Re: [rgui-dev] RE: Backend Debouch

[Laurent Julliard]

Tom Sawyer wrote:
> NOTICE TO ALL GUTOPIANS: THIS IS AN IMPORTANT THREAD, SOME ADDITIONAL
> INPUT WOULD BE VERY USEFUL. PLEASE EXPRESS YOUR OPINION AS THIS COULD
> FUNDEMENTALLY CHANGE HOW WE PROCEED.

Cannot agree more ;-) If one day my advice prove to be pure crap a 
couple of people will probably come to me and ask a few questions....

> hmmm...so back to the main point: SWT C source versus using Ruby-Gnome,
> SWin/VRuby, and RubyCocoa/RubyAEOSA. okay, here are, point by point, the
> questions i need answered:
>
> * who will do the "JNI" C translation for Ruby. i fear it is out of my
> league, as i haven't used C for over 10 years.
>
> * if we use SWT who will build the respective GUtopIa backend
> interfaces? I can hadle one of them (GTK's) while i hammer out the the
> primary meta-api, like i'm doing presently, but not all.

You know what. Take the lazy path for now. See if the existing MSwin and 
GTK backend can be used for an SWT like approach. If it works then fine 
keep them. If you realize in the course of this first prototype that it 
requires too many changes or doesn fit well with the SWT approach then 
you know what to do.

> * how do we integrate ParaGUI, Wise, or other backends if we use SWT?

As I said once you have proven the GuTopia approach is the right one I 
expect ParaGUI, Wise and other backends author to maintain the GUTOPIA 
backend

> * finally, is licensing an issue?

You mean, is GPL an issue ?? :-) You are going for GPL right (or may be 
LGPL)

> ~tom
>
> p.s. i read most of those readings you linked for me and also looked at
> the java code a bit. they make some fair arguments and i have already
> implemented some of those notions into gutopia. the java code is quite
> large, just a single widget had what appeared to be some odd thousnad
> lines of code. is this perhaps due to the one-to-one correlations? well,
> there was a lot of remarked docs in there too, so i'm sure that counts
> for a good bit. it just seems rather large to me, but perhaps that's
> just how it ends up when complete and i'm being nieve.

It is a big thing but but as you said it is also a piece of literacy :-) 
 Don't let the size impress you. Their approach is very mechanical as I 
said and it probably generates under-optimized code here and there. But 
as I said this is on purpose. They wanted the low level stuff to be 
simple-stupid and based on uniform interfacing and coding principle. In 
some way they applied the famous POLS principle.

Keep going, start something... and more people will join.

Laurent
--
Laurent JULLIARD - Xerox R&T/SSTC/XPA - Open Source team
>> Host your Xerox Software project on CodeX: http://codex.xerox.com
>> address@hidden community: http://xww.linux.world.xerox.com