Application holy wars vs religious reformation

[HALL Bill <bill -dot- hall -at- tenix -dot- com>]

I know Earl and some Techwhirlers find the application specific crusades and
holy wars to be tiresome. However, there are some fundamentally important
reasons to technical writers beyond the selection of writing tools why the
holy war periodically hots up despite all efforts to put out the fire. I
thought Sunday was good time to try to distil some of my thoughts on the
subject.

I have used keyboards for all of my writing since junior high school, when I
first learned to touch type on a mechanical type-bar typewriter and set type
by hand. Now, at 60+, I am assembling hypertexts being distributed at the
speed of light in CALS environments.

After majoring in physics, I earned my PhD in evolutionary biology from
Harvard and taught a wide range of biology subjects for several years in
Puerto Rico, Colorado, Maryland and Australia. My last major research
project as a biologist was a study of scientific revolutions as exemplified
by the methodology of comparative biology by comparison to the
hypothetico-deductive methodology adopted from the physical sciences. I paid
for my first personal computer systems (CP/M) by running an academic word
processing bureau, and gave up teaching to earn my living as a technical
writer (software industry) - and then as a specialist on application systems
used by technical writers (banking and defence industry). As a technical
writer, my understanding of communication problems between scientific
disciplines following different research paradigms (i.e., pre- and
post-revolution) has helped me to explain unfamiliar technologies to new
users.

My thesis here is that the year 2000 is near the point of inflection in the
most significant technological/conceptual revolution in human history. We
are beginning to manage and manipulate human knowledge externally to human
brains. Word processing and DTP systems represent paper paradigms of the
past, while structured authoring environments, such as FrameMaker+SGML and
ArborText's Adept Editor enable the future. It is no wonder that users of
the respective products often see discussions of their differences as holy
wars rather than sincere attempts to deal with a changing world.

To provide a framework for my assertion that we are on the 'cusp' of the
most significant conceptual revolution ever, consider 10 of the major
revolutions in human affairs. Each revolution fundamentally changed the
nature of humans in relationship to their environments and to each other:

Technological revolutions:
o use of clubs, stones (throwing and cutting) and levers to extend human
reach beyond anatomical limits -2 M years
o taming of fire to extend digestive human metabolism < -1 M years
o use of ropes and digging implements to control and manage non-human
organic metabolism for human benefit (agriculture) -12 K years
o the industrial revolution to extend organic muscle power (human and
non-human) with mechanical power (water power, external and internal
combustion) -250 years
o industrial automation to replace human control mechanisms with cybernetic
systems -25 years

Conceptual revolutions:
o evolution of memory and learning << -100 M years
o origin of speech and teaching to effectively transfer knowledge from one
human memory to another via the spoken word ~ -750 K years
o use of physical counters, tallies, writing and reading to record and
transmit knowledge external to human memory -10 K years
o invention printing and spread of universal literacy to make knowledge
accessible to the masses -550 years
o knowledge automation to manage knowledge externally to the human brain ~
-7 years

Obviously, some of my choices may be debatable, and I may eventually have
time to write the book about what they all mean, but two trends are evident:
the earlier revolutions were evolutionary adaptations involving small
incremental changes taking place over many generations. The later changes
have been revolutionary, sufficient to change the entire ecology of the
planet within the lifetimes of individuals. 

My mother-in-law, now 87, grew up in rural Australia where horses and
shoeleather were the primary means of transport and electricity was no more
than a theory. She still insists she has proved that people can live without
the telephone, automatic teller machines, dishwashers, computers and the
like.

I learned to write when individual letters were physically hammered onto
physical paper by striking typebars one at a time with levers, and small
print shops still set type by hand in composing sticks like those used by
Benjamin Franklin. As well as learning to type in junior high, I took
printing for my 1 year shop course requirement. Today I am specifying and
implementing business systems able to automatically assemble complete
ship-sets of maintenance procedures for a class of warships from a database
of commonly used knowledge elements. This specific system reduces the volume
of text required to be managed for a fleet of 10 warships to 5 (five)
percent of the data required when we were producing sets of documents for
each ship using word processing technology.

The important point I am trying to make with these anecdotes, is that the
way we produce, manage and deliver human knowledge has changed more
radically in one decade than it has over the time it took humans to evolve
from ape-like ancestors. We are currently implementing technologies which
will soon eliminate the need for humans to be involved in knowledge
management. As such, our knowledge related technologies will change at least
as much in the next ten years as they have in the last ten.

