%
\centerline{\bf The 9th Annual \TeX\ Users Group Meeting}
\medskip
\noindent
The Montreal meeting, held at McGill University in the centre of cosmopolitan
Montreal was was the first  TUG meeting to be held outside the {\sc usa}. 
There was no campus housing, so delegates were liberally sprinkled
amongst  various downtown hotels.  Lunch and dinner were not an integral part of
the proceedings, so I didn't get the same sense of atmosphere and contact as at
European meetings.  The only real opportunities were at coffee breaks --- when
grotesque chocolate croissants were on offer, or the afternoon breaks
when we were converted to health-giving fruit juices.  Everyone would
 disperse in small groups for lunch and dinner.

The TUG meeting programme  did in any event contain other
opportunities for social contact ---  although the
pre-conference dinner on the Sunday night rather
restricted conversation to be with those sitting
immediately next to you --- all of us wrestling with
eating food in the style of Old French Canada
in 1691, armed with only a knife. The cocktails
before the dinner were more informative, but 
involved a lot of peering at lapel badges in mere 10 point
CMR, to try and put faces to the famous (and not so
famous) \TeX\ notables.

The best opportunity for social advancement with a range of \TeX ies
was the wine and cheese party hosted by Lance Carnes of Personal \TeX\ Inc.
That was on the Monday night after we had all had a chance to sample the
daytime talks and get into the swing of the conference.

There were in all around 160 people attending the conference.   There were only
a handful from Europe --- 1 from Finland, 1 from Holland, and 2 from the {\sc
uk}. If you count Israel as `European', that makes 5 in total.
The rest were North American, bar 1 from Thailand, 1 from Australia,
and a contingent of 4 from Japan.

The main theme of the conference was `\TeX\ in the Production
Environment'.  Nearly the whole of the first day (Monday)
was devoted to this topic --- 10 presentations in all.
This theme was further subdivided into 3 `slightly
arbitrary' topics of `In-House Systems', `Case Studies' and
`\TeX nical Solutions'. Taking these 10 talks as a whole,
the main pointers that emerged in relation to \TeX\ and
Production were:
{\parindent10pt\parskip0pt 
\bi Integration ---
the pros and cons of integrating \TeX with older\slash other\slash current
text production systems.
\bi
Management ---
the need for good management to co-ordinate the computing and publishing
sides, with teamwork essential to manage a system of software, hardware
and people.
\bi
Macros ---
everyone was writing their own local macros.
\bi
Databases ---
the adaptability of \TeX\ to provide a useful link with a range of databases.
Advance planning essential to provide all the hooks: \TeX\ only used at the
end of a very long process. 
\bi
Schedules ---
the effect on and analysis of production schedules:
choosing appropriate documents to process through \TeX.
\bi
Fine Tuning ---
the disproportionate time taken in fine tuning the layout.
\bi
Training ---
the great need for good training in \TeX.

}

\section{\TeX\ in Production: In-House Systems}
There were 5 speakers talking on in-house systems for document production.
Mary McCaskil from {\sc nasa} described the setting up of their system ---
starting with dedicated word processors and culminating in the present set-up
involving \TeX\ input from IBM-PCs, Macs and Sun workstations,
with preview on the Suns, laser proofing on Lasergrafix and typesetting
on Autologic.  She did some analysis of production schedules, and found
it much quicker if figures and illustrations are pasted in by hand,
rather than automated.

Time required to produce 1 page of {\it draft} output with 1--4 equations
per page:
$$\vbox{\halign{#\hfil\quad&\hfil# minutes\cr
Text (typeset)                & 22\cr
Tables (typeset)              & 73\cr
Figures (captions \& pasting) &  5\cr
Corrections (2 cycles)        & 16\cr
}}$$
(This compared to 1 hour per page for a manual, as cited by Eric Jul of OCLC.)

She regarded the overall advantages of \TeX\ as being:
quick training,
satisfactory productivity,
outstanding typography and
device independence.
And concluded that management {\it must} co-ordinate computing and
publishing, with teamwork essential to manage a system of software,
hardware and people.

