summaryrefslogtreecommitdiff
path: root/info/digests/texline/no8/oddtypes.tex
blob: 92ca69242d4a73f316e2ad8e46269bedf905c60e (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
\centerline{\bf Odd Types}
\medskip\noindent
The BCS Electronic Publishing Group has a well-justified
reputation for its one day meetings at the  School of
Oriental and African Studies. The  meeting on
September 29th easily maintained that reputation, and,
for perhaps the first time, exploited its venue more
fully. The morning session was devoted to mathematical
typesetting, but the afternoon was concerned with
non-Roman text, and the accompanying problems and
solutions. Hence the (witty?) title, `Odd Types'.

David Penfold of the Charlesworth Group started off the
day by laying the foundations of mathematical
typesetting, covering some aspects of the history. Even
up to the 1950's, maths was set by hand, in cold metal.
The mind reels at the thought of the skills needed to
adequately typeset a moderately complex page of maths by
hand. David also described Monotype's `four-line'
equation, which was set from a keyboard. In my naivety I
had assumed this really meant four lines of equations, but
all it really means is being able to set something like
$${g_{b^2d_4}X^{a^1c_3}}\over{b_{f^el_n}Y^{b^kd_\gamma}}$$
which {\tt eqn} and \TeX\ users rather take for granted
now. Maybe there has been progress. It is always a
fascination to see how far technology limits our forms of
expression. Filmsetting on a matrix obviously poses
problems of positioning for maths. Similarly, when
symbols are obtained from a disc (before the days of
floppies), there are limits on the number of symbols
which can be placed on the one disc. To some extent,
Monotype seem to have dominated  math typesetting, and
many of us feel at home with their Roman typeface for
maths (No.\ 327 for text, 569 for maths) where every
superior and inferior has to be separate type, with a
different design size. As an aside, \TeX's favourite
typeface, Computer Modern, is a derivative of one of
Monotype's maths typefaces. This is perhaps not 
completely aside, since one feature which became apparent
was the inherent conservatism of the customer (and
perhaps the printing industry). With technical
typesetting, innovation in typefaces is hardly likely to
enhance the result. It was with this problem of
appropriate representation that David had begun his talk,
with three (fairly well-known) representations of
Creation: Genesis, Maxwell's equations, and Haydn's score.
All quite different, but in a sense saying much the same
thing, although to rather different audiences. David also
showed some rather good bad examples of mathematical
typesetting, but unfortunately would not name the
perpetrators. His conclusion of `no general solution' was
not entirely unexpected, and perhaps a blow to some of us.
\looseness=-1

Tibor Tscheke from St\"urtz (part of Springer-Verlag's
empire) discussed some of the problems which they had
encountered in adopting an electronic approach to their
journal production. It comes as no huge surprise that
author input on electronic input does not entirely solve
all problems. Authors do not necessarily understand the
remainder of the publishing process, and their own
assumptions may make life very difficult. Coupled with
the fact that the correction process may actually be more
expensive than key-boarding we see some problems looming
for publishers who are continually under pressure from
authors to accept electronic input (and give them higher
royalties). Since St\"urtz used Unix, it was no real
surprise that their initial attempts were with {\tt
troff}. But since they were using Lasercomp typesetters,
and {\tt troff} was really intended for the obsolete CAT
typesetter, this proved difficult. So to \TeX{}. One of
the recurrent problems with \TeX\ is its insistence on
knowing a lot about the character set you are using. And
most of that information was lacking for the Monotype
fonts which were to be used. So a great deal of work had
to be done by hand to create the information for the font
metric files. At least this is a one-time job. Some of
\TeX's font-dependency also became apparent, since some
of the standard macros do not work well outside
the Computer Modern Family, and had to be rewritten.
There was also a problem to be solved in screen
representation. Since we can see \TeX\ on the screen for
some kind of crude proofing, we have to adopt some scheme
of font substitution (we don't have screen
representations of the Lasercomp's fonts). Substituting
the Computer Moderns directly gave a rather
unsatisfactory result, with overwriting on the screen and
ragged margins. This rather upset people. In the end, the
approach was to generate new versions of the Computer
Moderns, but with the width characteristics of the
corresponding Lasercomp fonts. This ensured that
everything was in the right place on the screen, even if
typographically it was not perfect (since we're dealing
with an IBM pc, with not very good graphics anyway,
typographic purity is hardly a key issue). Of course, the
reason that \TeX\ is amenable to this approach is by way
of \MF{}. Part of \MF's beauty
is its power to allow modification of existing fonts
created through \MF\ --- hence the `meta' part of the
name. St\"urtz are also moving on to other fonts, like
Fraktur, phonetic, Hebrew and so on. One of their problems
is in finding a way of giving them away --- Monotype and
St\"urtz are still exploring this.

