UNIT 11 INDEXING TECHNIQUES

Structure
11.0 Objectives
11.1 Introduction
11.2 Derivative Indexing and Assignment Indexing
11.3 Pre-Coordinate Indexing System
11.3.1 Cutter’s Contribution
11.3.2 Kaiser’s Contribution
11.3.3 Chain Indexing
11.3.4 PRECIS (Preserved Context Index System)
11.3.5 POPSI (Postulate Based Permuted Subject Indexing)
11.4 Post-Coordinate Indexing
11.4.1 Pre-Coordinate Indexing versus Post-Coordinate Indexing
11.4.2 Term Entry System and Item Entry System
11.4.3 Uniterm Indexing
11.5 Keyword Indexing
11.5.1 Key Word in Context Indexing (KWIC)
11.5.2 Variations of Keyword Indexing
11.5.3 Double KWIC
11.5.4 Other Versions
11.6 Computerised Indexing
11.6.1 Meaning and Features
11.6.2 Manual Indexing versus Computerised Indexing
11.6.3 Advantages and Disadvantages of Computerised Indexing
11.6.4 Components of Computerised Indexing System
11.6.5 Categories of Computerised Indexing Systems
11.6.6 Comparison of Computerised Indexing Systems
11.6.7 Index File Organisation
11.6.8 Methods of Computerised Indexing
11.7 Indexing Internet Resources
11.7.1 Search Engine Indexing
11.7.2 Subject / Information Gateway
11.7.3 Semantic Web
11.7.4 Taxonomies
11.8 Summary
11.9 Answers to Self Check Exercises
11.10 Keywords
11.11 References and Further Reading

11.0 OBJECTIVES
After reading this Unit, you will be able to:
 understand the basic principles of subject indexing techniques;
 appreciate the differences between (a) derived and assigned indexing, (b) pre-
78 and post-coordinate indexing systems;
 trace the major contributions in pre-coordinate indexing systems; Indexing Techniques

 learn the different stages of intellectual operations and working of some important
pre-coordinate indexing systems such as, Chain Indexing, PRECIS and POPSI;
 understand the objective conditions that led to the development of post-coordinate
indexing;
 learn the different entry structure such as item entry as well as term entry systems
and the methods of post-coordinate indexing with particular reference to Uniterm
indexing system and post-coordinate searching devices;
 explain and apply the different varieties of keyword indexing;
 appreciate the differences between the manual indexing and computerised indexing;
 understand the concept of computerised indexing in terms of its features,
components, categories, and index file organisation;
 learn different methods associated with the generation of index entries with the aid
of computer;
 get acquainted with the indexing internet resources with particular reference to
search engine indexing and other associated concepts; and
 develop skills of using different indexing techniques for formulating different types
of subject headings.

11.1 INTRODUCTION
Subject approach to information has been an area of intense study and research in the
area of organisation of information, resulting in the generation of new theories and the
design of the corresponding new indexing techniques based on these theories. Indexing
technique actually originated from what is known as the ‘back-of-the-book index. Its
objective is to show where exactly in the text of a document a particular concept (denoted
by a term) is mentioned, referred to, defined or discussed. The ‘back-of-the-book
index’ may be either in the form of specific index or relative index. Specific index presents
the broad topics in the form of one-to-one-entry whereas the relative index is one
which displays each concept in different context. The best example of such an index is
the relative index of Dewey Decimal Classification. But the relative index is usually
unique to the text to which it points to and is quite difficult to maintain on a large scale.
Subsequently we have seen the development of pre-coordinate indexing model.
Till about the early fifties of the last century, the pre-coordinate indexing models were
the only ones that had been developed. In the subsequent decades, the post-coordinate
indexing models were designed and developed, the physical apparatus for these index
files also changed from the conventional index cards to different formats of post-
coordinate indexing models. With the advent of the computers, the keyword index
models like KWIC, KWOC, and KWAC were introduced, post-coordinate indexing
also became more amenable for computer manipulation. Most of the bibliographic
databases today have indexes based on post-coordinate indexing principles. The
following sections of this Unit present the major developments in subject indexing
techniques for organising the index file.

