On 1/12/2013 8:18 PM, Booth Martin wrote:
Hopefully this will be written well enough to start a discussion and
avoid religious wars. My goal is to fairly hear the experiences and
conclusions from those who have already traveled this path.
I will never travel this path. My end users have an intense dislike for
change. I have two broad classes of internal users: those who use a
5250 emulator and those who call down asking for printed reports. My
company is large enough that external users are supported by a separate
web programming team.
What works? What has a broad base of support? Where is IBM going with
the user interface? How are designers, analysts, and developers moving
forward with the i?
The IBM UI is the 5250 data stream. Everything else is 'unofficial' and
IMO this is the Achilles Heel of the midrange platform. We wait and
wait and wait for IBM to provide the 'integrated' solution and we wait
so long we let the rest of the pack pull clear away from us.
When I say 'we', I am thinking of end user companies. Software vendors
have a different lot. They need to reach out and grab customers with
products that are useful and attractive. It seems upside down, but IMO
the vendors, as recalcitrant as they are, have actually driven the
modernisation effort in the midrange arena.
I want to develop my own skills to get the job done, and I am seeing far
too many choices out there to feel comfortable with my ability to pick
the right paths forward. I have spent time working with: VARPG, OS/2,
GWT, Java, HTML, JSON, XML, Zend & PHP, EGL, and other offerings. The
more I learn, the more uncomfortable I become with my skill at making
choices; hence this desire to hear what others have to say.
This used to bother me, but not any longer. Waiting (effectively doing
nothing) is easily the worst choice. Pick one and go. Of the list
above, some are fundamental to all of the choices. Everyone should know
enough HTML, XML and JSON to at least follow an AJAX transaction around.
The real choices are between Java and .NET, with PHP and EGL at the
back of the pack. Pick any one of these four and you will always find
work.
So, what is _your_ opinion of where the future lies for RPG programmers?
RPG's future is the same as it ever was. RPG is a server-side language.
Companies owning a midrange machine typically aren't interested in the
expense of migrating elsewhere unless they are driven to it by the
expense of maintaining the current software. I occasionally hear
someone complain about the aging population of RPG programmers but I
think this is the marketplace at work. When we've all retired, the
young will learn RPG the same way we did - the seat of their pants. All
of the difficulty has thankfully been removed from RPG. No more
matching record, look ahead fields, stacker select, detail time, total
time - all gone. The kids'll do fine.
--buck
midrange.com
