\def\CJ{{\sc Cajun}}
\title{Malcolm's Gleanings}
\author[Malcolm Clark]{Malcolm Clark\\\texttt{m.clark@warwick.ac.uk}}
\begin{Article}

\section{Macsyma}
A brochure for Macsyma arrived the other day. On the back page of this
multi-colour leaflet was the statement that `Macsyma's math
expressions look just like those in textbooks'. I hear some of you
already `that's hardly surprising since Macsyma can output in \TeX\ 
format'. Well, yes it can, but what the advert was extolling was its
ability to use MS-Write (`which comes with MS-Windows') to create
`screen displays of large expressions', at which it `excels'. By this
time you will have worked out that I wasn't impressed by the example
they give. If I make a list of the infelicities that were displayed
you'll think I was making it up. If Macsyma thinks that textbooks look
like this it is clear that standards of literacy, mathematics and
attention to detail have declined irredeemably. I may have to retire
to Tunbridge Wells.

\section{As others see us}
In the production notes accompanying the Acrobat in Publishing
booklet, \LaTeX\ (or Latex) is described as a `mark up text processing
package'.

Rosemary Bailey pointed out to me that in a report entitled `The
Scientific, Technical and Medical Information System in the UK',
prepared on behalf of The Royal Society, the British Library and The
Association of Learned and Professional Society Publishers, \TeX\ is
defined as `A mark-up language, similar to SGML, compiler and output
software, first developed by the American Mathematical Society (AMS),
for complex mathematical papers.  The files contain standard ASCII
characters, and can therefore be transmitted over simple computer
networks.  Has the ability to cope with all mathematical symbols and
can provide high-quality output.'

Rather similarly, Allan Reese remarked in the UKTUG electronic digest
that when his letter to the Daily Telegraph was published, `\TeX' had
been changed to `a text based publishing system'.


\section{Colour}
I can't really see why there is all this fuss about colour. None of the
publishers I have spoken to show much enthusiasm for \LaTeX's (sort of) new
ability to allow us to place colour on the page. There are few enough STM
(Scientific, Technical and Medical -- the nomenclature for most
\TeX-friendly publishers like Elsevier, Wiley, etc) books with half
tones let alone full colour. Colour is expensive. In my own
institution the cost difference between a colour page and a monochrome one
is a factor of 12 or so. Fine for the odd page, but hardly enticing.
It is true that colour has its place: that place is usually a full
colour photograph or two in, for example, medical books, or the dust
jacket or cover of a book. Few journals routinely print colour, and
when they do, they often levy an additional charge.

So why are people so excited? There are possibly a couple of reasons, and they
have nothing to do with paper. The first explanation has been with us for a
long time. \SliTeX\ has always given the capability of creating slides in
colour, although I have never knowingly witnessed slides created in colour by
\SliTeX. But given current technologies, colour slides make some sort
of sense in two presentational forms: the traditional overhead slide;
and through one of these panels which allows the contents of a screen
to be projected. In the first case we still have the hassle of
printing, but in the second we have something potentially quite
useful. Arguably the screen is the cheapest colour system around. It
is quite expensive in capital terms, and its resolution is not great
(always a worry with Computer Modern which is deficient at 300\,dpi,
let alone 72--80\,dpi), but at least the colour rendition is `what you
see is what you get'. Once you go from screen colours to printed
colours you are in trouble, and things seldom come out as
convincingly. This is a well known problem, although there are few
well known solutions. But stick to the screen and you have something
quite convincing. This was brought home to me the other day when I
obtained the notes for an Acrobat seminar in paper and electronic
form. The paper form was in black, grey and white. The electronic form
was in colour. Here was an instant example of added value to the
electronic form -- at no real additional cost. What fascinates me most
is that this is essentially subversive. Recall that \TeX's clarion
call was to produce masterpieces of the publishing art: it seems clear
to me that publishing here implied publishing on paper: what we may
have is a reason to stick to electronic form because the paper form is
less rich -- and indeed, can never be as content-full.

One of my reasons for querying the usefulness of colour was the
arrival recently on my desk of {\em Cahiers GUTenberg}, the journal of
the French (speaking) \TeX\ group. This long-awaited volume (lateness
seems endemic in
\TeX\ based productions) is a colour issue, and goes into many of the details
surrounding PSTricks, Seminar, and colour in general. While I appreciate that
if you talk about colour, you really have to use it as well, I find the
appearance of the volume similar to the early days of DTP, when the naive had
just discovered fonts. 

