[rgui-dev] Say no to SWT! :)

[Leon Torres]

Ok, this has gone way too far. Man overboard! Stop the presses. Hit the brakes.
I'm gone for a weekend and decide to slack a bit, then I come back to see that
the revolution is trying to put SWT on the throne! What madness is this? What
happened to Leon Trotsky? Did SWT assassinate him in Mexico or what? A SWT
backend will change my expectation of gutopia completely. I'm expecting Leon
Trotsky. I will not let my sweet, dear ParaGUI binding be married to any other!

Who is this Leon Trotsky I speak of? Well, me of course. ;-) I'm named after
that guy, and my expectation of gutopia should be its golden rule. Oh, and
please forgive my arrogance here. I'm feeling very Russian today after
hitting the Popov. It didn't agree with me. Good advice: if you ever see a
plastic bottle of Popov at a party, run away screaming. Where was I? Da,
the golden rule of gutopia.

THE GOLDEN RULE OF GUTOPIA (as imagined by yours truly)

Binding tks to gutopia should not be so hard and take so much effort that I
decide to give up on it and take my business elsewhere or do without. All
other concerns are secondary.

This means, when working on binding:

1) Minimize the need to bind by hand, use SWIG and similar tools maximally
2) Implement as little or as much of the gutopia stuff as desired
3) Whatever's not implemented can be tacked on effortlessly
4) gutopia should be ready _now_ so I can do my bindings
5) As little C as possible, Ruby preferred (follows from above pts, actually)

So how does SWT destroy this expectation?

1) gutopia won't be ready _now_, probably _never_ (SWT is big effort man)
2) I have to write the gutopia-ParaGUI bridge in C, so much for swig help
3) I have to write the gutopia-ParaGUI bridge in C, so much for easy
4) I have to write the gutopia-ParaGUI bridge in C, so much for tack-on
5) I have to write the gutopia-ParaGUI bridge in C, so much for not using C
6) Well, you get the point :p

Forget it, I'll just use my swig bindings raw. Gutopia/SWT isn't scratching my
itch. And I believe that I'm not alone. Without developers willing to
grow it, gutopia won't grow. Remember, in open source, the developers come
first. Our goal might be to make the user's life easy, but we can't be so
altruistic that we overlook the motivating factors for developers!

Let me tell you something else. Gutopia will die if it isn't
developer-first and if it doesn't start happening _now_. It's simple
really. There are efforts to make a pure-ruby gui tk. Not having to
code in C is heaven. I'd rather help one such project plug into SDL
than waste time on gutopia/SWT. Once the ball gets rolling on the
pure-ruby tk, something very big will be born. Think TCL/Tk. Gutopia will
be rendered redundant. Until such a time though, gutopia has the chance to
be a bridge between now and the future RubyTk. It has the potential to
start converting lots of people _now_ rather than _later_. This is what
we should be aiming for. But that's a story for another thread.

Tom has suggested we look at Ruby/DL. I think he's on to something very big.
Let's all take a breather, smoke the peace pipe, whatever gets people in a
mood for contemplation, and take a good look at Ruby/DL. I am more than
willing to use Ruby/DL instead, especially if we can automate some of
the common binding algorithms. I'm thinking of using it for the
super-secret project mentioned above if gutopia fails my expectations.

Cheers,
Leon