Re: RPG V?

Hans Boldt wrote:

What do you think? Is there a need for RPG V? If IBM is indeed serious
about an RPG V, what might the rationale be?

I find the differences between RPG II and RPG III to be minor. I liked
externally described files, but I was using /copy for that anyway, so it
wasn't revolutionary. Subfiles instead of arrays was another
improvement, but my subfile programs weren't that much easier to
maintain vs the CCP code.

Moving from RPG III to RPG IV was a revolution, a stunning improvement
due to one thing: local variable scoping. Embedded SQL was interesting
but the ultra-warty pre-compiler back then made it painful to use.

If I were to get a revolutionary RPG V, it wouldn't look be in the form
of free form F specifications. Anything that can be handled by
CVTRPGSRC might be interesting, but it almost certainly won't make my
code tangibly better. For that, I'd need to see name space support,
full and unequivocal NULL support and full and unequivocal operational
descriptor support so that I can finally write my own universal BIFs. A
fully integrated SQL pre-compiler would be right up there too, but that
team has removed many of the pain points I used to routinely encounter
(thanks, Gina!)

Maybe what IBM really want to do is add a new parameter on CVTRPGSRC
called TGTLANG() or something, so that customers could finally convert
off of RPG onto C++ or something like that, I don't know. If I were
spending my $100 in the 'what do you want for RPG?' survey, I wouldn't
spend much more than a dollar on free form H, F, D, O specs.

Some improvement is better than no improvement, so if the choice were
RPG f-cubed (fully free form) or nothing, I'd take f-cubed. But it
wouldn't be my first choice.
--buck