\section{Acrobat again}
\subsection{Acrobat in Publishing}
By now you will have realised that I think that Acrobat is a good
thing. By and large my enthusiasm was vindicated at the Acrobat in
Publishing seminar held in London on May 16th at the Society of
Chemical Industry in Belgravia, London. It was there that the penny
dropped that there are advantages to including colour.  The meeting
took the typical form of a number of lectures touching on various
aspects of Acrobat technology and a few hands on demonstrations. In
addition we were given a 40-page hand-out of the talks, and a Mac or
MS-DOS floppy with the text in Acrobat format. One of the things which
the meeting lacked, although it was targeted `for people in the
publishing and printing industries' and was claimed to `explain what
Adobe Acrobat technology is', was much explanation of the components
of Acrobat. The notes did explain, but it might have saved much
confusion if an initial talk had just explained some of the buzz-words
and jargon which was about to be unleashed.

The first speaker (a replacement from the published programme) was
from UK Mail International (part of the group who publish the Mail,
the Mail on Sunday and the Evening Standard). To my surprise they turn
out to be very committed to electronic publishing (I really must stop
underestimating the right's ability to utilise technology:
Conservative they may be; conservative they are not). One proposal was
that we may expect to see compilations of back numbers of newspapers
on CD-ROM in Smith's (in Acrobat format). The less charitable
whispered that they couldn't imagine wanting back editions of the
Mail, but the point here was that for hardly any additional cost, the
newspaper proprietors have another product. Almost equally, one can go
from that and perhaps have the latest edition transmitted to you
electronically. They were talking in terms of a typical issue of the
Evening Standard taking about 10 minutes to download over a 64kbps
ISDN link. They also suggested that there could be added value by
including video clips (for example of a winning goal), which are
clearly not possible in the `standard' version.  What was never really
suggested was that I might just want specific stories, so that I might
make up my very own newspaper (this is a suggestion which has been
around for some years), or that the format that I might wish to read
on my screen might be different from the one that the sub-editors had
determined. At least it will be marginally easier reading a tabloid on
screen than a broadsheet (but no more easier to print).

Rosie Altoft from John Wiley discussed some of their experiences with
Acrobat.  Wiley used Acrobat in its beta development days. As you
should know, Wiley has been using \TeX\ for many, many years, and is
arguably one of the most electronically aware of our national STM
publishers. Part of Wiley's experience with Acrobat derives from their
association with the \CJ\ project, which itself is part of the fruit
of the journal EP-odd (Electronic Publishing: origination,
dissemination and design\footnote{Those wondering why the name EP-odd
  was chosen might reflect that the Electronic Publishing conference
  has a tendency to take place every two years, in even numbered
  years: a sort of EP-even.}) which they publish. \CJ\ and EP-odd will
recur in this report. I was not especially clear how much of an
advantage Acrobat was to Wiley's. Since they have been dealing with
electronic submission for years, any advantages seemed rather
incremental rather than revolutionary. The most exciting thing she
suggested was the ability to cut down the amount of time and reworking
involved in changes to cover design -- chiefly through the addition of
the `sticky notes' (or Postit notes) feature. But she did note that
their New York office had been involved in an experimental scheme to
allow college lecturers to create their own course material by
selecting chapters from a wider range of books, and having them
printed up into the course book, again through Acrobat.  This embodies
selection and print on demand (or, at least, very short run printing)
-- things which are definitely in the pipeline.

