Learning outcomes

1
Learning outcomes Understand some fundamental
database concepts Understand the basic
requirements for setting up a database and why
databases are necessary How indexer controls
retrieval results- recall, precision using
CDS/ISIS as example
Concepts for Review Indexing - pre- and
post-coordinate indexing Retrieval techniques
2
Introduction - Database Development Being looked
at primarily in the context of indexing and
retrieval in the library and Information studies
field. Definition A Database is a set of
electronic files with records that have the same
format or structure and it forms the basis for
retrieval of information in an automated
environment. As implied records are stored on
an electronic device such as the hard drive or
other direct access storage devices or magnetic
media.
3
Database Definition continued

• The records in each file are organized according
  to
• some common characteristic and are related or
• similar in some way.
• These files contain information that may be about
  
• customers/clients, equipment, suppliers, staff,
• inventory, names in a directory or in the case if
  
• libraries information bearing entities.

4
Database Definition continued

• Each record has information about one item be
  it person, book, equipment, article etc.
• Each record is made up of fields each field
  gives specific information about that item.
• Fields may be fixed length same length in all
• records or variable length different lengths
  in different records

5
Database Definition continued

• Fields may be sub-fielded. E.g. the author and
  publication
• data fields. See example of sub-fielding below.
• A Master number assigned by the software that
  manages
• the database identifies each record
• Each field has a field tag and sub-fields have
  delimiters and or
• codes.
• For example
• The DNA Story a documentary history by James
• D. Watson and John Tooze
• 245 14 a The DNA Story b a documentary
  history of
• gene cloning / c James D. Watson, John Tooze

6
A word about MARC record format and significance
(Brief explanation of what is a record format
and by extension what is the MARC
format.) Designed for communication of
bibliographic information between
computers. Structure Leader provides info such
as logical of record length, number of
positions assigned to a field and sub-field
identifiers status- new record, modified,
deleted type of record bibliographic or
authority cataloguing level in the record.
7
Significance of the MARC record format Structure
continued Directory with info about fields like
a map lists tags, length of fields, starting
position of the fields. Enables rapid computer
identification and processing of specific data
elements Variable length fields with 3 digit
tags, 2 indicator positions. MARC supports the
exchange of data among systems. A library using
this format can load records without
reformatting or extensive reprocessing.
8
SAMPLE OFCATALOGUE CARD
Brenner, Richard J., 1941-.25 GV943.25.B74
1990 Make the team. Soccer a heads up
guide to super soccer! / Richard J. Brenner . --
1st ed. -- Boston Little Brown, c1990. 127
p. ill. 19 cm. "A Sports illustrated
for kids book." Summary Instructions for
improving soccer skills. Discusses dribbling,
heading, playmaking, defense, conditioning,
mental attitude, how to handle problems with
coaches, parents, and other players, and the
history of soccer. ISBN 0316107514
12.95 1. Soccer -- Juvenile literature. 2.
Soccer. II. Title Heads up guide to super
soccer. II. Title.
9
SAMPLE OF USER VIEW IN OPAC
TITLE Make the team. Soccer a heads up guide
to super soccer! / Richard J. Brenner. ADDED
TITLE Heads up guide to super soccer AUTHOR
Brenner, Richard J., 1941- PUBLISHED 1st ed.
Boston Little, Brown, c1990.MATERIAL 127 p.
ill. 19 cm. NOTE "A Sports illustrated for
kids book. NOTE Instructions for improving
soccer skills. Discusses dribbling, heading,
playmaking, defense, conditioning, mental
attitude, how to handle problems with coaches,
parents, and other players, and the history of
soccer. SUBJECT Soccer--Juvenile
literature.Soccer. Copies Available GV943.25
.B74 1990
10
SAMPLE OF MARC RECORD FORMAT
Leader 01041cam 2200265 a 4500
Control No. 001 89048230 Control
No. ID 003 DLC DTLT 005
19911106082810.9 Fixed Data 008
891101s1990 maua j 001 0 eng LCCN
010 a 89048230 ISBN
020 a 0316107514 c 12.95 ISBN
020 a 0316107506 (pbk.) c 5.95
(6.95 Can.) Cat. Source 040 a DLC c
DLC d DLC LC Call No. 050 00 a
GV943.25 b .B74 1990 Dewey No. 082 00
a 796.334/2 2 20 MEPers Name 100 1 a
Brenner, Richard J., d 1941- Title
245 10 a Make the team. p Soccer b a
heads up guide to super soccer! / c Richard J.
Brenner.
11
MARC Record Format continued
Variant Title 246 30 a Heads up guide to
super soccer Edition 250 a
1st ed. Publication 260 a Boston
b Little, Brown, c c1990. Phys Desc
300 a 127 p. b ill. c 19 cm. Note
General 500 a "A Sports illustrated for
kids book. Note Summary 520 a
Instructions for improving soccer skills.
Discusses dribbling, heading, playmaking,
defense, conditioning, mental attitude, how to
handle problems with coaches, parents, and other
players, and the history of soccer. Subj
Topical 650 0 a Soccer v Juvenile
literature. Subj Topical 650 1 a
Soccer.
12
Indexing Records must be indexed for
retrieval The records in the databases are
indexed and the indexing system varies from one
system to another. Usually only specific fields
in a bibliographic database are indexed. The
whole field may be indexed or certain sub-fields
or certain words or phrases as in CDS/ISIS.
13
Indexing continued The Inverted File

• The indexed words are stored in a file sometimes
  referred to as
• the inverted file or the fast access file. It
  may be more than one
• file.
• The inverted file is an important file as this is
  the file which
• the software managing the database is directed
  to go to when
• a search of the database is initiated.
• The inverted file stores the indexed fields,
  words, or phrases
• along with the record numbers in which these
  occur. When a
• search is initiated, the system does not
  necessarily have to
• search each record sequentially to find
  information for a search
• request. The system saves time by first going to
  the IF for the
• Information.

