June 26, 1974 The success

Essays of an Information Scientist, Vol:2, p.82-85, 1974-76
Current Contents, #26, p.5-8, June 26, 1974
Number
June 26, 1974
The success of Current
tents
has stirred
up a
librarians
and
the artifkial
journals.
scientists
separation
who
Indexa
reference
true in the
(SCZ@).
In 1974 the SC1 staff
approximately
tations.
To
bemoan
of books from
This is especially
Science Citation
will process
5,000,000 reference ciprepare
the SC1, every
given in about
ticles
will
those
references
About
Book Connumber
of
be keyed.
500,000
About,
cite journal
20% cite books
ar-
difficult
quently
the
difficulty
is compounded
One may ask why all the concern.
The user of SC1 & SSCIrM does not,
presumably,
articles.
‘book’.
across
need to be told
However,
it, when
looking
up a journal
For many years we used to identify these ‘non-journal’
citations with
pointed
consistency,
usefulness.
and
its apparent
lack of
Few users have ever com-
mented upon our dropping the code.
Nevertheless,
it is now important
to
review the decision
because of the
advent of the Social Sciences Citation
Index’”. More than 5W0 of the references picked up from social science
journals cite books.
More often than you would imagine,
it is difficult
to distinguish
between
a reference that cites a book and one
that cites a journal. It is all the more
In addition,
librarians
asked
many
of SC1 to check on the
of information
some
client
about
wishes
the client
recall
citing
memory.
to
consulted
to cite. But we all
a book
The
librarian
corroborate
Unlike journal
One
be the case
presumably
he wants
a book
to cite.
why this should
the book
from
1 Thus
would like to use the Citation
wonders
can
verification.
section
accuracy
when
as we have previously
out, the SCZ has become a tool
for citation
that
is a
one may have come
article by the same author,
Index
that the
he or she is consulting
journal’ items that are not clearly identifiable as references to journals.
a special code, but we discontinued
the
coding because of its high cost, in-
at
by a lack of generally accepted citation
standards or a flouting of what standards may exist.
citation
‘non-
it has to be done
breakneck speed while keying for input to the 1S1 data bank. Not infre-
80%0 of
or other
when
26
citations,
our
or article
is often
memory.
book citations
in SC1 can often be less than adequate.
This was brought
82
home tome
when we
recently
book
reviewed
citations
counts
for highly-cited
ride. Today publishers relish the use of
journal titles that do not clearly identify
for
of citation
‘classics’z.
them as journals.
~s-
printed
If I allow myse\f the luxury of unlimited
central-processor
time (not
really possible
when handling
such
A single book or article may be cited
enormous
files) I could deal with the
problem in the following way.
crepancies
our
our conventions
in a study
between
magnetic
information
tapes
and
the
from
SCI appeared.
a dozen
or more
citing article.
times
by the same
In the case of books,
Given two entries under the same
cited author and the same cited year,
if
the author of the citing article does not
do
the
two
contain
the
same
cited
cite a page number, all of the repeated
titles,
citations
Though they may, they may still represent citations of a book or a journal.
are essentially
duplicates.
The
data-entry operator does not usually
remember this repetition as the article
is processed, and creates dozens of
as for example,
One feature
that
they
one
might
Pros tagkmcfins.
of most journal
contain
citations
a volume
is
number.
duplicate entries. During the final printout of the SCZ the computer
detects
these duplicates
and “unifies”
them
Unfortunately,
such heavily cited journals as J. Chem. Sot. London do not.
into a single line of information.
HOW
ever, the individual references remain
Thus, if we encounter
a blank in the
volume field we may reasonably take
in the data bank tapes. They may cause
the time
confusion
in any statistical
analysis
which uses the printed SC1 for veritl
that do not use volume
If the citing
author
does supply
we have another
and it must be carefully
a
problem
considered.
II
And therefore,
bnd
withotrl
algorithms,
book title
nate
superior
to our computel
consider the case when a
is identical
to
of a partiarises as to
Nature
or Freud’s
of the Chemical
Collected
works, don’t
we do the searcher a disservice by
eliminating
these clues. If we elimi-
some other clue, it is not possible to
unify what the human eye or brain
knows is a book title. If you feel
suddenly
numbers.
such citations
as citations
cular book, the question
as Pauling’s
the book title could be :
title.
a list of journals
citing articles in the Source Index. But
if it is a frequently cited book such
SC1, twelve separate lines will appeal
in the index. As far as the compute!
journal
basis.
whether we should give the cited page.
