Dick Balaska wrote:
>
> I was always surprised that M$ didn't "invent" springs 'n struts; it seems
> that s'n's would be a natural for Visual Studio.

It has surprised me as well. But MS doesn't invent, the acquire. Since
they didn't acquire Visix (although I'm told they made overtures once
upon a time), they were unable to invent the same feature.

And they wouldn't be able to patent it as their own. (For the same
reason Visix couldn't patent it - prior art.)

> (The problem with a Java
> implementation is that the class would have to tied to _some_ IDE.)

Not necessarily. Look at how other Java layouts work. The IDEs don't
produce resource files to be loaded at runtime. They generate code that
builds the dialog when executed.

IDEs from different companies can all generate code for SpringLayouts if
they want and the results should be portable, even if a competitor's IDE
can't parse that code for loading into its own IDE.

I will agree, howeer, that implementation of a springs-n-struts layout
is different from other kinds. Instead of having a single layout object
with a few ordered collections of widgets, your layout needs to generate
an internal object corresponding to every widget, in order to store all
of the connection information for all of them. This makes
implementation and debugging a bit tricky, and the code used to generate
these dialogs won't resemble other layout code very much.

It should be noted that this has already been done:

http://java.sun.com/docs/books/tutorial/uiswing/layout/spring.html

-- David