RE: RPG is Dead? Say it aint so!

From: Jon Paris

>> What you will see is a concerted effort to integrate RPG as a
server-side language into IBM's GUI strategy, which is centered on EGL.

While I agree that _if_ EGL (or whatever its name is today) lives up to
its
promise, it has potential for further integration with RPG. However, to
base RPG's entire UI future on a product that may flop just as badly in
the
System i world as its antecedents would be short sighted to put it mildly.

Well, first of all I don't see RPG's entire UI future is contingent on EGL
or any other IBM UI strategy. I've been writing JSP just fine without EGL,
and can continue to do so. And of course there's always RPG-CGI, or
integration with one of the scripting languages such as PHP.

There are in fact a myriad of ways to integrate RPG with the web.

EGL is great - but it still needs an AS (WAS or whatever) and that
requires
a level of expertise in the shop that many may not be able to justify.

This argument holds no more sway for WAS than it does for any other new
technology, Jon. PHP requires skills that many may not be able to justify.
RPG-CGI requires skills that many may not be able to justify. This is a bit
of a straw man.

Remember, EGL requires NO knowledge of Java and NO knowledge of HTML. It's
a completely WYSIWYG environment that allows you to generate an industry
standard EAR or WAR file that can then be installed into a web server using
a browser-based installer that is no more complex than the instructions to
install Zend PHP.

Even
if they could justify it, corporate dictates in many shops categorically
rule out WAS and insist on .NET. So what do those RPG users do? Stay
green-screen?

No, this is EVEN MORE of a reason to not use a proprietary RPG opcode, but
instead to write your business logic as encapsulated servers. Then leverage
those servers through the UI language of your choice, be it EGL or JSP, or
expose it through industry standards such as Web Services or stored
procedures.

The LAST thing you should be doing is writing UI code in RPG, especially
when we start incorporating rich client technology! Do you REALLY want to
be writing the code to fill a combo-box and handle the events in RPG? Are
you arguing for a sort of Visual Age for RPG type of language? Me, I'd
rather not tie my UI to a specific back end language, any more than I should
tie my back end to a specific UI.

I believe it is much more important for the compiler to surface (as an
API)
some form of interface perhaps based on existing I/O op-codes, perhaps on
new ones. Such interface to supply the buffer, and information relating
to
the layout and type of the data, etc. Think SPECIAL files with far
greater capability.

We fundamentally disagree here. I'd rather the compiler team spend time on
basic issues like overloading procedures as Aaron needs or reflection
similar to what Steve complains about. I already have a number of perfectly
valid UI approaches and I don't really want them wasting time on it,
especially since the UI requirements are going to change one, two, five
years down the road.

Make RPG the best server language in the world and let me use the right tool
for the UI job (whatever that tool may be). But that's just me.

Joe