11.2 DERIVATIVE INDEXING AND ASSIGNMENT

INDEXING
Indexing can be either ‘derived indexing’/’derivative indexing’ or ‘assigned indexing’/
‘assignment indexing’. Derived indexing is a method of indexing in which a human indexer 79
Indexing or computer extracts from the title and/or text of a document one or more words or
phrases to represent the subject(s) of the work, for use as headings under which entries
are made. It is also known as extractive indexing.
In derivative indexing, terms to be used to represent the content of the document are
derived directly from the document itself. Here no attempt is made to use an indexer’s
own knowledge of the subject or other guides, but use only the information which is
manifest in the document. Index terms are derived directly from the title or text of the
document. It requires least intellectual effort on the part of the indexer. Mechanical
devices and computers are used in abundance to carry out the burden of index
preparation as well as the tasks of matching these with the questions of users put to the
system. Examples of derivative indexing are keyword indexing, citation indexing,
automatic indexing, etc.
Assigned indexing is also known as ‘concept indexing’, because it involves identifying
concept(s) associated with the content of each document. It is a method of indexing in
which a human indexer selects one or more subject headings or descriptors from a list
of controlled vocabulary to represent the subject(s) of a work. The indexing terms
selected to represent the content need not appear in the title or text of the document
indexed. Here, an indexing language is designed and it is used for both indexing and
searching. Some notable examples of assignment indexing are chain indexing, PRECIS,
POPSI, classification schemes, etc.

11.3 PRE-COORDINATE INDEXING SYSTEMS

Basically all indexing systems are, by nature, coordinate indexing. The purpose of using
a combination or coordination of component terms is to describe the contents of the
documents more precisely. Many subjects can be expressed in a single term, e.g.
Libraries, Cataloguing, Management, etc. Others are expressed as a combination of
these, e.g. Cataloguing in libraries, Management of libraries, etc. When the indexer
assigns subject headings representing such compounds and arranges entries in a series
of classes according to the subject content of the document, the resulting system is
referred to as pre-coordinate indexing system. Here, the indexer coordinates the
component terms representing compounds at the input stage, i.e. at the time of indexing
in anticipation of users’ approach. Most of the classification schemes allow a measure
of coordination, either by including compound subjects or by providing facilities for
creating them out of simple elements. The same is true for readymade lists of subject
headings like Sears List of Subject Headings and Library of Congress Subject Headings.
Such classification schemes and lists of subject headings can therefore be regarded as
pre-coordinate indexing languages. Almost all pre-coordinate indexing models have an
a priori approach in choosing their semantic paradigms such as categorisation of
concepts, role indicators, relational operators, etc. Pre-coordinate indexing involves
coordinating (combining, pulling together concepts) followed by engaging in an act of
synthesis to build the index entries. In such a system, the most important aspect is to
determine the order of significance by following the syntactical rules of the given indexing
language.