[f the book in question is cited only
a few times a year, the matter is trivial;
the reader can examine the titles of the
a dozen different
pages of the saint
book have been cited, they are re
garded as twelve separate
citations
When the information
is printed in th~
is concerned
to examine
on that
Even if we should be successful,
one way or another, in identifying
all
cation.
page number,
dktinguish
a journal
83
would
space
would
such
specific
page citations,
we
save an enormous
amount
of
in the SCZ. Each cited book
have
a single
entry.
Though
such citations
in the
now
appear
SCZ, the presence
together
of the page
number causes them to print
different,
although
contiguous
items.
out as
cited
I
abetically, know enough about workers
in his field of interest to pick them
from a straight alphabetic list?
How often
of a book
have you used the page
as a starting
point
in an
Consider pauling’s book, which is
cited hundreds of times every year. I
SCZ search? Let me point out that the
problem is much less severe when we
have always assumed
are dealing with multi-authored
useful
that is would
be
for the user of the SCl to enter
the Citation
or chapter
Index
from a specific page
of such a work.
The user’s
interest
may be quite specific. That
specific subject may be discussed in
certain chapters or on specific pages.
Such a user will not want to find all
references to the book, but only those
related to a specific chapter or page.
with chapters
authors.
or sections
Most people
books,
by different
cite the chapter
author first, then the book title, pagination, and then (perhaps) the editor.
We assume that searchers will begin
with the name of the chapter author.
1ss any case there usually is no volume
number.
The page number is usually
known
by anyone
who is looking
for
Some members of our Editorial
Advisory Boards think it is the rare user
papers that cite that particular chapter.
lf we dropped the page number, librar-
who will start a search with a specific
page number,
lf they are right, our
ians and other reference workers would
lose the information
which they frequently
find useful in verifying reference
sources
for customers.
Un-
present process of including cited page
numbers is wasteful. It uses extra space,
adds to printing and other costs, and
separates- entrie~ that otherwise would
be arranged
citing
in alphabetical
author--with
order
all duplicates
by
elim-
inated.
Obviously
a book that is cited only
a few times a year presents no such
problem, even in a five-year cumulation.
But a heavily-cited book, and there are
doubtedly,
on occasion, the omitted
editor’s
name is a slight loss. For
example,
I have always assumed that
referring
to a book
turn
up all citations.
the
chapter
by Merton
It turns
authors
have
would
out that
also
been
cited. Their names must also be added
to my “Merton” ASCA@ profile
Let me point
out that the omission
many of them, if cited 20 times a year,
will appear 100 times in a five-y eat
of page numbers from most references
to books is so flagrant that one wonders
cumulation.
Will the user be able to
distinguish the subject interest of these
why we should have attempted
citing
will now understand
works
without
some indication
with
the
feel like
of page or chapter? wdl the title of the
citing article or the name of the citing
journal tell him what he really want!
out
to know? Would he, if we igrrorec
cited pages and listed authom- alph
to drop
matter
frying
I suspect
84
page
you
why I sometimes
King Solomon.
of the
to deal
at all. Perhaps
pan
As we say,
into
the compromise
numbers
the
fire.
will be
for single au-
thored
works
bers in
volumes.
and retain
page num-
considerable
we have done
SCI. But the continued
of the scientific
literature,
size and scope of interests,
tify them in the first place. If we stick
with our presqnt policy it will not be
the first time
of the existing
growth
the case of multi-authored
The problem is how to iden-
so after
soul searching.
economically
difficult.
It is a major
problem to keep ahead of inflation
labor and paper costs.
a separate SCZ Book Supplement
done
would handle all sorts of non-journals-
literature
to
integrate
with the journal
the
book
literature
it
has been an indirect assault on a major
problem in bibliographic
librarian
and many
that books,
especially
control.
scientists
Every
realize
multi-authored
works of all kinds, are not adequately
indexed. What needs to be done now
is to take a big leap forward
monographs,
conference
which
proceedings,
etc. Librarians will not find this separation difficult
to accept since it is
traditional in the library world. This
separation would of course be necessary
for printed indexes. But is should be
remembered
that the growth of our
on-line
by pro
in
Consequently
I can foresee in the
near future that 1S1 may embark upon
All of the above leaves open a very
important question. While the SCI has
much
in
makes this
services
means
that the book
cessing such works as source materialsin the same way we handle journal
and journal material will be treated
as one. The same could eventually be
articles. Until quite recently I was hope-
true
ful this could
for our tape
and ASCA
users.
be done in the context
veriGarfield, E. Precise bibliographical
1 fication with the Science Citation Index.
Current Contents @ No. 35, 2 Septe.nber
1970, p. 4-5.
85
.................. A core research library for
z developing
graduate
schools--the
100
books most-cited by researchers. CCO No. 1,
2 January 1974, p. 5-9.