Green screen - it's time is over

Man, I thought I was long winded.  The same arguments have been being
repeated over and over here, to what avail I'm not sure.  But I thought I'd
try to recap a couple of the salient points, and then try to make something
constructive.

1. IBM did not one day say, "Let's charge extra for interactive!"  The price
per CPW has been dropping for both interactive and batch machines.  It's
just that the batch machines are dropping much, much faster.  I don't know
IBM's philosophy, nor does anyone on this list, but I have a feeling it has
something to do with trying to get as much revenue as possible from the
green screen while at the same time moving away from it.  Which brings me
to...

2. Green screen is going away.  The old 5250 interactive feature, which made
the AS/400 pretty much unique among the midrange world, is gone.  Dead.
Kaput.  You can still try to hang on to the old architecture, but it will
cost you.  That's the reality, no matter how much wailing and gnashing of
teeth happens here.  If you think you can change it, I suggest you get
together a coalition of people and talk to IBM.  Constantly rehashing the
argument here isn't going to do anything to change the reality of the
current situation.

And that's what I want to focus on for a moment.

The truth is that the AS/400, now spelled iSeries, is still the best
business platform available.  But its uniqueness no longer resides in that
wonderful integrated 24x80 window that we grew up with.  Instead, it lies in
the ability to write powerful applications in languages with tight
integration to its incredible database.  Personally, I think IBM's direction
of pushing everything to SQL is ludicrous, but it is silly for me to whine
about SQL.  Instead, I need to embrace it as best I can and work it into a
realistic development environment.

What I can do is to try to stop the proliferation of a couple of bad
elements:

ODBC.  Plain and simple, ODBC is a horrid idea for anything but the
occasional data mining application.  If you need to update data, do it
through servers.  In fact, learn to love the concept of tiered designs, and
build your applications accordingly.  It won't cost you a lot, and once you
have, it will be wonderful.  It sure beats SQL, which I have shown
repeatedly to be far inferior in performance to record-level access for any
transaction-based updates.

J2EE.  Enterprise Java Beans simply have little place in most applications.
The overhead is excessive, and a standard way of defining business objects
and the methods that update them simply hasn't been developed yet.  Until
that time, EJB is simply extra overhead.

If we avoid these two things, the iSeries, especially in its server
incarnation, beats any other machine out there hands down in total cost of
ownership, and in reliability and scalability.  If we join together to
develop some standard interfaces that allow data and programs on the iSeries
to be incorporated into general n-tier distributed applications, then the
iSeries will easily take on all comers.

Or, we can continue to bemoan the loss of our beloved green screen.  We can
reminisce wistfully about nickel soda pops and drive-in theaters, while the
world zips on by with Internet enabled applications running on Microsoft IIS
talking to whatever database server currently isn't crashing or locking up.
We can get used to unreliable systems and long delays while "indexes are
rebuilt" or "servers are synchronized".  We can twiddle our thumbs and
remember the good old days as yet another mission critical system succumbs
to some hacker's latest love child.

It's up to us.  The future is here, and in the future I see, the iSeries has
a huge part.  But it's not going to be as just another ODBC server - it's
going to be as the central business logic processor of my networked
applications.  I may have Microsoft, I may have Linux - in fact I may have
many of those boxes, since I'll need failover for the toy operating systems
that run the Flash presentations that smooth-talking dot-com consultants
sell to management.  But when push comes to shove, my mission critical
systems are going to be written in RPG, run on DB2, and talk to the world
through secure messaging.

You with me?  If so, quit worrying about the demise of the green screen.
It's already happened, but we just haven't admitted it yet.  Right or wrong,
the brave new world is upon us, and it's up to us to bring our platform into
it.  And if we do... if we do, we'll have not only the best damned server on
the market, but a whole new architecture that may just roll back some of the
tide of bloatware that has tainted our industry and our profession ever
since the first release of Windows.

No, I may never write an entire operating system in 32KB again, but I can
fight to make sure that business rules aren't written in SQL and databases
aren't updated by Java Beans.  At least for a little while.  And that isn't
just tilting at windmills, I don't think.

Joe Pluta
www.plutabrothers.com