On Fri, Jul 18, 2014 at 1:06 PM, Nathan Andelin <nandelin@xxxxxxxxx> wrote:
This question is only hypothetical, because I'm sure that IBM will continue
to support and enhance ILE RPG for as long as IBM supports IBM i.
I'm sure of that as well.
If IBM were to abandon support for RPG, would you abandon IBM i?
This is a weird question, because it kind of depends on what you mean
by "abandon support". For example, if you just mean that IBM will
stop providing updates and fixes, and will stop providing technical
support, then the answer is a very simple and resounding: We would
not even abandon RPG, never mind abandoning the i!
There are businesses that are still using V5R2 in production, I'm
sure. Probably even V4R5, and maybe even older systems than that. In
other words, for these businesses, EVERYTHING is already unsupported,
yet they continue to use what they have, and probably even still write
"new" stuff with what they have.
So, to my mind, that's not a very interesting question. Equally
uninteresting is if you mean "what if RPG *the language* were to
suddenly disappear, and anything written in RPG suddenly stopped
working?". Clearly, lots of businesses would grind to a halt, and
probably most of them would not survive. The smaller ones in
particular wouldn't have the resources to migrate (even to something
else on the i) in a reasonable amount of time. (So even with advance
notice of the RPG apocalypse, they might not make it.)
That leaves the funky situation of: "what if RPG *compilers* were to
suddenly disappear, but any existing RPG program objects continued to
function as before?" Well, that would definitely force a lot of folks
to start migrating. The question is whether they do it piece by piece
(organically rewriting programs from scratch in a new language only as
bugs are found and new features are needed), or wholesale. If
wholesale, then do they migrate off the i completely?
Where I work, I think we'd stick to the i and rewrite gradually. I
don't think how much we "love" the i is really much of an issue. If
we're being honest, most platforms out there are pretty mature and
pretty competent. It's not like Linux-based business are crashing and
burning and flocking in droves to the i for safety. Even
Microsoft-based businesses are frankly doing just fine, thank you very
much. The issue is that migration is an expensive, painful, and
time-consuming endeavor, no matter how you slice it.
What would we rewrite in, given that I said we'd stick to the i? The
most obvious (but boring) choice is probably Java. Personally, I
already do a significant chunk of my development in iSeriesPython, and
I would be very sad if I had to give that up. We'd almost definitely
depend more on SQL than we do now. I think the biggest thing we'd
have to deal with is weaning ourselves off the green screen. Which is
something we're already trying to do, but it's going extremely slowly
because the urgency is not there. By urgency, I mean at all levels,
from technical to cultural, from the big boss to the users to us. It
probably *would* take RPG-compilergeddon for us to really get in gear.
John Y.
midrange.com
