Re: Compile error - why?

["M. Lazarus" <mlazarus@xxxxxxxx>]

Doug,

At 10/8/02 01:42 PM, you wrote:
> >  My real question is why was a restriction put there in the first
> >place?  Level breaks handle numerics just fine.
>
> Some would argue this is not the case.  Level breaks on numeric fields
> consider
> only the absolute value and ignore the sign.  Many people forget this detail,
> since level breaks are often used on things like a customer number or order
> number, and these tend to not used negative numbers anyway.
 You're right - since I don't think I've ever used level breaks on a real
"number" field I don't think I've run into the absolute value issue.


> > Or it can just use a buffer character comparison (I would imagine that
> >that would be the most efficient) and ignore field type!
>
> That breaks existing rules (see above) so couldn't be used for packed and
> zoned
> fields.  Or did you just mean for binary fields?
 We were discussing binary fields, but it would be confusing if it behaved
differently than packed or zoned.  But it still begs the question: "What
technical reason can there be to  restrict level breaks on binary numbers?"


> >I would like to venture a guess that there are *far* more closet (or
> >open) "cyclers" than the crowd or IBM would care to admit!   :-)
>
> Let's not start that again...
 Sorry, I forgot to put in the standard "no flames, please don't start the
cycle wars again, etc." disclaimer.

 -mark