11.3.1 Cutter’s Contribution

Charles Ammi Cutter first set forth the rules for alphabetical subject headings in a
systematic way in his Rules for a Printed Dictionary Catalog in 1876. He used the
term ‘subject cataloguing’ instead of ‘subject indexing’. It was C. A. Cutter who gave
the idea of specific subject entry. He also tried to systematise the rules for compound
80 subject headings which consisted of more than one term as a phrase and where the
subject heading was composed of a name of a subject and the name of a locality and so Indexing Techniques
on. Cutter’s concept of specific subject heading was quite different from what we mean
today. He had in mind a set of stock subject names and every book had to be
accommodated under the most restricted subject which contained the subject of a
book. For compound subject heading, Cutter laid down the rule that the order of the
component terms in compound subject heading should be the one that is decidedly
more significant. But Cutter could not prescribe how one will come forward to decide
which one is more significant. The question of significance varies from user to user. The
decision in respect of ‘significance’ was left to the judgment of individual indexer, which
was subjective. Some specific rules/guidelines as furnished by Cutter in his Rules for a
Printed Dictionary Catalog are mentioned below:
1) Person versus Country
Entry will be under person in case of single / personal biography. Entry will be under
country in case of history, event, etc. For example the biography of Sachin Tendulkar
would entered under his name whereas the biography of Indira Gandhi would be entered
her name and India- history also.
2) Country versus Event
Entry of an event would be under event if it is proper noun, e.g. Jalian Wala Bagh. Entry
would be under country if it is common noun. e.g. Freedom fighters in India, entry
would be under India.
3) Subject versus Country
In scientific subjects, entry will be under subject qualified by place. e.g. Oceanography,
India.
It has been stipulated to prepare entry under the name of place qualified by the subject
in subject areas like History, Government and Commerce, e.g. India—Moghul Period.
For Humanities, Literature, Art, etc., adjectival form of subject headings were suggested,
e.g. Indian Painting, Moghul Architecture, etc.
4) Between overlapping subjects entry will be according to the importance of the
subjects. It is to be pointed out here that the decision regarding ‘importance’ was
left to the judgment of the indexer.
5) When there is a choice between different names, Cutter prescribed the followings:
a) Language—If there are two languages out of which one is English, entry will
be under English. Here the rules appear to be biased towards English
language.
b) Synonyms—Entry will be under one word with reference to others.
c) Antonyms—Entry will be under one word with reference to others.
6) In case of compound subject headings, Cutter prescribed the following rules:
a) A noun preceded by an adjective, e.g. Organic Chemistry, Ancient History,
etc.
b) A noun preceded by another noun used as an adjective, e.g. War Prisoners,
Flower Fertilisation, Death Penalty, etc.

81
Indexing c) Use direct form in case of a noun connected to another noun with a preposition,
e.g. Patient with heart disease, Fertilisation of flowers, death penalty, etc.
d) Use direct form in case of phrase or sentence used as the name of a subject,
e.g. Medicine as profession.

11.3.2 Kaiser’s Contribution

Kaiser started from the point where Cutter left. In 1911, Julius Otto Kaiser (1868–
1927), a special librarian and indexer of technical literature, developed a method of
subject indexing system known as Systematic Indexing. Kaiser defined indexing as
“the process by which our information is collected and made accessible” and claimed
that it constitutes “the main work of organising information”. The definition highlights a
cardinal tenet of his theory of systematic indexing —that, within the framework of a
business library with which he was attached, the primary aim should be not to index
‘documents’ but ‘information’, which Kaiser took to be the various ‘facts and opinions’
(i.e., informational units) encoded in the texts of documents. He viewed systematic
indexing as a two-step procedure: The first step was to analyse a subject so as to
distinguish constituent concepts associated with the content of the given document into
two fundamental categories of facets: “Concretes”, and “Processes”. ‘Concrete’ refers
to things, places and abstract terms, not signifying any action or process; e.g. Gold,
India, Physics, etc. “Process” refers to mode of treatment of the subject by the author
(e.g. Evaluation of IR system, Critical analysis of a drama), an action or process
described in the document (e.g. Indexing of web documents), and an adjective related
to the concrete as component of the subject (e.g. Strength of a metal).
Kaiser also laid a rule that if a subject dealing with place, double entry (‘Concrete—
Place—Process’ and ‘Place—Concrete—Process’) is to be made. For example, index
entries for ‘Manufacturing of petrochemicals in West Bengal’ would be:
PETROCHEMICALS—West Bengal—Manufacturing; WEST BENGAL—
Petrochemicals— Manufacturing. The second step was to synthesize constituent
concepts into indexing statements formulated according to the strict rules of citation
order—Concrete is to be followed by Process.
Thus, according to Kaiser’s systemic indexing, all indexing terms should be divided into
fundamental categories “concretes”, “countries”, and “processes”, which are then to
be synthesized into indexing “statements” formulated according to strict rules of citation
order. Some examples of subject headings according to Kaiser’s systematic indexing
are furnished below:

Documents Categories Subject Headings

Indexing of films Concrete—Process Films—Indexing
Strike in India Country—Process India—Strike
Libraries in Nepal Concrete—Country Libraries—Nepal
Manufacturing of Concrete—Country— Copper—
copper in Assam Process Manufacturing—Assam

Kaiser’s systematic indexing did not make any provision for entry under the ‘process’
term and as a result it failed to satisfy the users’ approach by the ‘process’ term. The
concept of ‘time’ was also left by Kaiser. It is to be pointed out here that Kaiser was
perhaps the first person who gave the idea of categorisation in subject indexing.
82
11.3.3 Chain Indexing Indexing Techniques