The second talk  came from Tom Renfrow, also of {\sc nasa}, but from a different
division. He described the
production of technical manuals and other documentation,
all based very much on the use of databases and locally
written macros. He stressed
the need for advance planning when building databases, so
all the hooks are there, and that \TeX\ only makes an
appearance at the very end of the long process.  His
system allowed very long documents (around 600 pages) to
be turned around within a week --- and this included lots
of graphics and tables, all of which were automatically
included and automatically numbered, with
cross-referencing of section numbers (up to 15 levels of
sections).

This talk was followed by one given by David Ness of TV
Guide.  Again they found \TeX\ very adaptable for use with
large amounts of database information.  Their system pulls
out macro definitions from the database relevant to
particular documents, such as docket numbers, date or time
modified, or status.  Tools in the database can provide
summaries of all documents by purpose\slash date etc.  The
system also remaps macro definitions, sorts the numbers
and manipulates the data.

The fourth talk was by Laurie Mann (female) from Stratus
Computers. They produce families of computing manuals ---
such as on Fortran, Basic or PL/1 --- whereby each manual
has the same source text, such as the description of a
subroutine, but macros will automatically insert the
appropriate computer language name (as in Fortran)
together with examples of use, and correct number of
parameters to commands. The |\languagecase| macro for
example has arguments for each of 6 supported languages,
and can also be used to `turn on' all 6 languages, such as
for a review.  This she terms `single sourcing' resulting
in fewer unique files per manual.

The final talk on in-house systems was from Jean Pollari
from Rockware International, describing government
standard documentation. Again the stress was on lots of
locally written macros to simplify text input, and the
need for training in \TeX.

\section{\TeX\ in Production: Case Studies}
The conference then moved on to 3 case studies of \TeX\
in Production. These ranged from Eric Jul of OCLC who
described the problems of integrating \TeX\ with
older\slash other text production systems. Training was
again of paramount importance, plus the need to
differentiate between documents that would\slash could be
produced in \TeX, and those that are better produced by
other systems --- as determined by the effect on
production schedules.  He regarded 1 hour per page to
produce a manual as acceptable, but there was a
disproportionate amount of time occupied by one person
fine tuning the layout or editing the text.

At the other end of the scale of case studies was Robert
Harris from Kennel Club Yearbooks, whose main problems
stemmed from mistakes made in matching up the photograph
with the correct pedigree entry.  But again he stressed
how well \TeX\ was suited for documents using information
extracted from databases.

In between these two talks was the third case stude --- by James Mooney ---
on his experiences with textbook production.  He regarded
\TeX\ as having far greater control than \LaTeX\ (in fact
\LaTeX\ was barely mentioned in the conference as a whole,
although it did boast the largest Birds of a Feather
group).  James Mooney used separate sets of macros for
draft printouts, as compared to the typeset version of the
textbooks. The draft versions had simpler layouts and
extra spacing for reviews. His main problems related to
correction and fine tuning --- he found it hard to process
only selected pages, and that minor changes could disrupt
following pages even if the page break hadn't been
altered. He concluded that \TeX\ should become more
co-operative and allow itself to be fine tuned more easily
--- he found it particularly difficult to amend \TeX's
vertical spacing, and floating inserts were occasionally
out of sequence.

\section{\TeX\ in Production: \TeX nical Solutions}
The two final sessions on \TeX\ in Production related to
`\TeX nical Solutions'.
David Ness from TV Guide re-emerged to describe how they
had solved problems such as filling an exact number of
pages with text, or piecing streams of text together with
an output routine.  They solved one particular |\parshape|
problem by using APL to write \TeX\ to produce very long
|\parshape| commands.  Their approach attempted to convert
content driven typesetting into layout driven typesetting,
again using an elaborate live database.  They regarded
\TeX\ as an ``assembly language program for text''.

