[Gambas-devel] Re: Gambas and Visual Basic

[Rob]

On Monday 14 March 2005 09:53, Ahmed Kamal wrote:
>  Well MS did drop support for old VB6 apps with VB.net. Everything
> changed.

Well, Fabien was talking about vb6 programs not running at all under 
new Microsoft operating systems, not just having to rewrite your 
source for vb.net.

> If you could support VB.net, we would be embracing the
> future. Everybody is porting old code now.

No, if we ever claimed to support VB.net we would be chasing 
taillights.  Besides, the Mono project (specifically mbas) is already 
taking on that challenge, even if they don't have an IDE yet.  It may 
be that once they do, its popularity will make Gambas fade into 
insignificance, but I personally (and some of my clients) feel the 
same way about Mono as we do about Java.

But I'd love to be able to support VB5/6 and VB.Net, if I'm the one 
who ends up writing the translator.  It's just kinda likely that I'll 
do it badly, what with never making it past CS 101.

>  I always thought (Like Stallman) we should target some generic VM.
> Either mono (stop screaming), or www.parrotcode.org (This is really
> a good option IMHO), opens up tons of perl/python code.

I would be excited about a parrotized Gambas, but I question whether 
Gambas would still be Gambas at that point.  Much of what we have 
written would basically need to be flushed down the toilet and 
rewritten.  I would certainly rather use Gambas' Qt bindings than 
Perl's, for example.

Also, I don't know if you've noticed, but components are loaded with 
meta-information used in the documentation and elsewhere in the IDE 
to make the end user's life easier.  I don't even know if you can 
reliably enumerate installed Perl modules, much less enumerate the 
classes, methods, constants and properties embodied by each one.  
Ditto Python, Java, and .net/Mono.  Maybe it's there and I just don't 
know about it.  Being able to tick off a box to include a reference 
to a component rather than typing "#include" or "uses" or "import" 
etc. is kinda important to the RAD concept in a way that those who 
haven't spent their entire career using RAD tools usually don't 
understand.

Worse, those of us who have written components would need to re-learn 
to write components against whatever VM ended up being the chosen 
one, and from experience in other projects where the underlying 
technology was changed midstream, I can tell you many of us just 
won't bother (though maybe other people would join in.)

I think those who would like to target a different bytecode should 
grab the Gambas compiler source and just do it.  Make a proof of 
concept, post it, and show us why it's better.  If you're proud of 
your work and we're not receptive, you can always fork.  Free 
software projects thrive when the developers are excited about 
working on them.... VM advocates, now's your chance to make us all 
excited.

Rob