A successful lisp machine?

Reginald S. Perry perry@zso.dec.com
Tue, 29 Apr 1997 17:35:50 -0700

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

> I think there are so many cases/applications where "the web browser
> is the desktop" that we can ignore all the rest and we will still be
> a huge success.

> As I've said, what I want might be viewed as a very narrow focus.
> But I think that is why it can succeed.  It is trying to do just one
> thing better than anyone else can, which is utilize just the web
> browser for the user interface for applications.

At first glance I would agree with this, but I think that what happens
is that the special purpose tool slowly starts to be used for
(shoehorned into) the general purpose. I think that this applies to
more abstractions than programming languages. So I would bet that all
of the desktop stuff gets dumped into the browser so that the browser
is the desktop. Therefore I would say, dont go that way build the
desktop and give all of the desktop objects global extent so that a
particular object can be anything and live anywhere but be represented
on your desktop. 

So while I think that trying to narrow the focus is good, and the
browser is of major importance. I dont know intuitively how to prevent
people from attempting to get more power out of a thing than I
originally meant it to have. Especially if that thing is written in
lisp and intrinsically extensible. I will have to think more about
what you wrote.

> A lisp persistent object web-server is the right tool for this job.
> All the other approaches to this job are using the wrong tools
> (e.g. PERL CGI scripts) and rapidly become hopelessly complex to
> maintain and evolve when the application grows to any non-trival
> size.

This is very true. I would say lets use CL-HTTP as the base. All sorts
of interesting stuff is going into it.