> You are talking about a tool/IDE not Java.
This is true. But I have never used a GUI builder that was worth much.
One point in favor of JBuilder is that it doesn't die if you have the
audacity to touch its generated code.
> > ridiculous amount of Java code--constructors, adjustments,
> > assignments--totaling at least a hundred SLOC for something I
> > could have
> > done in fifty lines of TCL/Tk/wish. (And that's counting
> > modifiers like
> > -width 50 as SLOC!)
> It could be written [but never read] in one line of APL, but why count.
But in the case above, the much smaller wish was also more readable. You
want a button, you just say "button" and you list the properties that
matter to you. Any properties you don't list are automatically computed
to look good (usually) and they work the first time (usually). With the
AWT, I need to declare and initialize containers, configure a layout
manager, define and initialize things that I don't even understand before
it will even compile. And then it doesn't work.
Before anyone says I'm just defending what I know against what I don't
know, let me say I don't really know TCL/Tk either. It's just that of all
the (few) GUI methods I've tried (including C/X11) it's the only one I
could produce something in on the first try, or the second, or the third...
So to my viewpoint, the Javanaut is a mediocre language, a decent but slow
execution environment, and a mediocre set of libraries, all being used to
sell each other!
> > lines, Java took many. (Ada wasn't there, but would have
> > been about the same size as C).
That part wasn't GUI, it was just an aside. Where C uses printf with a
format string, Ada uses Text_IO.Put (Item, Fore => ___, Aft => ___); C++
took only a little more work, but the Java was absurd. OF course, it
could be that the person who posted it was just trying to slam Java by
picking the worst of several possible approaches...