\def\wysiwyg{{\it wysiwyg}}
\bigskip
[The following article is the text of a presentation
given at the Royal College of Art on December 8th, 1988.
The theme of the meeting was `Training for
Desktop Publishing', and it was intended as a vehicle to
inform `management' about some of the features of dtp
which vendors rarely reveal. In the event, many of the
other speakers were representatives of various software
and hardware vendors. But having written it, it seemed a
pity not to put it out into the world. Some of the
article comprises fairly straightforward background
material. I believe that this is an area which we seldom
address well (myself included). So many people have a
totally inaccurate view of \TeX\ and \LaTeX. Just glance
at what the computer press has to say about \TeX, should
its denizens ever deign to review it: clearly journalism
requires neither research nor serious application. But as
long as we fail to reach out and communicate, the fault
is as much ours. We have nothing to lose except \TeX.]
\bigskip
%Arthur Keller, 1985. 
% %The use if \TeX\ offers not only the potential for
%producing material of the highest quality, but also the
%possibility of producing unreadable or un\ae sthetic
%documents. Elegant typesetting is not a substitute for
%clear, refined prose. 
\let\big=\bf
\centerline{\big\TeX, training, and a few other windmills}
\section{\TeX: background}
Let's first of all establish what \TeX\ is, and how you
use it. \TeX\ is a computerised typesetting system which
has been under development since about 1978, and which
ceased development in 1982-ish. Any work done since then
has been purely bug fixes.

It was created by an academic, Professor Donald Knuth of
Stanford University. Knuth was dissatisfied by the
quality of typesetting which his publishers were managing
to achieve, and basically, he decided to do it right.

His own books are largely mathematical in nature, and he
gave a great deal of attention in \TeX\ to setting
technical text. Even now, there is no computerised
typesetting system which can hold a candle to \TeX\ for
setting mathematical equations. Some of us would be bold
enough to argue that its typographic excellence goes
beyond mere mathematics. And this despite the fact that
the whole of \TeX\ resides in the public domain. Anyone
can adapt or adopt its algorithms (or even its code) and
make them part of their own product. There are a number
of commercial organisations who will supply their own
particular implementations of the software, and who will
provide the backup services normally associated with
software suppliers -- help, advice, consultancy,
teaching, etc.

\TeX\ was developed on a Digital TOPS-20 system at
Stanford: it was developed on a mini-computer with
probably less power than an Macintosh II or a 386 pc.
Despite the fact that it started out on traditional
multi-user computer systems, it was sufficiently well
written that it could be transferred to all sorts of other
machines.

Today \TeX\ runs on almost every flavour of computing
engine. The minimum requirement is that the machine
supports 16-bit arthmetic. This encompasses the IBM pc
(in all its variations), the Apple Macintosh (in all its
varieties), the Atari ST, the Amiga, and  even
more obscure personal machines. \TeX\ is not fussy about
where it lives. The micro
versions of \TeX\ are not cut down versions of mainframe
programs. They are functionally equivalent in every detail.
\TeX\ not only runs on a variety of
machines, it produces identical output for identical
input, no matter what machine you choose -- no matter
what output device you choose. \TeX\ has been driving
phototypesetters and laser printers since 1978. 

But \TeX\ isn't particularly a desktop publishing device.
To a user of \TeX, using a Cray supercomputer, the
desktop would have to be very strong. To another, using
his Macintosh at home, it would look pretty average. But
both would have the same facilities at his or her
fingertips. This gives \TeX\ users a very mobile quality.
When they move from machine to machine they don't have to
relearn \TeX, although of course they might have to
relearn the details of the operating system and their
editor. \TeX\ would probably rather be called a document
preparation system. It's really designed for producing
books, scientific papers, and other paraphanalia of the
academic world. Despite that, it is also used to produce
manuals, newsletters, magazines, the odd letter and quite
a lot of lecture notes.

Given that it was designed by an academic, with rather
academic purposes in mind, it isn't altogether surprising
that \TeX\ spread to other people in universities. In
general they each had a fairly specific need: they wanted
to be able to type in their own work -- why did they
want to do it themselves? Not really out of any feeling
of altruism, but simply because typing scientific
material is often fraught with errors for the uninitiated
(always assuming that you could read the handwriting). A
typist who does not know that this squiggly thing is a
$\zeta$, and this one is a $\gamma$ may have difficulties
distinguishing $\sum$ from $\Sigma$. One way out of the
fairly inevitable  revision/edit cycle is to do the job
yourself.

