Re: Compiler directive

On 8/9/06, Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx> wrote:

> INCLUDES are 1980s technology. If IBM has any skills remaining in
> languages and ILE I vote they enable service programs to contain meta
> data ( if that is what it is called ). The definitions of  prototypes,
> constants, structs should all be bound in service programs with the
> compiled procedures.

I'd like that even better, but it's not the way ILE works today, and there
are hurdles that would have to be overcome before they could begin to make
it work that way.  (For example, the fact that CL doesn't support
prototypes, and the fact that different languages use different calling
conventions.)

Furthermore, it means a departure from existing standards.  Sure, IBM can
change the way RPG and CL work, but COBOL and C have to adhere to
standards. You're talking about a change to all ILE languages, and as
such, they'd have to depart from those standards.  No other C compiler on
the planet gets the prototypes from the objects they're bound to, and I'm
pretty sure it's not ANSI C.

Leave C behind. Just like those software visionaries Bill Gates and
Anders Helsberg did at MSFT in the late 1990s when they bet the farm
on .NET

http://channel9.msdn.com/ShowPost.aspx?PostID=10804

I think it'd be great if IBM would overcome those hurdles and make the
capabilities available, but I think they'll just say "if you want that
sort of capability, use Java."

True, but IBM as a corporation is lazy. The MSFT CLI approach is
better than Java because it doesnt lock you into a single language.
The C++ that MSFT provides is CLI compliant. C++ and Java have a lot
of differences. I have not tried it, but my guess is ANSI C code will
be compiled by the C++.NET compiler.

-Steve