RE: Problem Found! Need solution

["Joep Beckeringh" <joep@xxxxxxxxxxxxxxxxxxx>]

Joe,

Think of it this way:

Defining a variable by number of bytes is a very computer oriented way
of thinking. The limits for integer variables are strange from a human
point of view (excluding nerds like us for a moment).

Defining a variable by number of digits is a very human oriented way of
thinking. The resulting storage is messy from a computer point of view.

We may think of RPG as a programming language, but originally it was
more like a tool. Some of its heritage still shows.

Joep Beckeringh

BTW: I am so glad that the first element of an array actually has number
1 in RPG!


> -----Oorspronkelijk bericht-----
> Van: rpg400-l-admin@midrange.com
> [mailto:rpg400-l-admin@midrange.com] Namens Joe Pluta
> Verzonden: dinsdag 17 september 2002 18:26
> Aan: rpg400-l@midrange.com
> Onderwerp: RE: Problem Found! Need solution
>
>
> > From: Joe Pluta
> >
> > What I asked was why a 4B field is the same size as a 5I field,
> > but not the
> > same size as a 5B field, which is even more confusing than the
> > 4B/BINARY(4)
> > mismatch.  You've alluded to the fact that it has to do with decimal
> > precision.  A 5I field can hold any decimal representation
> of a two-byte
> > binary field, which can in turn hold any four-digit decimal
> number (4B).
>
> A little further thought on the subject yields the following:
>
> B fields can have any precision, because they are used to
> represent actual
> application data, just ilke packed or zoned fields.  The size
> of a type B
> field is the MINIMUM size of a binary field required to hold
> a value of that
> precision.  Thus, a 4B field can be held in two bytes, while
> a 5B field
> requires four bytes.
>
> The I/U fields, on the other hand, are used specifically to
> define a binary
> field.  You specify the minimum number of digits required to
> hold all the
> possible values for a binary field of a given size.  That's
> why I/U fields
> can only be specified for 3, 5, 10 or 20 digits,
> corresponding with 1, 2, 4
> and 8 byte binary fields.
>
> At least that's my read on the situation, and I'm stickin' to
> it <grin>.
>
> Joe