Philip Smith of Nottingham University Computer Science Department,
{\em Using Fonts in Acrobat} added to my sum total of Acrobat
knowledge in a number of ways.  Basically he was describing how
Acrobat handled fonts, but not harping on about the Multiple Master
Technology, which he rather took as read. At present Acrobat can
handle both \PS\ Type~1 and Type~3 fonts, and TrueType fonts. In
passing, in an earlier edition I suggested that Minion and Myriad had
been renamed Adobe Sans and Adobe Serif. This appears not to be the
case. I have at least one document which has Minion, Sans and Serif.
Philip cleared up one point which had been worrying me: how does
Acrobat handle non-Latin fonts? Basically it embeds. Embedded fonts
are those which, for some reason, Acrobat decides to include with the
document, so that rendition is possible. How does it decide? It will
not embed the `standard 14': these are Times, Helvetica and Courier in
their four variations, plus Symbol and Zapf Dingbats: it will embed
fonts which do not use the Latin (ISO Latin~1) character set: it will
always embed Type~3 fonts; all others will be approximated through
Multiple Masters. Almost. You can force the `real' fonts to be
embedded if you use Distiller (one of the Acrobat suite). You will
appreciate that there is a legal issue lurking in here. Can I legally
distribute an Acrobat file which contains an embedded font which has
been licensed to me, but which may not have been licensed to the
recipient? Firstly, all Type~1 fonts in the Adobe Type Library (which
may include those licensed from ITC, Linotype and Monotype) may be
`freely' embedded in Acrobat files (that's a bit woolly to me);
secondly, Adobe considers its encryption to be good enough to prevent
the unscrupulous from extracting the fonts and using them for other
purposes. But beyond that you do run the risk of violating copyright
law.  The other piece of key information which Philip gave was how to
obtain the information about which fonts Acrobat is using for a
document. Hold down Shift$+$Ctrl (on Windows), or Shift$+$Option (on a
Mac) while selecting the {\sl Document Info} item from the {\em File}
menu. A slight catch is that this is a running total: you either have
to view the whole document page by page first, or do a search for a
word which doesn't exist (that forces processing of each page).
Another catch is that if you are viewing your second or third
document, their fonts will also be listed.

After lunch, Ian Chivers of Kings College London discussed the use of
Acrobat with Ventura. He was concerned with its use in an academic
institution, and particularly for the production of large multi-author
documents.  He noted that Acrobat (in common with other Windows
products) was resource hungry, requiring something of the order of a
33\,MHz 80486DX with 8\,Mbyte of memory for serious work. A Distiller
run on a 40 page document took 4 hours on a 20\,MHz 80386SX with
5\,Mbyte of memory.  I was interested to see that Ventura was taking
Acrobat quite seriously to the extent of providing the hooks to
generate `bookmarks' for tables of contents and indexes, which Acrobat
can subsequently use.

Leon Harrison (again of Nottingham's Computer Science Department) described
\CJ. This acronym stands for {\em CD-ROM Acrobat Journals Using
  Networks}\footnote{The acronym \CJ\ had been established when
  Acrobat was actually called Carousel, hence Carousel Assisted
  Journals Using Networks, but what's in a name?}. It is in fact a
collaborative venture with John Wiley \& Sons and Chapman \& Hall. A
couple of interesting features emerged in this talk. The major product
is EP-odd on CD-ROM. EP-odd is archived in \LaTeX. The text remained
constant in the archive, but the macros evolved. However, they were
not themselves archived, and when it came to rerun the articles,
discrepancies became apparent. The extra value which Acrobat form can
add includes links between documents (or to the table of contents,
etc) which can be embedded in the \LaTeX\ macros. There are some
difficulties, since forward references require that they know exactly
what point they are to refer to (a common enough problem in \LaTeX,
solved through the {\tt.aux} file, but requiring some more subtle
maneuvering in the \PS\ which will become Acrobat, apparently).
Line art had been redrawn and discarded, requiring some scanning in
from page proofs. One of the other journals in the project is {\em
  Collaborative Computing}, which is re-keyed into 3B2, although
authors may submit in \LaTeX\ (given the algorithmic similarity
between 3B2\footnote{A Santa Barbara beer to the first person to give
  me the correct explanation for this name; revealer collects, of
  course.} and \TeX\ this is a bit sad).

The last paper was from the urbane Conrad Taylor, who discussed some
of the design issues which were highlighted by Acrobat. He made a
number of points on displaying documents which apply quite widely: you
can seldom display the whole page and read the body type (I can,
because I have an A4 screen and {\em Textures}, but not all the world
is blessed in this way -- however, Conrad was actually talking about
newspaper formats here, and only would-be newsletters use A4 format,
so he's right), and therefore have to zoom and scroll. This becomes
tedious. If the document has colour you need 24-bit colour support.
Rendering and redrawing can be time consuming, especially for
graduated tints and complex vector mapped graphics. If it is indeed a
newspaper, the large size prohibits hard copy at the size for which
the pages were designed. The diagrams have a level of detail
appropriate for litho printing, but not for the screen.  He made the
observation that it would be more effective to reformat a newspaper
before distributing it in Acrobat format, to take account of some of
these difficulties. Conrad went on to give an example of designing for
paper and the screen. Admittedly, the example he provided will come as
no great surprise to (\La)\TeX\ people, but it is interesting to see
how far his typesetting tool, FrameMaker, has come. Basically he
employed some generic  and was able to take the same marked up
document to produce a screen oriented version and a paper-oriented
version. What made this interesting was that he did the conversion
live, and that Frame supports similar tools to Ventura to allow the
implanting of useful links to support table of contents and other
navigation aids. Of course, when I say `tables of contents', I don't
just mean that they exist, I also mean that they are electronically
linked to the sections to which they refer. It was most agreeable to
see Conrad defending and promoting the use of generic markup.

