Re: open source rpg compiler -- RPG400-L

Steve Richter wrote:

Hans, if you are not part of the decision making on open sourcing of the rpg
compiler, I am not going to badger you.  If you are, then that is another
story!

What I am suggesting is that IBM release the source code of the RPG
compiler.  Also, do whatever has to be done to make sure that any derivative
compiler would require the underlying websphere.  This assures that IBM does
not lose any revenue from the sale of an RPG offshoot compiler.

Once the source is released in this manner, why would IBM have to
"coordinate and validate the open source changes"?   Esp if the derivative
language does not have RPG in its name?

This is, of course, an interesting debate. But the idea of opening up the source of the current RPG IV compiler is impractical for lots of reasons.

First, the "proprietry" language the compiler is written in is C. In fact, the same C compiler that customers use. Since everyone knows and loves C, that's not a limitation to RPG as an open source project.

So, what's wrong with the idea? First, let's have a look at how a couple of other open-source languages work. Both Perl and Python are managed by a central core group of developers who maintain and enhance the code. They also review and approve changes submitted by others in the general community.

Although I'm sure some RPG fans would love to work with the RPG compiler source themselves, the community as a whole would not be served well by multiple RPG compiler products available, each compiling a slightly different version of the language. Most RPG shops would still choose to stick with one "official" language. Likewise, in the Perl and Python communities, although there's always the possibility someone might fork off their own version, in practice it never actually happens. For example, in the Python community, changes are subject to the approval of the creator of the language, who's often referred to as the "Benevolent Dictator For Life" (AKA, BDFL). Although some might disagree with his pronouncements, everyone agrees that a single reference language is beneficial, and so no one has any desire to fork the language.

And so, with an open source RPG compiler, we'd have a community of developers providing patches to the compiler and a central core group of developers reviewing, merging code, and testing the resulting code to produce an approved RPG release. The lab would need to hire additional people to manage that whole process!

Then again, in the Perl and Python communities, how many programmers actually do submit enhancements or fixes to the base code? I suspect relatively few programmers have the time to invest in learning the details of the language implementation, let alone the time to spend in investigating any possible change. Most enhancements provided by the community are in the form of class libraries. But that can happen whether the base code is open-source or not. The internals of the language are not needed in the implementation of a class library. Likewise, in the RPG community, enhancements can be (and are being) made by anyone in the form of service programs.

So what's my point? First, although some people would love to implement their own changes, the community as a whole would not welcome a forked language. Second, the time it would take to understand the compiler code would be a major barrier for most. Thirdly, a knowledge of the compiler internals is not even necessary to produce "enhancements" in the form of service programs.

Cheers! Hans