14
INVERTED FILE
15
Explanation of headings on previous slide
Word The term that is indexed Postings
Number of occurrences or postings of the term
Document The document in which the word in
found Location The location on the storage
medium where each term within the document is
found.
QUESTIONS

• If above were the complete inverted file, then if
  you searched
• the word communication, how many postings would
• the system give in response to your query?.
• 2. If searched for the word camera, the system
  how many
• postings would the system give in response to
  your query?.

16
TEXT RETRIEVAL/SEARCHING
Note that generally, the software for managing
the database, allows for text searching as well
for retrieval. A database is the underlying
structure for retrieval tools. This however
takes much longer than searching indexed fields
or words or phrases and its necessity or
efficacy depends on the type of database. In a
patent database, it may be necessary to do text
searches In bibliographic databases, though
some allow for text searching, the practice is to
index the contents of only specific fields as it
is expected that users will search for
information that is found in only certain fields.
17
System retrieval of indexed records from the
Inverted File
Each record is structured so that information
about the fields and sub-fields and the record
length is stored in a type of directory called
the header or leader. This is necessary for
retrieval purposes. E.g. in the MARC record
format When a search is done on a database, and
the IF directs the system to a given record, the
software checks the directory for direction to
the part of the record where the information is
stored.
18
Different types of databases Reference are those
which point the user to another source such as a
book, a person or an organization. A referral
database points the user to persons or
organizations to refer to for the given
information. A bibliographic database points the
user to the item which has the information. It
may not give the location of the source item.
Surrogate record A catalogue database points the
user to a items in a specific library or
libraries which have the information
19
Types of Databases continued A source database
leads the user to the actual source of the
information itself. In other words, they s lead
you to the electronic document itself. Full-text
databases such as those for newspapers,
journals, dictionaries, directories,
statistics Multimedia databases with video,
pictures, sound, animation and text.
20
Bibliographic databases These grew out of the
printed versions and in fact computers were
first used in their production to make it easier
and quicker to produce the printed output and
the printed product still exists in many
cases. One advantage of having a database is that
many different products can be produced from the
database. OCLC WorldCat is a database that
supports online searching, shared cataloguing,
catalog card production, and other services.
21
Bibliographic databases continued Databases can
also be linked so that an output can be had from
more than one database. The user can tie data
together from different databases together by
retrieving data from multiple databases in one
query. For example, each borrower in the patrons
database can have books in the catalogue
22

• Bibliographic databases advantages
• A bibliographic database enhances productivity in
  libraries.
• An electronic database can be accessed from
  many locations
• no need to get up and walk to the catalogue.
• Creating reports, lists or other outputs can be
  easily done and
• maintained. Problems of misfiling eliminated.
  Updating easily done
• Many access points for searching possible
• End-users have opportunity to browse records
  electronically at
• own convenience
• A single electronic record can serve many
  purposes.
• One original no copies

23
To develop a database, what is required? 1.
Identify fields Name fields according to
convention of the software Identify sub-fields
needed What is the length What is the field
type Which are date fields which are text,
numeric For any field data integrity controls
needed e.g. messages, constant data, drop down
lists
24
Requirements for developing a database
continued 2. Determine indexing requirements.
Consider users In CDS/ISIS the indexing
convention is Decide which fields are to be
indexed The main indexing techniques are 0
by line 1 index each sub-field separately 2
index words or phrases entered if between
angled brackets 3 index words entered
between slashes // 4 index each word in the
field (Create Stop-word List of words which are
not to be indexed)
25
Requirements for developing a database continued
Determine indexing requirements A note on
Database output measurements NOTE INPUT
determines OUTPUT Method of indexing determines
to a large extent level of recall, precision
general measures of the effectiveness of
your information system. The effectiveness
depends on the input - Indexing
levels Exhaustively i.e. indexed every facet
or aspect possible Specificity used specific
terms in indexing rather than broad terms If use
approach of summarization, not specific If use
in-depth indexing more specific
26
Requirements for developing a database
continued Database output measurements
continued Notice however that exhaustivity may
increase recall at the expense of relevance
whereas specificity may increase relevance at
the expanse of recall. Recall ability to
retrieve the maximum number of relevant
documents Precision the ability to screen out
irrelevant documents Full text offers high recall
but low precision
27

• Requirements for developing a database continued
• 3. Determine input screens and outputs needed
  whether
• screens for display, print or electronic files.
• What fields needed for data entry by the
  various data entry
• personnel in different sections
• How much and what type of information user
  needs on screen?
• What information needed for administrative
  purposes such as
• evaluation of services?
• Considerations
• Always bear user needs in mind when developing
  a database.
• Plan for future growth.
• No database will be perfect work with what you
  have in the
• best way possible.

28

• SUMMARY
• Database - electronic records of same structure
  and format
• which are related in some way and are organized
  according
• to some common characteristic. Necessary for
  retrieval of
• information.
• Bibliographic databases in libraries usually in
  MARC format
• DB has indexing features which allow for
  efficient and fast
• retrieval
• Different types of databases
• Requirements for developing a database consider
  users
• Identify fields
• Determine indexing features
• Determine out and input screens and other
  reports
• Plan for future growth