Ranganathan’s facet analysis of subject provides a kind of representation of subjects

by transforming multidimensional relations of subject into a modulated layer of linear
representation. Ranganathan is credited with the invention of chain indexing, an
economical system of providing access to the terms in classification schedules without
replicating the hierarchical structure of the classification in the alphabetical index.
Ranganathan’s Chain indexing technique was devised as a complementary and
supplementary tool to classification schemes. However, due to the efficiency and
economy, this technique can effectively be made use of in deriving alphabetical subject
indexes for any indexing/abstracting services.
Definition and Use
A chain is a string of terms organised in a particular sequence based on the classification
scheme that the chain adopts. The sequence of terms is pre-coordinated according to
the syntactical rules of the given classification scheme. Chain indexing is a method of
deriving subject headings from the chain of successive subdivisions leading from general
to specific level needed to be indexed.
According to Ranganathan, chain indexing is a
“procedure for deriving class index entry (i.e. subject index entry) which refers
from a class to its class number in a more or less mechanical way.”
A note is also given with the above definition as
“Chain procedure is used to derive class index entries in a Classified Catalogue,
and specific subject entries, subject analytical, and ‘see also’ subject entries in a
Dictionary Catalogue.”
Chain indexing was used in the British National Bibliography (BNB) in the 1950s
and 1960s until it was replaced by PRECIS-indexing.
Chains and Links
The concept of ‘chain’ is the foundation of chain indexing. A chain is deemed to be a
structural manifestation of a subject. The term ‘structure’ in this context refers to the
parts constituting a subject and their mutual interrelationship. It is a modulated sequence
of sub-classes or isolates ideas.
Since the chain expressed the modulated sequence of sub-classes more effectively in a
notational scheme of classification of subjects, this method takes the class number of
the document concerned as the base for deriving subject headings not only for specific
subject entry but also for subject reference entries. The nature and structure of the
classification scheme used to classify the subject of the document controls the structure
of the subject headings drawn according to the chain procedure. The concept of ‘chain’
becomes operative only after the concept of a set ‘links’ about the structure of the
subject is conceded. A chain should comprise a link of every order that lies between the
first link and last link of the chain. The different types of links in chain indexing system
are discussed below:
 Sought Links (SL): Sought links denote the concepts (at any given stage of the
chain) that the user is likely to use as access points.
 Unsought Links (USL): Unsought links denote those concepts that are not likely
to be used as access points by the user.
83
Indexing  False Links (FL): False links are those that really do not represent any valid
concept, mostly these are connecting symbols or indicator digits.
 Missing Links (ML): Missing links represent those concepts that are not available
in the preferred classification scheme, these are inserted by the indexer by means
of verbal extension at the chain-with-gap corresponding to the missing isolate in
the chain whenever there is such a need.
Steps in Chain Indexing
The following steps are to be followed in chain indexing for deriving different types of
subject headings:
1) Construction of the Class Number of the Subject of the Document
Classify the subject of the document by following a preferred classification scheme. A
class number constructed according to a scheme for notational classification will form
the basis for applying the rules for chain procedure for deriving subject headings.
2) Representation of the Class Number in the form of a Chain
Represent the class number in the form of a chain in which each link consists of two
parts: class number and its verbal translation in standard term or phrase used in the
preferred classification scheme.
3) Determination of Links
Determine different kinds of links: Sought Links (SL), Unsought Links (USL), False
Links (FL), and Missing Links (ML).
4) Preparation of Specific Subject Heading
Derive specific subject heading for the specific subject entry from the last sought link
and moving upwards by taking the necessary and sufficient sought links in a reverse
rendering or backward rendering process. If the subject includes a space isolate, time
isolate or a form isolate, break the chain into different part(s) at the point(s) denoting
space, time and form in the class number. In such a situation, specific subject heading is
to be derived from last SL of first part in reverse rendering process and then by second
part, third part, etc., if any, in the similar process. Then, the components of derived
subject headings are to be arranged in the sequence of their derivation from each part
of the chain of the class number.
5) Preparation of Subject Reference Headings
Derive subject reference heading for the subject reference from each of the upper
sought links. This process continues until all the terms of upper sought links are exhausted
and indexed.
6) Preparation of Subject Reference Entries
Prepare subject reference entries or ‘see also’ references from each subject reference
heading to its specific subject heading. When a subject heading starts from last sought
link denoting space or time or form, prepare ‘See’ references instead of ‘See also’
references from subject reference heading(s) to specific subject heading.
7) Preparation of Cross References, if any
Prepare cross references (i.e. ‘see’ references) for each alternative and synonymous
84 term/heading used in the specific as well as subject reference headings.
8) Alphabetisation Indexing Techniques

