OO-RPG was ->RE: Qualified sub-procedures -- RPG400-L

>Are you suggesting that an OO-RPG could be a "better way"?  Sorry, I
>can't see that. Unless we copy outright an existing OO language.

>From what I've heard C# is Java except it takes out some of the things in
Java that are frustrating or confusing.  Could RPG do the same thing, yes.
Doing an exact duplicate of another language would be foolish, but taking
each languages attractive points and coming out with a shiny pearl would be
the solution.  I am not saying "develop the end-all-programming-language",
just that there are some darn useful things in other languages that I wish I
could make use of without having to switch back and forth.

>Why are there so many OO languages? Why are there some many types of
>motor vehicles? Because some are better suited to particular tasks
>than others.

Exactly!  And aren't the RPG programmer's particular tasks changing?
Couldn't OO concepts be used today in RPG?, Absolutely!  I won't argue that
RPG hasn't gone through a lot of revisions through the years to accommodate
change.

>I'm not sure I see your point. I've always argued that you should
>use the appropriate tool for the task at hand. RPG is good at some
>things, such as implementing business logic. But, not so good at
>other things. (Others have argued that OO is not even suitable for
>business logic. I don't agree, but I'm not going to debate *that*
>point in this thread!)

My point is that the compiler needs to change with the business needs.  Why
did IBM allow you to have time to put in free-format?, because it was seen
as a benefit to RPG even though there are many many other languages that
companies could have switched too to get that same functionality.  See what
I mean?


>Unfortunately, there's a big learning curve in learning OO, and even
>bigger if you want to learn it properly! Just having OO features in
>a language does no good if you can't use them properly. And there's
>a big learning curve in learning particular class libraries, which
>are an important aspect of good OO languages. In the time it would
>take to learn how to use an OO-RPG, you could learn Java, or any
>other OO language.

My opinion on this would be that you can be very productive in an OO
language and use only the aspects of it that you feel benefit your
programatical architecture.  If someone doesn't know how to use interfaces
in Java do they care as long as they can create their own data types.

The latter part of your statement is most likely very true, but they would
at least have a head start if they already knew some ILE.  But that would
get back to RPG running much faster than Java on the iSeries.  That is
probably one of the main reasons I haven't utilized Java to it's full extent
is because of it's separation and slowness on iSeries.

>In many ways, there's no reason you can't implement OO-like designs
>today in RPG, if you understand OO.  You could come up with some
>sort of OO framework in RPG.  For example, you could have a
>framework where you allocate objects such that the object is
>prefixed in storage by a couple of pointers, including, say, a
>pointer to a class descriptor.  You could define procedures that
>invoke methods on the objects, and the framework could search for
>the appropriate procedure based on the class hierarchy you set up.
>Implementing a storage manager with an effective garbage collector
>might be a bit tricky, and mixing RPG variables with objects might
>be a bit of a problem, but should be doable. (However, I see that as
>more of an academic exercise, and not really a practical one.)

This is exactly why I want it to be at the compiler level;-)  It sounds like
you already know how you want to do it:-)


Aaron Bartell
OO-RPG Programmer