Path: utzoo!utgpu!watmath!clyde!att!ucbvax!WATSON.BBN.COM!dan
From: dan@WATSON.BBN.COM (Dan Franklin)
Newsgroups: comp.society.futures
Subject: Re: desktop of the future
Message-ID: <8901092247.AA05892@multimax.encore.com>
Date: 9 Jan 89 22:49:30 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 51

Maybe we could talk about something more interesting?

I find the question of the "desktop of the future" a very interesting
one.  But comments so far have only described new hardware
developments like the DataGlove and small variations on the desktop,
like the "virtual room" (or set of rooms).

By themselves, these developments will not lead to breakthroughs in
computer use.  They'll never lead to the kind of stuff you see in
Cyberpunk books, where (apparently) computer hackers routinely survey
a "landscape" containing hundreds of computers, each with thousands of
files, looking for interesting information (or weaknesses :-).

What kind of future user interface SOFTWARE would lend itself to this
kind of operation?  Dealing with massive complexity--being able to
find the needle in the haystack--is clearly going to occupy more and
more people trying to cope with the "information explosion".

I can even speak from experience.  In my current work I often browse
among files in almost a hundred directories, with each directory
containing a couple of dozen files.  I'm trying to figure out what
part of a rather large software system is responsible for some errant
behavior that just turned up that morning, or what part of the system
would break if I made a certain change.

It helps a lot that I can represent each file or directory as one line
on a big screen, and I can order the files by date, type (filename
suffix, that is), or other salient characteristics.  I can even do
this for the contents of many directories at once.  (Try *that* with
MacOS!)  But it's still very clumsy, and I do this almost every day!

What I want is a way to look at hundreds or thousands of files at
once, in such a way that possibly-important properties, whatever they
are, spring out at me.  Maybe show each file as a small blot, and let
the color and shape of the blot reflect whatever properties I'm
interested in at the moment: files created more recently than 24 hours
ago, files containing calls to "system", or even object files that are
out of date with respect to the source files needed to recreate them.
These are all selections I've made (or wanted to make) as I went
browsing.  In a 3D world, I could imagine having the "interesting"
files literally stick out of the mass.

Or maybe I'm interested in a continuum, rather than a binary question:
how old is each file? How big is it?  How many functions are in it?
What files are the most complex by some (user-defined) complexity
measure?  Is there a correlation between that complexity measure and
revision rate?  And so on.  I could imagine lots of questions I'd want
to ask of these thousands of files, if only I could ask them easily
and get instantly comprehensible answers.  THAT'S where the future is!

	Dan Franklin