The final speaker on \TeX\ in Production was the sole
Australian representative --- he described how and why a
trade typesetter chose \TeX\ in preference to newer DTP
systems or older single supplier typesetting systems.  He
listed the commercial advantages of \TeX\ as a front end
as:  {\parindent10pt\parskip0pt
\bi
Quality ---
provides high quality typesetting and captures some of the old craft.
\bi
Maths ---
excellent for mathematical typesetting (the best).
\bi
Hardware Adaptability ---
availability on a wide variety of computers, and from a wide range of
suppliers.  Users can start with a cheap MS-DOS system and work up to
larger systems without losing any of the old work.
\bi
Drivers ---
various typesetter drivers, plus 300\,dpi.  Arbortext
drivers for Compugraphic, Autologic and PostScript
(Linotronic). 
\bi
Documentation ---
good quality\dots
\bi
TUG --- for information sharing.
\par}
The rest of the TUG meeting contained more varied talks on \TeX, plus
the usual exhibitor presentations. 

\section{\TeX\ Training}
There were two talks specifically on training in \TeX\ --- the one by
Berkeley Parks of Washington State University
concentrated on tips to get started --- ranging from
progressively moving the |\bye| command lower down the
file to isolate errors, to innovative uses of |\halign|
for complex mathematical layouts.  

The other training talk was by Alan Wittbecker, the new TUG employee with
the unenviable task of constantly travelling America teaching a succession
of 2, 3 and 5 day courses.  He listed several factors as important for new
users to gain productivity:
{\parindent10pt\parskip0pt
\bi
relevance of the commands to users' needs.
\bi
simplicity --- ease of understanding
\bi
efficiency --- a typist should understand the command --- simple and mnemonic.
\bi
consistency --- produce the same quality from different versions of \TeX.
\bi
quality (standards) --- portable between devices, style file changes, macros.
\par}
This brought in the more general topic of applying standards to all TUG
courses, so that having completed a course a student is known to have covered
certain sets of commands and topics, irrespective of where he or she was
taught.  Alan provided a suggested syllabus for each of the 5 day beginning,
intermediate and advanced \TeX\ courses.  Comments were invited.
Doug Henderson also provided a similar list for the Metafont course.

As a TUG course instructor myself, I had some initial
comments and queries. It wasn't clear as to the
flexibility of the syllabus for, say, a 5 day intensive
beginning\slash intermediate course as compared to two
separate 5 day courses.  The timescale would inevitably
mean omitting some of the topics --- it would therefore be
useful to differentiate `essential' from `additional'
topics in each course.  There was also very little maths
covered in either the beginning or intermediate courses,
whereas this is frequently the initial impetus for people
wanting to learn \TeX. From my experience, students regard
the maths as very straightforward as compared to the
intricacies of text-type layouts.  As well as providing
more maths, it is obvious that the courses should be
tailored to people's needs, particularly if it is an
in-house course.  I was also surprised to find several
topics in the proposed syllabus that were not |plain| \TeX,
such as Stephan's |\hboxr| command, which could easily
cause confusion.

Another teaching proposal was a `self-test' --- a series
of questions to help the wavering student to decide which
level of expertise he or she was at, before booking for
the appropriate course.

The one talk in the whole conference that specifically
referred to \LaTeX\ was by Shawn Farrell of McGill
University (the local organiser), who suggested ways of
choosing between \TeX\ and \LaTeX.  The usual advantages
and disadvantages of each was cited, and he concluded that
there was a place for both \TeX\ and \LaTeX\ for the right
user.  Inevitable maybe.

\section{Non-English \TeX}
Two other main topics were covered in the conference ---
Non-English \TeX\ and \TeX\ and \sgml. I unfortunately had
to miss the Japanese contingent talking about Mathematics
Textbook Publishing with Japanese \TeX. However Jacques
Goldberg's talk on Semi-\TeX\ was very amusing, talking
about the problems of mixing right to left Hebrew with
left to right maths. He runs his system on IBM-PCs and
VAXes.  On the VAX he has set up the DEC VT keyboard so it
can switch between Hebrew and English on screen, but not
display both at once. He mentioned the \XeT\ program which
needs changes to \TeX\ and the \dvi\ drivers.  He suggested
writing a \dvi\ t\dvi\ DVI convertor to allow any existing
driver to take \XeT\ output.   He is also still working on an Arabic \TeX.