Unfortunately Mike Glover, from Icon Technology, who was
to have talked about Formulator could not be present.
Formulator is a Macintosh tool which allows equations to
be typeset in a {\it wysiwyg} fashion. The resulting
equation may then be pasted into a MacAuthor document, or
may be represented in \TeX\ format, for inclusion into a
\TeX\ document. Contrasting reviews of Formulator have
appeared in \TeXline\ 7, the November issue of {\sl Desktop
Publishing Today} and the October issue of {\sl Apple
Business}. It is hoped that Mike will be able to present
Formulator at some future date.

The afternoon moved into even more exotic areas, starting
with Dominik Wujastyk's description of typesetting some
South Asian languages. There is perhaps one key point to
be made about the afternoon's presentations: the fonts
were really designed for Western use, rather than by the
native speakers of the languages. In a sense, the markets
are rather small, and we are moving into areas where
electronic (or desktop) publishing really can make an
impact, permitting things to be published which could not
have been done by traditional techniques. Dominik is a
Sanskritist by calling, and a curator at the Wellcome
Institute by profession. Some of his work there has been
in publishing Sanskrit editions concerned with the history
of medicine. Taking Sanskrit as an example, there are
several ways to represent this in typeset form. We can
use the Roman alphabet in a standardised transliteration.
This is obviously attainable with the present levels of
word processing. Or we can move to `real' Sanskrit fonts,
when we input with a standard keyboard (i.e.~Roman
alphabet) and somehow massage this into the Sanskrit
alphabet. Dominik showed examples of these approaches. 
One area which he also touched on was the use of
preprocessors, where, prior to the `typesetting' stage
(naturally performed by \TeX/\LaTeX), the input (here
Note Bene/Xywrite) was transformed. Again there are
several ways to do this, from the obvious --- write your
own program, to the slightly more subtle --- configure the
editor's printer driver to do the transformation for you.

Paul Thompson of SOAS gave an account of the approaches
which could be employed in setting Chinese. As we know,
Chinese (and we can include variants, from the pure
unmodified Chinese, or Hanzi, through the hybrids, the
Japanese Kanji and Kani, and the Korean Hancha and
Hangul, to the modified Vietnamese), has a rather large
character set. Just how large is unresolved. The Chinese
national dictionary will use about 60,000 different
characters, but only about a maximum of 15,000 are needed
for modern styles, and the Chinese standard character set
is a mere 7,000. But even this, in two different styles
and at 4 standard sizes, gives 56,000 characters. Clearly
there are a few storage problems for electronic
representation. Whatever problems are there are made less
straightforward by the presence of three different
standards --- Japanese, Taiwanese and Chinese.

We are probably all aware that there are certain
assumptions made in the way we handle characters
electronically; operating systems and applications tend
to depend on the assumption that character sets are not
larger than the {\sc ascii} set. One way of addressing the
problem here has been to use 2 {\sc ascii} characters to
represent chinese script. 

Again the cultural biases were apparent when Paul
described the evolution of techniques to handle text
creation (i.e.~typing). One reasonably successful
technique was based on telegraphic techniques, where a
pair of `western' characters represented a chinese one. A
large number of telegraph operators were very familiar
with this technique, and although it took some time to
learn, it was fast and of fairly low cost. Many of the
newer techniques which had been tried were either slower
or more expensive. On the other hand they tended to be
faster to learn. In general the trend seems to be towards
using a `standard' keyboard and inputing phonemes, at
either the syllabic or entire word level. With
improvements in artificial intelligence this appears to
be a viable route.

Lastly, Michael Mann described a number of case studies
drawn from various departments within SOAS. As a hardened
markup man, I was forced to admit that {\it wysiwyg} could
sometimes come up with a reasonable published product.
It was perhaps in the area of very odd scripts
that the versatility of the Mac (in the right hands)
shone through. Some of the fonts had to be designed by
hand, but effectively what they replaced was doing the
whole job by hand --- at last electronic publishing was
doing something which was otherwise impossible. It was
interesting to see that hyphenation problems were still
present in Amharic, but since no-one was very clear where
the hyphens ought to be anyway\dots\ The limitations of
technology were again apparent by the problem of having
to use monospacing to accommodate the Mac's handling of
accents. The school also uses Ready Set Go!{} in an Arabic
version --- that means the menus are in Arabic too ---
highlighting another Mac-factor. Although the Mac can be
`localised' that means it becomes completely localised.
Michael even touched on subjects like Microsoft's
{\sc rtf} (rich text format) which is usually a fairly
well-hidden secret, and the problems in finding out about
Apple's Scriptmanager. All of which added weight to his
contention that Babel still reigned.
\smallskip
\rightline{\sl Malcolm W Clark}