Re: RPG to Java Conversion

[TitanRebel <TitanRebel@xxxxxxxxxxx>]

Wow, I guess I need to check my email more often.  Looks like I've got 
several questions to answer.

Why did we replace some business logic and the interface at the same 
time?  Well, it just worked out better that way.  A lot of the business 
logic that we put in was new or "enhanced".  So the timing was right. 
In the future we will pick and choose which projects we should 
replace/rework some or none of the business logic.

Why didn't we use client side Java?  For the obvious headaches 
associated with deploying/distributing Client/Server apps.  Java 
Webstart was brand new when we started, and is a technology that we 
would like to research in comparison to Applets.  So, if anyone has 
experience with Webstart and Applets, please let me know what you think 
the pros and cons of each are.

Why not use Visualage for RPG?  Well, if I remember correctly, when we 
evaluated it, it generated Client/Server code.  This is a model that we 
wanted to avoid, as stated above.  I don't know about it's ability to 
generate Java.

Why not use Transidiom?  Well, the "no per-seat" license isn't as 
advantageous for us.  We are a software shop and a hardware reseller. 
So each of our installations would have to absorb the cost of the 
Transidiom server, increasing the price of our software.  While a 
partial conversion to Java is costly as far as training and time lost 
during training, it is an investment in the future.  We can change the 
way we create new software, so that we don't have to depend on a 3rd 
party tool to support "aging" programming methodologies. -- With that 
said, we don't plan on dumping JWalk and RPG, and we can't rule out 
using tools like Transidiom and Mr. Pluta's tool for applications that 
we don't want to spend the time converting by hand.  We have many, many 
applications that could possibly be implemented differently.  The 
applications that we have chosen to convert are ones that we want total 
control and flexibility over.  Our RPG applications are always being 
changed and enhanced, and that won't change.  We don't have applications 
that stay the same.  We do have parts of an application that are mature 
and don't seem to change much anymore, and we are going to try to 
leverage that.  I guess my point is, if we are always changing 
application code, what 's the difference if we do it in RPG or Java. 
The only negative one that I can think of is the initial learning curve 
associated with the change (and I can think of several advantages). 
That is what my initial query was about.  Does anyone know of "best 
practices, success stories, etc..." literature of taking a few RPG 
developers to Java?  I did NOT mean to give the impression that "we are 
converting tomorrow and never looking back".

Ken Kennedy