So the spread of \TeX\ is largely a grass roots one. It
has spread because there was nothing else which was
available (at least, not generally). It has spread
amongst a sector of the population with rather limited
spending power, and with a distinctly unglamorous public
image. There has been no media hype to tell us how
wonderful \TeX\  is, and how we can't possibly manage
without it. Despite its unparalleled excellence in
setting penalty copy, you don't see it featured in
any expensive Seybold Reports. Nobody is getting
rich out of \TeX\ (well, hardly anyone -- and certainly
not Knuth).


\TeX\ is also unfashionable in the sense that it is not a
direct manipulation \wysiwyg\ system. You type page layout
commands within your text, whose effect you will not see
until you run the text through the \TeX\ program. Now
this is quite an interesting contrast to the current wave
of dtp systems like Ventura, Pagemaker, Fleet Street
Editor, and so on. If we examine the roots of these
systems we can see that we can identify two
starting points. There is the typing root, and there is
the typesetting root. Of course, the typewriter was
designed originally to emulate printing (which itself was
of course designed to emulate the work of scribes and
copyists). Nevertheless, we can note immediately that the
notion of `what you see is what you get' is deeply
embedded in typewriting, but that in the typesetting
tradition, only those who could read intaglio mirror
images of signature printing could guess what they
were likely to get. Both traditions have co-existed for
some time. 

There are many who would argue that it is not only
irrelevant but positively distracting to see the layout
of your text as you type in. Many believe that content
and structure are inherently distinct. Certainly, in
typesetting a book, it would be rather irritating to
have to recall the exact details of the layout of each
chapter opening. 

Before I finally get down to saying something about
training, let me quote the following from Arthur
Keller, one of Knuth's associates: {\sl ``The use if \TeX\
offers not only the potential for producing material of the
highest quality, but also the possibility of producing
unreadable or un\ae sthetic documents. Elegant typesetting
is not a substitute for clear, refined prose.''}

\section{Training}
My own experience with \TeX\ over 5 or 6 years indicates
that there are many forms that training can take:
\smallskip\noindent
$\bullet$There are people who will use \TeX\ without any
training whatsoever. Armed only with the book (\TeX\ has
a rather good 500 page book which goes with it) they will
quite happily produce their thesis or book. This level of
sophistication is no challenge to them at all. They
probably have more difficulty with the mechanics of typing
than any other aspect of the problem.
\smallskip\noindent
$\bullet$There are many who learn from others. It seems
that once you have a `critical mass' of users, they
become self sustaining. Individually they may not know a
great deal, but by pooling information, they solve most
problems.
\smallskip\noindent
$\bullet$Many people like to go on courses. Those who go
on a course with a problem to solve, like a thesis to
write, assimilate the information much more quickly than
those along for the ride. Directed learning has no equal.
\smallskip\noindent
$\bullet$The best courses allow people to use the system.
Now this actually presents a problem, since once you
encourage people to sit at a micro or terminal, you have
to assume some level of competence of keyboard and
computer. In my own courses, this has been the single
biggest problem. I teach \TeX\ on IBM PCs, in an
environment where we also have multi-user
mainframe machines. Mainframe people, unfamiliar with
micros can find them awkward and unfriendly. Some
positively relish learning a new system, other dread it.
Given that we are covering new material, we cannot afford
to let people be left behind because they can't use the
editor. Lose them at the beginning, and you've lost them
for good. Another consequence of practical work is that
classes have to be reasonably small -- you can't spend
ten minutes with each member of a 20 strong class. And
practical work tends to be slow. 
\smallskip\noindent
$\bullet$Lets dispell the myth that
says that you can't teach something this complicated to
secretaries and typists. What nonsense! This is
usually the refuge of some academic (read `manager') who
doesn't really understand what is going on, feels he/she
ought to, but does not want to be upstaged by people to
whom he/she feels intellectually superior. But courses
must be tailored to their audience.

\smallskip\noindent $\bullet$People who want to do
training  usually do better than those who are drafted. 
\medskip
\section{The windmills}
There is much more to `publishing' than merely putting
the words down on the page. The whole history of
professional publishing demonstrates that the
`origination' side of things, although not without its
own special problems, is only one part of process. Within
the desktop and electronic publishing world the following
points may be a made.


