Biology Reference
In-Depth Information
In 1970, Edgar F. Codd, working for IBM in San Jose, California,
wrote a paper describing a new system for organizing data records.
“A Relational Model of Data for Large Shared Data Banks” set out
a scheme by which the representation of data could be completely di-
vorced from their physical organization on a disk or tape: “Activities
of users at terminals and most applications programs should remain
unaffected when the internal representation of the data is changed. . . .
Changes in data representation will often be needed as a result of
changes in query, update, and report traffi c and natural growth in the
types of stored information.”
7
Codd's idea was to organize the data into
a set of tables that were related to one another by “keys” that linked
data across tables.
8
For instance, a library database might contain in-
formation about all the topics in its collections. Such information could
be spread over multiple tables, as in this example (a library system with
just four topics):
BOOK_ID
AUTHOR_ID
BOOKNAME
CALL_NO
LIBRARY_NO
1
1
Making Sense of Life
XYZ
1
2
2
Simians, Cyborgs, Women
ZYX
2
3
2
Primate Visions
YXY
1
4
3
Nature and Empire
YYZ
2
AUTHOR_ID
FIRST_NAME
LAST_NAME
1
Evelyn
Keller
2
Donna
Haraway
3
Londa
Schiebinger
LIBRARY_NO
LIBRARY_NAME
LIBRARY_ADDRESS
1
Library of Congress
Washington
2
New York Public Library
New York
Columns with identical names in different tables (“AUTHOR_ID” and
“LIBRARY_NO”) are linked together in the database. To fi nd the au-
thor of
Primate Visions
, for example, the database must fi rst look up
that title in the fi rst table, retrieve the AUTHOR_ID (the “key” for the
table of authors), and then look up the corresponding author in the
second table. The query itself creates a link or “join” between two ta-
bles in the database. Codd's paper also suggested a “universal data sub-
language” that could be used to query and update the database. Such
a language would refer only to the names of the tables and the names
