| [Return to Home Page]
2000) [Printed in "Reality Module No.17" as part of "Freeform Futurology
FREEFORM FUTUROLOGY (3)
(A casual series of articles exploring various aspects
of our evolving society)
Smashing WindowsTM - The Ascent of Non-linear
"We'll develop from the Internet, through the Intranet (the Internet's
face in your own Local Area network), to the Internet within your own
(Article for the Amiga User's Group - July 1996)
We manage projects at work - this includes a collection of reports and
resources for websites. The Department's Operating System of choice is
WindowsNT release 4.
The Windows Explorer utility shows a bewildering array of directories,
sub-directories & sub-sub-directories.
This structure is the offspring of old DOS though now the file and directory
names can be long & meaningful. (How much harder it would be to manage
and keep track of thousands and thousands of computer files if we still
had to cope with the old 8.3 filename convention!)
But now it is becoming more and more apparent that the old hierarchical
filing-system model is inadequate.
To explain the limitations of the model & the solution to its problems
I shall compare books in libraries with webpages in cyberspace.
Computer files are treated like books in libraries - they are given one
unique identifier (a filename or a Dewey number) and are kept in one special
place (among the Astronomy books or in the 'Correspondence' directory).
Libraries have the advantage over harddrives - books can be given subject
headings from anywhere in the spectrum of knowledge - and a physics book
with the added subject 'Chemistry' will be found on a subject keyword
search of 'Chemistry' and will be called up along with all the chemistry
books even though it is kept apart from them.
In Windows Explorer we can make an educated guess at the appropriate
directory for a file (navigating a hierarchy we hope we understand), or
we can use the slow tedious find function to scan files for keywords.
But in a filing-system that is not entirely your own - with inherited
files and directories, and files and directories created by people whose
minds work differently from your own - it is easy to get lost & confused.
This confusion is the result of the outmoded concept of the hierarchical
directory system, and the limitations of a linear (one process to one
place) approach to file storage. Like how the old 8.3 filenaming convention
became unworkable when your collection had grown to hundreds of unique
files - tasks and processes in the digital workplace have gone too complex
to work efficiently with current file storage conventions.
To begin to describe the alternative non-linear approach I shall now
consider the Web.
When librarians took to cataloguing websites they realised that unlike
a book a website didn't have to live at one place and didn't need to have
only one Dewey number - it could live at multiple locations and have many
Of course the Web is chaotic (with local pockets of order) - and neither
fully indexed nor capable of being fully comprehended. It is like a huge
huge harddrive with files from everybody on it, but with a tremendous
advantage over the harddrive on your computer - nonlinearity! Anything
can be linked to anything else - like a great big gigantic mindmap.
The invention of metadata serves to act like a library catalogue card
held within the webpage itself - data about data (literally) - an invention
to allow the extraction of meaning from bits & bytes. I'll come back to
this a bit later.
The old hierarchical filling system suffers from the limitations of
linearity. It is designed under the (now known to be false) assumption
that every computer file has one purpose and has one ideal parent directory.
Another minor aside - consider email. We can sort email messages in
lots of useful ways - by sender, by recipient, by topic, by date - all
useful at different times. An email message is accessible by multiple
paths - the date path, the topic path, the sender path, etc. With email
old-fashioned linearity begins to break down. Only begins, because
although we can have email directories, we can only have a message in
one email directory.
The sender, recipient, topic, date headers in an email message are data
about its data - metadata, and it is useful to be able to sort messages
in these ways.
The metadata in webpages serves a similar function and will let clever
Internet search-engines select webpages by such criteria as their authors,
their geographical coverage, their date of composition, and their subject
keywords. It is theoretically possible to search for webpages from the
University of Melbourne published after June 1999 on plate tectonics.
Metadata and search-engines (and the associated automatic indexing of
file contents) are some of the keys to non-linear file management. I'll
describe this a bit more.
Consider this example - you wrote a report for Freda Berk a consultant
for Mega Industries, on February 29th about your terrific new non-linear
filing system for PowerMacs - which you developed in a cooperative venture
with Moon Macrosystems.
Where would you store this file?
You might choose directories with these names:
- Correspondence - February 2000
- Freda Berk - Correspondence
- Mega Industries - Correspondence
- Reports - Non-linear Filing System
- Non-linear Filing Systems
- Projects - Apple PowerMacs
- Moon Macrosystems - Cooperative Ventures
With the old (current!) linear filing system you would have to
choose one directory for your report - unless you went to the additional
trouble of storing copies of the report in each directory - wasting valuable
In September 2000 when you want to check out the report you will hope
you remember the name of the directory where you stored it!
Clearly life would be easier if all the information about the report
was recorded in metadata - and you had a lovely search engine where you
could type in 'Freda Berk' and get a list of all the Berk-related stuff
and then re-sort it by date and topic as easily as with email messages.
Existing Microsoft applications have a sort of primitive metadata facility
- i.e. document ownership details, revision number, etc. - but to work
properly a metadata system should as far as possible be automatic and
There is another requirement for an efficient non-linear filing system.
That is interprocess communication, and application data sharing. How
would an intelligent metadata agent know to index by the name 'Freda Berk?'
Because Freda Berk will be a name in your address book as will the company
names 'Mega Industries' and 'Moon Macrosystems.' It will know to index
the term 'Non-linear filing system' because it will be the name of a project
on your Project Management System. It will index by 'PowerMac' because
that machine will be listed on your Reference Platforms document.
When metadata systems are manual people can forget to fill the information
in, when they are automatic they don't have to.
Application data sharing also means that if Freda Berk sends you an email
with a new job title you will be alerted - and your address book will
be automatically updated.
Re: The Internet on your own machine? I suspect that in the future
(say by 2010) computer application programs will fade into the background
- they will no longer be the concern of the user but of the computer's
adaptive Web-interrogating operating system and the application-developers.
Your concern will be data files and messages. Your screen won't be filled
with icons for software packages - but with icons with names such as:
that will lead you to indexes of files - and each file, in non-linear
fashion, will live in many places.
It is obvious that WindowsTM - as it now is, will not survive
this paradigm shift. Microsoft may survive if it realises that we are
heading into a future where the barriers between different software applications
and between different data formats will be eroded. This erosion of barriers
is essential for efficient information processing in the 21st century.
(e.g. What good is a brilliant metadata agent if it can't read your WordTM
document? To work properly it needs to understand how the information
is stored in the file.)
The non-linear age has begun to enfold us, and will affect the way we
think and the way we approach problems. No longer will we have to start
at the beginning and plod steadily forwards. We can sketch signposts at
any point along the path, and can return to past or future to reiterate
whatever we will whenever we will. We do not have to start writing our
book at chapter one. This added flexibility will enrich our whole mental
processing, but we will have to learn to let go of old linear habits which
(like hierarchical filing systems) cripple our efficiency. We can fly
across the plains, and then mentally flip back to tackle the walk at the
beginning - before patching together the whole journey. I welcome these
[Top of Page]
- In This Series -
(1) What Can & Cannot Be Done - The Limits of
Futurology (April 2000)
(2) The $20 Computer (April 2000)
(3) Smashing Windows(TM) - The Ascent of Non-linear Thinking
(4) Nu Plastic Yu! (February 2001)
(5) Nu Plastic Yu Tu! (April 2001)
(6) Artificial Minds? (AI Revisited) (August 2001)
(7) Video-On-Demand (June 2002)
(8) Changes (June 2002)
(9) The Implications of Immortality (June 2002)
(10) Cheating in Education (April 2003)
[Top of Page]
[Return to Home Page]
Copyright © 2000 by Michael F. Green. All
Last Updated: 18 July 2011