Re: Genexus development - opinions?

["Jon Paris" <Jon.Paris@xxxxxxxxxxxxxx>]

> > But the question then arises, why not just have GeneXus go to machine
code instead of an intermediate language? 

Because:

A) Then you'd have to do an enormous amount of work to port to another
platform.  While the amount of work needed is still significant, the basic
Java/COBOL/C/C++ generated would be the same.  So you are much better off
letting the compiler worry about the machine level.  In fact, as you know if
you think about it, the RPG complier doesn't even generate machine code for
the exact same reason.  It generates the generic W-code so IBM only need a
single "compiler" (i.e. the translator) to turn it into machine language.

B) It's hard to do.  The old RPG compiler generated MI (again not really
machine code) and even with the help of macros to assist in code gen it is
not an easy task.  

I've not experienced using Genexus myself - but those I've talked to who
used it were very happy with the results.  Just like with Synon and other
similar products like ProGen - you're not supposed to look at the code it
generates so who cares what it looks like.


Jon Paris
Partner400

www.Partner400.com
www.RPGWorld.com