The knowledge automation revolution was enabled by the development of word
processing and electronic typesetting (e.g., desk-top-publishing)
technologies which saved texts in electronic formats including markup codes
controlling fonts and layouts rather as physical type composed in a chase
tray or forme to produce impressions on paper. The beginnings of the
revolution began when it was realised that these electronic texts could be
automatically indexed for retrieval by applications like Verity. However,
word processing, DTP applications and Acrobat are still inextricably tied to
the document-is-paper paradigm and that the important function they serve is
to make it easier to format the information on paper.

The revolution began with the publication of SGML (ISO 8879 Standard
Generalised Markup Language, 1986). As Techwhirlers would know, HTML and XML
are direct and somewhat more specialised offspring of the SGML parent. SGML
(and especially XML) for the first time separated content from format and
introduced the concept of marking up content by its function in the document
structure independently from any formatting which should be applied to it.

The most significant technological development in the conceptual revolution
was the invention of SGML object management applications able to assemble or
parse and decompose virtual documents from or into independently managed
elements of marked up text and other document objects at any arbitrary level
of document structure. A closely parallel series of developments was also
taking place in the engineering world, with configuration or Product Data
Management (PDM) systems based on STandards for the Exchange of Product
information (STEP), which linked product related information and texts to
conceptual components in a 3D hierarchical model of the product structure.
The combination of such tools with automated workflow management systems and
document assembly processors provide applications the ability to make
decisions and trigger actions based on textual content of particular
documents. A further step is to build artificial intelligence into
'documents' such as Interactive Electronic Technical Manuals (IETMs).

XyEnterprise's Parlance Document Manager and Sherpa's PDM system seem to
have been the first truly commercial implementations of object management
systems to manage the content of virtual documents, and these have been
followed by a number of other systems with similar functions in the
engineering, documentation, and eCommerce areas. As noted in a couple of
contributions last week, my organisation is currently implementing RMIT
University's SIM system - which I believe represents today's state of the
art in such technology.

The application wars between MS Word and FrameMaker users in this space
result from the great conceptual divide between word processing and DTP
products - conceived to make it easier for authors to chip words into stone,
vs tool kits designed to help authors create and visualise the structural
framework of documents to make them work better in an increasingly automated
hypertext world. 

Obviously, if a hypertext author fully understands how his/her text is to be
used, any kind of tool that captures the text electronically within the
appropriate markup codes can be used. It is just that the task is made
easier if the concepts of structured authoring are built into the tool.

In any event, my study of scientific revolutions shows that most scientists
involved in them never really understood what is going on. To many on
opposite sides of the divide, the arguments were holy wars over incompatible
dogmas. Even though they are supposedly rational scientists, their beliefs
only shifted via processes akin to religious conversion. Thomas Kuhn's book,
"The Structure of Scientific Revolutions" first made clear the underlying
issues that resulted in holy wars rather than rational discussion and
reform.

On the other hand, if you understand that the problem lies with
incommensurably different fundamental concepts, and work to consciously
understand these fundamental concepts in their own terms, it is quite easy
to understand and rationally work with what seems incomprehensibly holy or
mystical to many.

Where tech writing is concerned, Whirlers will increasingly be capturing
knowledge in forms suitable for automated processing and delivery. Display
formatting is likely to become a completely separate speciality from
writing. Tools like FrameMaker+SGML, Adept Editor and BladeRunner already
cater for this requirement to some degree, but all are still made much more
complicated (and expensive) than they need to be by their unavoidably
associated formatting functions. 

What is needed is for a tool maker to come up with a simple to use
browser-like editor that fully understands content markup; and to provide
independent systems able to implement structure related style templates for
the viewing browsers. W3C is obviously going down this track as far as the
standards are concerned, but I have yet to find a structured authoring tool
that is actually easy to write in. Adobe is well placed to turn FrameMaker
into such a tool, but they apparently don't have a particularly clear vision
of the future either.

As far as the application holy wars are concerned, I think we will all
benefit from rational discussions of how fundamental differences between our
tools affect our writing activities and the management of our written
products, and spend less time cursing the other religions. (Which won't stop
me from cursing monopolistic dictators who foist broken tools on
unsuspecting users.)

Bill Hall
Documentation Systems Specialist
Integrated Logistic Support
Naval Projects and Support
Tenix Defence Systems Pty Ltd
Williamstown, Vic. 3016 AUSTRALIA
Email: bill -dot- hall -at- tenix -dot- com