Re: JavaFX viability was->Re: Pete's web5250

Aaron Bartell skrev:

>Is it at all possible always to get your core business on a single
platform?

It wouldn't be an easy task but I know it can be done in most cases.\001
And before you think I might be saying 100% RPG green screens I will head

I'm not saying that. I just know that many ways exist to get to data inside an AS/400. Unfortunately very few of them are available ON the i except green screens.

interfere with the longterm viability of your core apps.\001 At my
previous employer I got all excited about Java because it had a nice
library of things that appeared to make things like web services and web
pages easier.\001 At the time it was a 99% RPG shop.\001 I decided, in my
idiot youth, that it would be an ok introduce Java to this shop and write
a portion of a core business process (tax calculation web service) with
Java+Tomcat.\001 Nevermind the fact that Brad Stone (e-RPG fame) was

We have had very good experience with letting Java provide a network connection point, massage the XML payload into a normalized form, and sending THAT down in a member file to a Cobol program with a rudimentary XML parser, pulling up the response and using XSLT to transform it back to an XML response as expected in the other end.

This allowed us to keep the logic in the Cobol program.

literally sitting a cube away - I wanted to taste the Koolaid because
everything I read was that Java would be the next big thing and I didn't
at that time know how to accomplish the task with RPG.\001 Well, since
then I now get calls from my previous employer to fix that Java whenever
it breaks.\001 Java never took off within their walls and now it is just a
pain in their butt.\001 If I would have taken a measured approach and
evaluated what introducing Java+Tomcat would mean then I would have
realized that RPG would be a much better decision for the shop (and very
do-able).\001 It is all about making the right decision for each
individual shop based on what their core technologies are.

It is always a trade-of. Doing everything inhouse in RPG is nice for smaller projects but I expect a scaling problem similar to what chefs experience. It is easy for a great chef to create a great meal for 4 persons, but hard for a great chef to create a great meal for 4000 persons in 5 localities and so on.

Personally I am certain that we should do everything network related in Java, and everything business related in Cobol as that gives the most for the money.

As I understand from your past posts you are a fan of not being tied to a
vendor and prefer to find a happy medium by combining a bunch of open Java
technologies together even if it takes some manual plumbing on your part -
I would consider that NOT to be the norm but also consider it a direction
that more shops should take vs. switching languages/platforms just because
one single new part of technology makes something slightly easier.

I think that basically MY problem is that I don't like having control over important components I use to create long lasting solutions for my employer (which is why I tend to dislike closed source solutions), and I don't like being tied to a particular product - open source or not. I like to be able to choose :)

Thinking that over I believe that Sun has just carried the RFC tradition of publicly available informal standards on, with their various java specifications, which has specifically been geared towards being able to create pluggable components. See the servlet specifications which has resulted in that web applications really ARE 99% portable, and the JavaServer Faces specification which has resulted in several independent implementations providing a choice :) and all this is running on a JVM where you usually have several to choose from.

Hopefully this clarifies my point of view. Being able to choose, means being able to pick the appropriate component for a given situation. My web applications can run equally well under an embedded Jetty behaving as a Cobol component, as well as part of a large WebSphere installation.

I might be misplaced or misinformed but I truly believe that one-size-doesn't-fit-all and that goes even in the i-world.

One of the statements I make in the RPG related sessions I give at
conferences is this "There is very little RPG can't do."\001 The same
could be said for almost any language out there these days as long as the
developer holding the language isn't afraid to tread into the unknown.

Sounds like the "Turing complete"-concept :) You are in good company there.

Anyways, hope that gives you an idea of where I am coming from.

Hopefully me too.