All in all, a most useful meeting, attended by close to 100 people.
The venue was good, with excellent facilities both for the social end
of the meeting and the presentations. The group might usefully
consider using this location, if we can fill it!

\subsection{pdf or dvi?}
It was at this Acrobat meeting that I started to wonder if Leslie
Lamport's notion some years ago that \TeX\ should produce \PS\
rather than dvi was not correct. I had always rejected this notion,
partly on the grounds that \PS\ was a proprietary system, that
many non-\PS\ printers were out there, and that \PS\
screen previewers were few and far between. Well, \PS\ is hardly
proprietary any more: there are so many clones, and the details have
all been published; there are still lots of non-\PS\ printers,
but the availability of GhostScript for all the main platforms (Mac,
Unix and pc/Windows) means that this is not a complete barrier.
Similarly, the use of GhostView allows \PS\ to be viewed on the
screen (invaluable for those pesky EPS inclusions). The use of
GhostScript and GhostView does involve an extra step, but could impose
a degree of standardisation which could save much effort. I would
argue that the \LaTeXe\ support for graphics is almost exclusively for
\PS\ graphics, acknowledging the pre-eminence of this system for
serious work.

But selecting \PS\ as the `ultimate' output format does not go
far enough. It should be Acrobat (or more correctly, portable document
format, pdf).  (\La)\TeX\ should produce pdf. Adobe already produces
Acrobat viewers for Mac, DOS and Windows. At present they make a small
charge, but I'm fairly confident that they will soon be part of the
operating system, or given away with so many applications that we can
assume their ubiquity. I was given a Seybold CD-ROM in Acrobat format
with an Acrobat reader for Windows at the seminar. I already have one
for the Mac which I was given at the launch of Acrobat in London last
year. 5D Solutions is producing a freeware Acrobat reader for Unix.
One of the advantages of the Acrobat reader is that it will allow you
to print to \PS\ and non-\PS\ printers -- and if the
document has been created by Distiller, that means that your embedded
EPS will also be printed out. In other words, we have a \PS\
interpreter in software (just like GhostView and GhostScript). I've
already commented on Acrobat's font substitution. Acrobat supports a
hypertext framework (pdfmark) which allows navigation through the
document. As yet it does not support intra-document links, but that
may come in time.

If the NTS (New Typesetting System) project has any imagination, it
will see beyond the narrow confines of creating a system to create
even finer masterpieces of the publishing art and will eagerly embrace
the technologies present here to create a system for practical
examples of the {\em electronic} publishing craft. The opportunities
are there. We can only hope their minds are not yet closed.

\subsection{Size isn't important}
What is the difference in size between {\tt.tex} {\tt .dvi}, {\tt.pdf}
and {\tt.ps} files? I compared only one file, a draft of the one which
contains this column. There are a number of things to watch. 
A {\tt.pdf} may contain embedded files, which will obviously make it
larger. I used Blue Sky's Type~1 Computer Modern in my preparation. In
theory it should not have been embedded, and my checks indicate it was
not (I viewed it on another platform which does not have these fonts
-- in fact, which does not have \TeX\ on it).  The {\tt.pdf} figure is
from using pdfWriter, not Distiller. I would expect Distiller to
produce slightly more compact code.  The {\tt.ps} figure is from dvips
on a Unix box. \PS\ is a notoriously difficult beast to tie
down, since what you are probably measuring has more to do with
optimisation decisions made by the drivers' authors.

