Re: /FREE RPG compiler errors

["Peter Dow" <maillist@xxxxxxxxxxxxxxx>]

Hi Steve,

> In mine and dare I say Bjarne Stroustrup's opinion Gary is wrong on this
> point.  One of the objectives of good program design is to abstract away
as
> many of the underlying details of the objects of a program as possible.
In
> C++, the action of "A = B" depends on what A and B are.  All the coder
knows
> ( and needs to know ) is that the value of B will be assigned to A.   Same
> thing with the RPG MOVE ( should be COPY but who wants to quibble? ).
When
> you see a MOVE stmt, you know that the value of factor 2 is being copied
> right justified to the result variable.

I'd have to say it depends on the situation.  Sometimes a coder _needs_ to
know a whole lot more than "that the value of B will be assigned to A". With
a MOVE stmt, it's not even necessarily true that the _entire_ value of
factor 2 is being copied to the result variable (i.e. factor 2 is larger
than the result field). It also has the conversion aspect as well as the
copy (numeric to character, numeric to date, etc.).  Abstracting too much
away reminds me of debugging an optimized program...


> The problem with not including MOVE in /free is that IBM did not provide
the
> ability to write a proc version of MOVE.  What is needed are data pointers
> in RPG, that is a pointer that also contains the attributes of the data
> pointed to.  Or an OPDESC that actually has a complete set of operand
> description information.  Or allow programmers to write their own BIFs.

I like the idea of OPDESC providing a complete description of the operands.
With that I'd think you'd pretty much be able to write a procedure with the
same capabilities as any BIF.

Regards,
Peter Dow
Dow Software Services, Inc.
909 793-9050 voice
909 522-3214 cellular
909 793-4480 fax