Les Tyrrell wrote:
> 
> "Ken G. Brown" wrote:
> >
> > I have also wondered how close Xanadu <http://www.xanadu.net/> is to
> > being a usable system. It sounds like a lot of good thought has gone
> > into it. It has gone Open Source <http://udanax.com/>, the source is
> > in Smalltalk and Squeak is mentioned as a likely candidate for
> > porting <http://udanax.com/gold/index.html>. A Smalltalk to C++
> > converter is mentioned.
> > Has anyone looked at Xanadu in any depth? What are it's drawbacks?
> > Why isn't it going places in a hurry? Or is it?
> >
> >     Ken
> 
> I would focus only on Xanadu's virtuality. Nelson makes very compelling
> arguments about what the virtuality should be.  Understand that first.

The virtuality isn't that hard to grasp, really. It's basically one big, 
happy universe of freely interlinked and transpublished documents.

> The existing implementation is very tough going, and the released code
> is missing a very large chunk of Xanadu, not to mention instructions
> for starting it up.

The Xanadu implementation now is mainly a bunch of hacks to
overcome a lot of the technical difficulties that arose when
trying to implement the Xanadu idea.
 
> A new ground-up implementation seems to me to be the only viable
> approach, once one understands the issues that Nelson raises.
> That would take a fair bit of time- but I don't think that this
> is anywhere near the technical challenge it was when Xanadu began,
> or even when the 92.1 version was built.

Well, I don't think it would be *that* hard anymore to come up
with a new implementation that supports the basic ideas. 
Technologies like Jini & JavaSpaces offer a lot of the 
infrastructure needed. Mixed together with XML-technologies
you could get quite far. Throw in Smalltalk-based CORBA
implementations as well and you could even supercede the original 
idea ;)

Greetz,

RS