\begin{center}\footnotesize
\begin{tabular}{|lr|}
\hline
\multicolumn{1}{|c}{file}&\multicolumn{1}{c|}{bytes}\\
\hline
\tt.tex&32682\\
\tt.dvi&46728\\
\tt.pdf&103248\\
\tt.ps &115412\\
\hline
\end{tabular}
\end{center}

\section{Editor nods}
Neither our revered and esteemed editor, nor John Bowsher, need lose
sleep that the \TeX\ logo is restricted by Knuthian fiat to Computer
Modern. As long ago as 1986 (\TUB, 7(2), p.101) Knuth had recognised
that the kerning and lowering amount for the logo were font specific,
even within CM. He went on to say `the plain \TeX\ macros are
specifically oriented to Computer Modern fonts. Other typefaces call
for variations in the backspacing, in order to preserve the logo's
general flavor'.  He then goes on to note that he has typeset the logo
in a variant of Times Roman for his {\em Computer Journal} paper `and
the standard \verb|\TeX| macro worked fine.'

This seems to suggest that (a)~Knuth had long ago realised
the problem, and (b)~he does not feel that the
\TeX\ logo should be restricted to CM (sigh).

\end{Article}
\endinput
\section{Offizin}
Whenever I pontificate about publishing with
\TeX, someone will always bring me to earth by pointing out that the
proceedings of the 1988 \TeX\ conference in Exeter took an
interminable time to hit the bookshops.  The figure is about two years
(I was busy\dots).  It was therefore a pleasant relief to receive {\sl
  Offizin} a few months or so ago.  This is a production of {\sc
  Dante}, the german-speaking
\TeX\ group. It is a publication designed to disseminate some of the
lectures given at the group's `\TeX\ days'. I worked out just when I
presented the paper which is produced in translation: it was February
1991. That makes the \TeX88 book look much less laggardly! Of course,
what I had to say, about {\em \TeX\ in Europe and America}, is
hopelessly out of date, but when it appears in my list of
publications, no-one will know that!

Putting this schadenfreude aside, it is an
interesting volume. It should be the first in
a series, a series published by Addison Wesley
(Germany). According to other bits of Addison Wesley,
they don't do conference proceedings, so someone did
some fancy footwork to get this through. Well done.

One quote I managed to extract was `typography
has its experts, but they have no audience'.



\section{Despair?}
Has \TUB\ sunk? It is now May 23rd and no sign has been seen of the
first edition of 1994. When last year's final copies came out more or
less on time I had supposed that it had finally managed to get its act
together and was to be produced on a regular and reliable basis.
Clearly I was deluded. What is the problem? I refuse to accept the
usual story that it is a complex journal and that to achieve the
standards required the devoted and underpaid or unpaid editorial
volunteers have to devote limitless time and energy to it. \TUB\ is
dying at the altar of quality. If the journal is to have any
credibility it has to come out regularly. Maybe it really is too
complex and \TeX\ is not really up to the production. Commercial
publishers -- to whom we direct much encouragement to use \TeX\ --
could not allow themselves to be sucked into this cuckoo's nest.  TUG
has to try to be realistic and trim the sails of \TUB\ so that it can
leave port. There are enough enemies of TUG, inside and outside the
user group, who wish to see it dismembered, and who do not need to be
able to point to \TUB\ to see graphic demonstration (or
non-demonstration) of the health of the whole organisation.
\section{Euro\TeX}
The publicity for the forthcoming Euro\TeX\ meeting in Poland has
dropped the short passage which described last year's Aston TUG
meeting as one of the Euro\TeX\ series. You may wonder why Aston was
not a Euro\TeX\ meeting. After all, it was a \TeX\ meeting in Europe.
There is no body which chooses a site for Euro\TeX. It has been a
voluntary and piecemeal choice which seems to have worked, to some
extent. When I organised the meeting in Exeter in 1988, I didn't have
to ask anyone, although I had volunteered to organise the meeting when
I was in Strasbourg in 1986. But basically the reason Aston was not
Euro\TeX\ was that the President of one of the European groups decided
it must not be. If I recall correctly\footnote{On reflection, I
  suspect this suffered from retelling and translation. Surely no-one
  could be so arrogant? The general flavour should be that he was not
  in favour of such a move.}, he said he would `instruct his members
not to attend' if it were called Euro\TeX. I find it all rather sad.
Here was a great opportunity for TUG and the European groups to be
seen to be working together.  The cynical might suppose that was
exactly what he didn't want to be seen to happen.