Michael Ferguson then described his Multilingual \TeX\
yet again, followed by Kauko Saarinen who listed his
experiences with \TeX\ in Finland --- largely to do with
keyboard character deficiencies, especially in relation to
\LaTeX.

\section{\TeX\ and \SGML}
Moving on to \TeX\ and \SGML, Stephan Bechtolsheim
described his modifications to the {\sc emacs} editor to safely
edit \TeX\ files and translate \SGML\ into \TeX. The mods
included error checking to prevent the more trivial
mistakes, prior to running the file through \TeX.  He uses
{\sc lisp} to program in such features as inserting pairs of
braces with the cursor positioned in between, or easily
checking the nested braces for such functions as
superscripts of superscripts.  Inevitably there are
problems in that the editor is fairly complex and uses a
lot of CPU time.  It doesn't run on PCs, but other
alternative editors (called epsilon and joe) will
apparently soon be available for PCs (joe will also be
available on the Mac, and is free).

Lynne Price ran over the basics of using \SGML\ and \TeX\
to produce user documentation.  Her main aims are
markup minimisation and automatic error checking.  

\section{\TeX nical Support}
The final talks of the conference were a varied bunch,
banded together under the heading `\TeX nical Support'.
Ken Yap,
talking about \dvi\ previewers, described possible
alternatives to the typical low-res 80\,dpi screen
preview.  These included generating fonts at the required
screen resolution using Metafont, and bit sampling where
sub-samples of printer fonts are used for the screen
(e.g.~reduce each 4$\times$4 pixel block to 1~pixel). 
This cuts down on the number of additional font files
required, but needs CPU intensive raster processing. Other
partial remedies included abandoning the idea of full
pages on screen, or using higher resolution screens at say
160\,dpi, or grey scale fonts, where brightness is traded
for resolution.  The only `full' solution was a proper
hi-res screen.  The main problems with previewers as a
whole were seen to be how to preview printer resident
fonts, and that printer |\special|s weren't previewable.

Robert Kruse from Nova Scotia then outlined some Software
for Technical Book Production --- comparing the author's
perspective (flexibility, simplicity, minimal logical
markup) with that of the publisher (design, copy editing
and markup, co-ordination, page makeup and typesetting),
and also with the macro programmers' perspective (the need
to accommodate flexibility, modular macros, consistent
syntax, robustness).  He suggested a Pre\TeX\
pre-processor to simplify \TeX\ for the author, such that
quotes would automatically be converted, and the italic
correction automatically inserted.

Mike Schmidt then described his CAP\TeX, or industrial
strength \TeX, which is a menu-driven control program with
access to a spell checker, index filter etc, and specific
to technical publishing in larger organisations such as in
the aircraft industry.  Features include font controls,
predefined table templates and structures, control over
figures and graphs, mulit-column layouts and automatic
generation of tables of contents and indexing. The system
assumes many technical writers and therefore uses minimum
markup and enforceable document styles.  It uses |plain|
\TeX\ and is 99\% structure oriented.  For a slight
variation on a theme, the \TeX\ files were read {\it into}
a database.

The last speaker was Paul Muller with \FASTeX\ --- an integrated front end
to \TeX\ that uses an editor with {\sc ascii} input and output, together
with a \TeX-compatible macro system, and typesets with simultaneous
screen preview.

\section{Exhibitor Presentations}
Exhibitor presentations included a description of TEXT1 
from Washington State
University --- a package of macros including indexing and 
table-of-contents. It is available
on IBM-PCs, but must surely be usable elsewhere.  They also
have availably the International Phonetic Alphabet, as
advertised in TUGboat.