\smallskip\noindent
$\bullet$Producing the typeset
quality output is usually the beginning of the problem.
This requires amplification. Using the nuts and bolts of
some system can almost always be mastered, whether
intuitively, by rote, or with some understanding. The
mechanical elements seldom present an overwhelming
problem. What presents the problem is, to paraphrase
Beatrice Warde's words, `making the type invisible'. As far
as technical typesetting is concerned, the type ought to
be transparent. It ought not to be a barrier to
understanding. That's one reason that technical texts tend
to be set in a rather bland variety of a `modern' serifed
typeface. It's what we are used to. If we were preparing
advertising material, our objectives might be rather
different. There is nothing competitive about a maths
book. The competition was between `shall I read the book
or dig the garden', not `which article shall I read
first', as might be the case in a magazine.
\smallskip\noindent
$\bullet$Insufficient attention is given to design. We
all seem to think we know what books should look like. Our
level of criticism seldom advances above `I know what I
like'. \TeX\ has made some attempt to tackle the design
problem, by having a series of highly structured
templates which will guarantee generation of an
acceptable design for books, articles,
reports, and so on. Not one of these could be described
as exciting, but we are not in the business of exciting
by our design. We are out to excite and inflame by the
contents. Nevertheless, I still see students determined
to emulate the typewriter. If you are preparing a thesis,
the regulations do tend to assume that that is the
medium, so there is some excuse.
\smallskip\noindent
$\bullet$Where are the typographers? One of the real
reasons that much desktop publishing is excerable is that
typographers have been shirking their responsibilities.
They probably were not involved in the desktop revolution.
The apocryphal tale is told that Hermann Zapf (of
Dingbats fame), who designed Optima, was not even asked by
Adobe for his comments on their digital version. But to
many typographers, and I'm using the broad term which
encompasses type designers as well as the people who are
concerned with the placement of the mass of type, the
electronic revolution seems to have induced panic, rather
like a rabbit fascinated by a car's headlights, and they
seem unable (with a few notable exceptions) to communicate
with us, except perhaps to tell us that we're doing it all
wrong. 

That typographers had nothing to do with the design
of the Macintosh interface is manifest just by the
Underline fonts, or by some of those dreadfully unreadable
Shadow fonts. Even when the Linotype fonts  on
the LaserWriter are used, the `base' set universally
available encourage people to mix Times and Helvetica as
their `serif' and `sans serif' fonts -- not a happy
pairing (it is even worse if you need to use a monospace
font, since you will then almost certainly be using
Courier -- the three do not blend well).

The total lack of clarity in font sizes on the
Macintosh is another example of control by people embedded
in a typewriting approach.  Why will packages allow you to
use  excessive measure, easily? They should allow you to
do it, but only by perseverance. Almost every DTP package
should carry a government health warning!
\smallskip\noindent
$\bullet$Since we've got managers and decision makers
here, let's add that you shouldn't be taken in entirely
by the claims of the package. In the hands of an expert,
the package, whatever it is, will produce stunning
results. Why then can't your staff achieve the same,
instantly? Obviously you know the answer to this, or you
wouldn't be here. But there is more to this than mere
training. Any technological changeover leads to some
reduction in quality. When Gutenberg started running off
his 42-line bible about 1460, there were many who thought
that quality had been compromised (we don't
actually know when the first book was printed by
Gutenberg, or even when the first bible was printed).
When hot metal typesetting was ousted by film setting,
many felt that standards were in decline (like Donald
Knuth). The truth of the matter is that standards do
decline, because there needs to be a build up of
experience, practice at doing the job. With any luck,
this doesn't take too long. 

A major drawback in the
computing world is the distressing frequency with which
new versions of the software appear. Usually they have
fixed a whole host of bugs, and introduced a whole host
of new ones. You will also have a range of new features
which no-body knows how to use (or even what they are
really for). It is probable that the first versions of
PageMaker and Ventura (and \TeX\ too, for that matter),
were rather limited.  From a software suppliers point of
view, in a highly competitive market, enhancements are
needed to help attract media attention, and to keep the
product firmly in focus. 
Fortunately from my point of
view, \TeX\ is out of this vicious spiral. I can use it
instead of relearning it.
\smallskip
\rightline{\sl Malcolm Clark}