On Fri, Feb 1, 2019 at 1:50 PM Barbara Morris <bmorris@xxxxxxxxxx> wrote:
I'm not even sure that RPG programmers would want it. Currently, you can
use the same copy files whether you start from free-form or fixed-form.
If there was a new compiler that only supported free-form, presumably it
wouldn't be able to use fixed-form copy files.
Right, that would be part of the point. If people wanted to use
fixed-form copy files, or indeed fixed-form anything, they would use
the RPG IV compiler.
I agree it's a lot of cost for development and testing. Any
calculation of whether it is cost-justified has to include a hard look
at things like how easy it is for folks at IBM to work on the new
compiler for the lifetime of the platform, and how much the "new"
language would improve the ability to recruit new talent and extend
the lifetime of the platform (in other words, how much value it adds
to the "brand").
I am not saying these would necessarily come out in favor of RPG V. I
actually tend to agree that they probably wouldn't. And of course
those things are inherently uncertain, even harder to pin down any
actual dollars than next year's sales projections, which are already
fuzzy and not guaranteed to prove correct. So it's a huge risk. I get
that. Companies making money hand over fist can afford to make those
kinds of bets; IBM doesn't seem to be in that position.
Given limited (read: relatively small) resources, I think investing in
PASE, basically what they're doing, is as good a route as any. I know
several of us, including Jon, would agree that improving the
interoperability story between QSYS/ILE and PASE would be very