Re

[dwig@xxxxxxxxx]

Andrew Pam writes:
> Does anyone on this list have any questions about Xanadu, or any suggestions
> on how to support and promote the ideals?  Would anyone like me to repost our
> aims and objectives to kick off some discussion?

Well, I'll try to get things going by taking a bit of a Devil's advocate
position.  I should preface this by saying that my knowledge of Xanadu is
mostly based on the late 80's version of "Literary Machines".

I've been working in software product companies for the last decade, after
earlier experience in applied research environments.  From the latter
viewpoint, I can see the beauty and promise of the Xanadu architecture and
ideals.  From the former perspective, it almost looks like Xanadu is an idea
with a great future behind it.  The realities of the information systems
marketplace have evolved around it, adopting little but the basic idea of
the embedded link.

Possibly the two most successful hypertext systems today are the Microsoft
Windows help file and the WWW, for similar reasons.  The former, because
Microsoft made it ubiquitous on Windows platforms, and the latter because
anyone with a machine connected to the Internet can get free software to
access it and add to it (the success of Hypercard on the Mac was similar).
Of course, there are significant tradeoffs between them.  For example, one
Windows browser, Cello, provides its online documentation by a Winhelp file;
another, Netscape, by a set of centrally located Web pages.  People who have
tried both tend to prefer Cello's approach, at least partly because help is
available even when the net isn't.

We could argue ad nauseum about the technical advantages of Xanadu over
existing systems, but it would be pretty much academic.  Practical questions
for Xanadu include:

- Does it provide any advantage readily perceivable by the community of
  hypertext users, one that would lead them to spend time, effort, and bucks
  to switch to a Xanadu-based solution?

- Can it be implemented and marketed in a fashion that minimizes the pain of
  adoption for the average owner of a PC and/or Unix LAN system, who may
  already be using such things as WWW, Lotus Notes, and Winhelp?

- Are there components of its technology that can be extracted and added to
  popular hyptertext systems, furthering the Xanadu vision in a piecemeal
  fashion?

- Are there parts of the architecture that could be used to guide the 
  evolution of of currently successful systems, thus bringing them closer to
  the Xanadu vision?

- Can its architecture be adapted to an open systems approach, in which it
  may be used on everything from a single local machine to a WAN?

I'm not sure whether these are the truly crucial questions for the viability
of Xanadu, but they're in the ballpark.


Don Dwiggins				"The truth will make you free,
Mark V Systems, Inc.			 but first it will make you miserable"
dwig@xxxxxxxxx				 -- Tom DeMarco