Re: Client-Side Scripts

Paul Burchard (burchard@horizon.math.utah.edu)
Mon, 13 Mar 95 01:33:09 -0700

> "Dave Raggett" <dsr@hplb.hpl.hp.com> writes:
> Some people have suggested Safe-TCL but I advocate a way
> that lets us duck choosing a single scripting language by
> instead defining a portable virtual machine (vm) that
> the scripts are dynamically compiled into.

My last reply didn't do a good job of expressing my admiration for
this idea. What I was attempting to say was, circumventing the
"language wars" alone makes this a very attractive approach!

It seems like a happy medium between the extremes of the
fully-interpreted and the fully-compiled approaches to client
extension, and I've updated my Interactive Standards discussion page
<URL:http://www.geom.umn.edu/hypernews/get/interactive/index.html> to
reflect that view.

My main concern about this approach is that if we have to start from
scratch, it's a major undertaking, and we only have a window of
perhaps 6-12 months to get an _open_ de facto standard out there.
That's a lot of byte-compilers to write, and vm language to design,
in a short time.

Perhaps fortunately, Sun's WebRunner project has done a lot of the
groundwork. But if Sun wants it to matter, they need to figure out
how to get Java tools into hands of a lot of people quickly (i.e.,
not just as part of their own experimental browser).

--------------------------------------------------------------------
Paul Burchard <burchard@math.utah.edu>
``I'm still learning how to count backwards from infinity...''
--------------------------------------------------------------------