846
TAMPLIN AND MINKLER
Electronic computers are not oracles, and it isonly from human brains
that the analytical approaches will come. The coder will mostlikely
become one of the experts who can best analyze and integrate the data
that have been stored for retrieval. He must be a working scientist.
Furthermore, it is important to reemphasize that information retrieval is not the prime purpose of the Information Integration Project.
Rather, the project has been developed to expedite the future retrieval
of literature read by members of the group. Thus our objective is not
to read, catalog, and create an encyclopedia of the existing literature.
Rather, the literature is to be examined only to the extent that we may
be able to adequately define the state-of-the-art knowledge concerning
the release of radionuclides to the biosphere from nuclear explosives
and to point out the crucial gaps in this knowledge that require experi-
mental efforts. Therefore our selection of the literature to be examined
will always be dictated by our analysis of the overall mission problem.
The selection of literature on collateral information will not be made
on the basis of interesting problems but only upon the requirements of
the mission problems.
Within our present concept the system will consistof four separate
files, each serving a special function:
1. Bibliographic file. This file will contain the bibliographic information for each article entered into the system, i.e., authors, title,
journal or report number, volume, pagination, and date. As an article
is entered into the system, it will be assigned an accession number.
This file will be ordered on accession number, and the accession number will be carried on entries into all other files.
2. Reference file. This file will contain the accession number and
the bibliographic citation (journal or report number, volume, pages,
and date). The journals will be assigned a numerical code; each report
will be entered by its own alphanumeric code. The file will be ordered
according to source. Before a new article or report is requested or
assigned an accession number, the file will be searched to determine
whether or not the same data are already in the system.
3. Data file. This file will contain the alphabetic and the numeric
information abstracted from the documents which is coded into a threedimensional matrix. It is expected that in the X,Y plane there will be
100 rows and 100 columns. The 100 columns are intended to represent
disciplines or problem segments, whereas the rows will be logical
breakdowns for
respective columns. Thus the system is planned to
contain 104 primary index points. We plan that the system be openended in the Z direction, but we expect that no index point will have
more than 100 subcategories. This means that we plan to have about
10° coded storage slots.
4. Table file. This file will contain the tabular data from the docu-
ments. Tables will be assigned consecutive numbers, and note of the