A successful lisp machine?

[Marcus G. Daniels]

>>>>> "KM" == Kelly Murray <kem@Franz.COM> writes:

MM> but they make crappy window systems! You want to edit lisp code in
MM> a form's text entry window? Not me! 

Likewise.

KM> Well, first off, recall that emacs is also a web browser, and so
KM> you get a text-edit box under emacs, which does paren matching at
KM> least.  It can also be extended to identify the contexts as lisp
KM> code for even better support.

That's great for developers, but how are you going to implement an
interface to something like Excel without using the JVM, or installing
new native code?  And if it comes to installing new native code, it seems
to me to make more sense to install an all new program than to fuss
with plug-ins.

Again, I don't think installing new programs is a big deal; users do
it all the time with stuff like Acrobat and RealAudio.  I think
adequate free software exists to develop such a browser. 

Some ideas:

  o Help out with converting Emacs to multithreaded Guile (Guile will
    have Tk support).  Result:  ability to send Common Lisp, Scheme and 
    bytecodes at the browser.  Emacs already has the socket support.  XEmacs
    already has more widgety features.  
  o Write a new browser in Lisp (you may recall Gilbert Baumann 
    described some work he is doing like this).
  o Grab some other free browser, add lots of extensions to HTML or
    add a `Lisp command' extension, and drop in a a free Lisp implementation.
  
KM> Having a screen full of mouse-sensitive items and being able to
KM> execute a lisp function passing parameters for every mouse click
KM> can be used to do quite very powerful interfaces.

Yes!

KM> Also recall that in my vision that there are NO SOURCE FILES.  One
KM> doesn't edit a source file.  You edit classes, methods, functions,
KM> macros, interfaces, etc, which are part of a code/application
KM> module.

I think just having something like NeXT's Interface Builder, but having Lisp
`NIB' files would be nice.