now tells me to NEVER, NEVER, NEVER do any type of
> > >floating point math in RPG.
> >
> > Exsqueeze me?!?  Where in Carl's sample program did he use floating
> > point numeric variables?  He was using decimal numeric variables
> > exclusively.
> >
> > BTW, you don't really expect 30 digits of precision from floating
> > point variables, do you?  I agree with you - until you understand
> > the difference between the floating point and decimal numeric types,
> > please don't use floating point.  (Study Chapter 10 of the ILE RPG
> > Reference manual.)
> >
> > Cheers!  Hans
> >
> > Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com
> >
> > +---
> > | This is the RPG/400 Mailing List!
> > | To submit a new message, send your mail to RPG400-L@midrange.com.
> > | To subscribe to this list send email to RPG400-L-SUB@midrange.com.
> > | To unsubscribe from this list send email to 
> RPG400-L-UNSUB@midrange.com.
> > | Questions should be directed to the list owner/operator: 
> david@midrange.com
> > +---
> 
> +---
> | This is the RPG/400 Mailing List!
> | To submit a new message, send your mail to RPG400-L@midrange.com.
> | To subscribe to this list send email to RPG400-L-SUB@midrange.com.
> | To unsubscribe from this list send email to 
> RPG400-L-UNSUB@midrange.com.
> | Questions should be directed to the list owner/operator: 
> david@midrange.com
> +---
> 
+---
| This is the RPG/400 Mailing List!
| To submit a new message, send your mail to RPG400-L@midrange.com.
| To subscribe to this list send email to RPG400-L-SUB@midrange.com.
| To unsubscribe from this list send email to RPG400-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---