\centerline{\bf Book Reviews}
\smallskip
\noindent
Every so often something good plops through the letter
box. Both the books which I received recently fall into
the `good' category. The first of these was a
particularly delectable item: not only is it by someone I
know, but it's a book about
\sgml\ that is readable. Eric van \hbox{Herwijnen's
}`Practical \sgml' (to be published by Kluwer Academic
Publishers Group) is a very thoughtful and balanced
approach to his subject. Briefly, the background is this:
Eric works at CERN in Switzerland. One of CERN's problems
(besides physics), is in exchanging documents between an
array of different machines both at the Swiss site and at
other similar locations world wide. The approach has been
to encourage the use of \sgml. The book is therefore an
outgrowth of that approach, and therefore covers not only
the {\sl meta\/}-language, but also the software which
may be employed to write with \sgml, and the problems
which were encountered. The  {\sl Practical} of the title
is well-earned.

He takes other systems seriously -- giving fair and
reasonable attention to \TeX\ (so he should, I taught
him!), although his example of using \TeX\ to construct a
memo I found a little idiosyncratic. In a similar way, he
examines how the same formula for the decay of a
particular particle could be expressed in \TeX, |eqn| and
two flavours of \sgml\ DTD (AAP Maths and  ISO). Since
one of the criticism's Eric makes of the \sgml\
applications is that `compound symbols such as $(m_{\eta
c})$ and $(Q_c)$\dots have a meaning to physicists' I was
surprised that he had not redefined them to have a name
which reflected that meaning. Similarly, if you have to
distinguish between `a right arrow and a decay reaction',
why not give them different names? The symbols can remain
the same, but the coding could be different. This is the
advantage of markup, that you can make the intention clear.
But the criticism that Eric is making here is really of the
DTDs, which are currently inadequate.

One of the appendices discusses the problems which were
encountered in writing a book on \sgml\ using \sgml. This
is a particularly illuminating section and does much to
redress the excesses of other promoters of \sgml. 

This is a
valuable and welcome addition to the literature on \sgml.
\begintt
@book{EvH,
title={Practical {SGML}},
year={1990},
author={Eric van Herwijnen},
publisher={Kluwer},
isbn={0 7923 0635 X},
price={\pound24.90},
pages={307}}
\endtt
\smallskip
\subsection{Introduction to \TeX}
The other book was Norbert Schwarz' `Introduction to
\TeX'. I bought the German version last year (along with
two other German \TeX\slash\LaTeX\ books). This is a
translation prepared by Jost Krieger, a Dutchman. Now
there's something interesting. The first non-Knuth \TeX\
book in English is a translation from the German by a
Dutchman. And the second may be Bechtolsheim's (another
non-native English speaker). Hint. It does read a bit like
a translation too. I don't think that does any harm. Once
or twice you have to reread a sentence or paragraph to
understand just what is being meant, but on the whole Krieger
has done a fair job. It is very much a book on \TeX.
There is very little about documents. It carefully, and
most often clearly, goes through the bulk of the plain
\TeX\ commands (one or two of which sent me scurrying
back to my {\sl\TeX book} just to check that they really
did exist). Very thorough. But they do tend to present
things without giving a clue of the generality or
usefulness. I did like the fact that the name Baskerville
appears by page 33. 

Many of the examples are extremely practical and useful.
Inevitably there is a tendency to introduce some things
in a brief way in order to develop something more
interesting. Although macros don't appear until Chapter
7 (starting on page 95), they have in fact cropped up
several times in the preceding chapters, with only
rudimentary explanation. The first reference to a macro I
found on page 15, and the first definition is on page 21.
I suspect that developing \TeX\ in a `linear' manner is
near impossible (after all, Knuth had to adopt a sort of
recursive approach), but sometimes I felt the reader had
to take quite a lot on trust. The alternative would be to
illustrate points with trivial examples, and I would have
criticised that too. Perhaps the single strongest feature
is the list of commands, together with an explanation
which forms the `end' of the book. It really is useful to
be able to flick through that, like an glossary of
commands, and then refer either to \TB, or to this book
for more details. A little hypertext version of the
glossary for my Mac would be very useful. 

The authors note the use of |\tracingstats|.
Unfortunately, as Knuth points out (page 303), and  they
fail to, not all production versions of \TeX\ heed the
value of the tracing commands. This could conceivably
confuse the inexperienced. In view of Chris Rowley's
comments elsewhere in \TeXline, it was interesting to
read their comment (page 96) `A superfluous pair of
braces rarely has an adverse effect..' (sic). Although the
authors claim that the book should be `a bit more
internationally oriented than a book of English (sic) or
American origin', they don't suggest A4 alternatives to the
built-in |\hsize| and |\vsize|.


If I have a major criticism, it is one which is really a
criticism of  the publisher (the inevitable
Addison-Wesley). I can see why the problem arises, but I
don't think it an adequate excuse. The book was
inadequately proof read. I suspect the authors themselves
typeset and proof read the book. This is not what we
employ publishers for! There are too many places where the
punctuation should have been tightened up (too numerous to
mention, but note the comma which begins a line on page
37 and the opening quotes round the wrong way on page 150), or
where the english should have been corrected (page 184,
`dipthong' for `diphthong'; page 84, `incontinuity' for
`discontinuity'). There are several cases where * is used
instead of $\times$. Surely the asterisk has not yet displaced
the $\times$ for a multiplication sign? What makes this more
culpable is the inconsistency of using both operators. The
table on page 12 seems curiously incomplete, with one
conversion factor given merely as the decimal point. 
Small things, but they make a great deal of difference. 

And surely we don't have to print from laser printed output?
Bromides don't really cost that much more -- if Addison-Wesley
is having difficulty finding a printer, I'd be happy to take
care of the printing from \dvi\ for them. It seem such a
tragedy that the amount of effort which Knuth went to in
providing us with \TeX, the effort that Norbert and
Jost went to in writing this particular book, should be let
down by the publisher who inadequately controls that part
of the process they traditionally handled -- copy-editing
and printing. There might be some criticism too of the
driver used, since it shows up those irritating {\it
dis\/}continuities in the horizontal bar of the square
root sign and the over- and under-braces which
appear to bedevil so many drivers.

This is a prime example of the orthogonality of form and
content. The content is excellent, but the form could do
with some attention. To dedicated \TeX ies, who know
what's going on, the form will not detract, because the
content is so strong. For the uninitiated, for whom this
book is really intended, the capabilities of \TeX\ may be
obscured.

Leaving the relatively minor criticisms of content aside,
this is a fine book. Their discussion on tables (chapter
6), the list of `common' error messages, and the glossary
make the book value for money. How much? A mere
\quid19.95 in paperback, for a bountiful 278 action-packed
pages.
\begintt
@book{NS,
title={Introduction to {\TeX}},
year={1990},
author={Norbert Schwarz and Jost Krieger},
publisher={Addison Wesley Europe},
isbn={0 201 51141 X},
price={\pounds19.95},
pages={278},
notes={see also Schwarz'
        original in German}}
\endtt

\rightline{\sl Malcolm Clark}