Barry Smith, of Blue Sky Research, assured everyone
that \TeX{\sc tures} was  under  stable
ownership, and that new developments were proceeding
apace.  He is moving into the graphic arts field by
providing an interactive {\it back} end to \TeX{\sc tures},
whereby typeset text can be skewed or rotated through an
interface with Adobe Illustrator --- a sort of back end
automatic paste-up on screen.  He is also extending into
colour printing on dot matrix, and full colour separation
on the typesetter.  Lucida fonts are also in the offing
(including maths)  plus all the 75 CM typefaces as
PostScript outlines --- although the results at 300\,dpi
are apparently not as good as Metafont.

Mr \PCTeX, the charismatic Lance Carnes, described the very complete selection of
software available for \TeX\ from Personal \TeX\ Inc.  There are now over 60
products --- drivers, macro packages, Metafont 1.3 and an interface to
Bitstream fonts (over 30 typeface families).  He compared
maths output using Bitstream Times with Computer Modern maths.
He also announced the PTIView Previewer.
Personal \TeX\ Inc.\ also now have a phototypesetter --- a Chelgraph
IBX2000 all the way from England. It provides 2000\,dpi, onto paper or film,
and can set any Metafont or bitmapped character.  They have developed their
own driver and will provide a typesetting service from October, with 300\,dpi
laser proofs also available.

Richard Kinch of Kinch Computer Co.\ swept the field by giving away
complete copies of his Turbo\TeX --- a Pascal translation to C for the IBM-PC,
portable to Unix.  It has a virtual memory subsystem added onto the DOS
implementation and used a swap file on disc or the RAM drive.

Bob Kaister of K-Talk Communications described Publishing Companion,
which translates Word Perfect to \TeX.  Other convertors are now also
available for other word processors as an add-on.  I believe Ewart North of
Uni\TeX\ sells
this in the U.K.  K-Talk also sell Math Editor for the PC for
{\it wysiwyg} equation generation, which outputs a \TeX\ file.

Stephan Bechtolsheim then did a wonderful job of selling himself: his master
work, `Another Look at \TeX', has now reached 975 pages, in two volumes, and
still growing.  In the absence of a publisher he sells the book at \$57.

Arbortext also came up with some new announcements.
Publisher version 1.0 was released in November 1987.
Version 2.0 is now due (August\slash September 1988).
It boasts faster performance, copy and paste between the document and
Sun Windows, a re-engineered editor and higher compatibility with `raw' \TeX.
It has \TeX\ and \LaTeX\ modes (??), with equation\slash table\slash
graphics editors, as well as integration with \SGML.
Other possible Publisher platforms should be announced this Autumn ---
this could include Apollo, DEC, HP, Mac II and OS/2.
Arbortext have also taken on Micro\TeX\ from Addison Wesley.  They intend
using Micro\TeX\ to build a second generation product under a new name ---
$\mu$\TeX.
This will include a module for an editor (based on {\sc emacs}), \TeX\ and
Preview, with simplified keyboard mappings and a streamlined installation
program.

\section{TUG European Co-ordinator}
Ray Goucher had nominated Malcolm for
the new  post of European Co-ordinator, which was agreed in committee. 
I had been requisitioned to
take his place at the committee meetings in his absence.
The actual functions of the post
still aren't clear, but are to be discussed between Malcolm and Ray.
In essence the post revolves around acting as a focus for improved two-way
communications between America and Europe, and stimulating TUG awareness
and membership in Europe --- acting as a representative and platform for
European views, whilst keeping tabs on all European \TeX\ activities.

\section{The 1989 TUG Meeting}
And now for next year\dots\ \ Dates for the 10th Anniversary TUG meeting
have now been finalised. The meeting will be at Stanford (where else)
on August 21st--23rd 1989, with the usual courses before and afterwards.
This meeting will be the 10th annual meeting of TUG. The keynote
speaker will be Donald Knuth, who will present a paper entitled ``The
Errors of \TeX''. Be there.
\smallskip \rightline{\sl Cathy Booth}