Optical scanning for retrospective conversion of information
| Date | 01 June 1986 |
| Pages | 328-331 |
| DOI | https://doi.org/10.1108/eb044721 |
| Published date | 01 June 1986 |
| Author | Morten Hein |
| Subject Matter | Information & knowledge management,Library & information science |
ARTICLES
Optical
scanning
for
retrospective
conversion
of
information
MORTEN
HEIN
Library Inspector
State
Inspection
of
Public
Libraries
Copenhagen
Denmark
Abstract: The
method
of
doing retrospective conversion
by
means
of optical
scanning and computer
formatting is
discussed.
An historical
survey
of research
in
such applications
is
given
as
background.
Applications,
named OSCID,
are
considered
as
important
for making low
priced machine-readable
data
representing older
material.
OSCID
applications could be
of
special interest
for
information systems within
humanities.
Introduction
O
SCID is an acronym used to
cover a series of applications
which will be discussed in this
article. OSCID stands for Opti-
cal Scanning for Creation of
Information Databases.
An overview of the electronic information
scene in so far as library collections are
concerned reveals that electronic information
means NEW information, with the coverage
of in-house databases normally going from
the point at which each individual library
and/or information service installed an auto-
mated system.
With the growth rate of information
services this means that information available
on such databases will normally be a dozen
years old or so. The only services that differ
from this are the bibliographic utilities like
OCLC, Utlas and RLIN.
Considering the profile of subject coverage
by electronic information services, it is a
known fact that databases normally contain
information on high-brow and high-tech
aspects in science, applied science, business
and economics. These are the areas where
one can expect a high information turnover
— and very little interest in 'older' (i.e.
pre-1970) information.
The motivation for efforts in making this
older information available is lacking, and
thus too little attention has been paid to
developing low cost applications for retro-
spective conversions of already existing
material. This creates a danger for such older
information which is very often still worth
while using. This applies especially to
subjects in humanities for two reasons:
— humanities rely more intensively on older
information, because there is a lower ratio
of obsolescence.
— humanities are, economically speaking, a
low budget area which can hardly afford
the costs of present day techniques for the
creation of machine-readable infor-
mation, let alone the costs of retrospective
data conversion.
A prolonged version of today's trend will
give humanities a lower priority on the infor-
mation market, since the general information
development and interchange in humanities
will not keep
pace.
Humanities will then have
to rely on spin-off information created by
their rich cousins and thereby be forced to
use perhaps mis-shapen electronic infor-
mation in a strange mix with old-fashioned,
second-class services.
Obtaining machine-readable
information
There have been several investigations into
creating cheap machine-readable infor-
mation. The best solutions are known from
strictly bibliographic systems. The appli-
cations can consist of resource sharing e.g.
shared cataloguing, so that participating
institutions bring correspondence between
their own traditional information and the
machine-readable version established and
paid for by other institutions. This saves
original cataloguing and keyboard work and
is the virtue of all multi-user/multi- institu-
tion systems.
The conversion can be a matching process
in a dialogue form or
as
a batch
process.
After
this match the machine-readable 'foreign'
version is usually accepted instead of one's
own version, typically printed or typed. Of
course local information can be added, but
that is a costly affair. In theory downloaded
or transferred information should be as good
as the local 'original',
as
information could be
supposed to be objective — but it is not.
There are different ways of collecting infor-
mation and of interpretation and indexing —
take simple problems such as forms of names
for instance. Detailed presentations of infor-
mation also tend to be language dependent,
thus problems are caused by trans-language
resource sharing.
The ideal situation would be to possess
applications that could make the local infor-
mation machine-readable at a low cost. And
an even better version would be subsequent
computer matches between this low-cost
328 The Electronic Library, December 1986. Vol. 4, No. 6.
To continue reading
Request your trial