Merge specific subject entries, subject references (i.e. ‘see also’ references) and ‘see’
references and arrange them in single alphabetical sequence.
The above noted steps in chain indexing are demonstrated below with illustrative example:
0) Subject of the Document
Researches on Child Psychology in India
1) Class Number of the Subject of the Document
Class no.: 155.4072054 [according to DDC, 22nd Edition]
2) Representation of the Class Number in the form of a Chain
100 Philosophy, Parapsychology and Occultism
150 Psychology
155 Differential and developmental psychology
155. ———
155.4 Child psychology
152.40 ———
152.407 Education, Research related topics
152.4072 Research
152.40720 ———
152.407205 Asia
152.4072054 India
3) Determination of Links
100 Philosophy, Parapsychology and Occultism [USL]
150 Psychology [SL]
155 Differential and developmental psychology [USL]
155. [FL]
155.4 Child psychology [SL]
152.40 [FL]
152.407 Education, Research related topics [USL]
152.4072 Research [SL]
152.40720 [FL]
152.407205 Asia [USL]
152.4072054 India [SL]

85
Indexing 4) Preparation of Specific Subject Heading
Research, Child psychology, India
5) Preparation of Subject Reference Headings
Research, Child psychology
Child psychology
Psychology
6) Preparation of Cross References
India, Research, Child psychology
7) Preparation of Index Entries
Research, Child psychology, India 152.4072054
Bibliographical description and abstracts of the document are to be furnished
under the specific subject heading.
Research, Child psychology 152.4072
See also
Research, Child psychology, India
Child Psychology 155.4
See also
Research, Child psychology, India
Psychology 150
See also
Research, Child psychology, India
India, Research, Child psychology
See
Research, Child psychology, India
8) Alphabetisation
Arrange the above entries according to single alphabetical order.
Advantages
A major advantage of chain indexing is that it ensures the collocation of aspects of a
subject which have been scattered in the classification scheme (i.e. distributed relatives)
because last link in the class number is always the first link in the chain of subject index
entries.
It offers general as well as specific information to all information seekers by deriving
subject headings from the chain of successive subdivisions that leads from the general
to most specific level.
It is more or less a mechanical system and it is economic also.
86
Criticisms Indexing Techniques

 Too much dependency on classification schemes

Chain indexing could be operative effectively only if they are backed up by a good
structured classification scheme. If chain indexing is based on a structurally defective
classification scheme, the subject headings drawn according to chain procedure will
naturally become defective.
 Disappearance of Chain
Chain disappears in each stage of deriving subject reference entries and thus it results in
the loss of full context of the content of the document.
 Lack of Specificity
Chain indexing provides only one specific entry and others are subject references.
 Unsuitability for computerisation
The formation of ‘Chain’ is very much a human intellectual operations which is beyond
the capability of the computer to manipulate.
 Running from pillar to post
It makes direct access to specific subject heading possible only for those who knew the
specific subject heading for the given document, but in most cases access to specific
subject heading was possible at the cost of running from pillar to post since only one
entry is specific subject entry and others are cross references in chain indexing system.
Self Check Exercise
Note: i) Write your answers in the space given below.
ii) Check your answers with the answers given at the end of this Unit.
1) Distinguish between Derived indexing and Assigned indexing.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
2) Discuss the techniques by which J. Kaiser has solved the problems of indexing
compound subject.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
3) What is the epithet of the term ‘chain’ in chain indexing? Why is the class number
of the subject taken as the base for deriving subject headings according to chain
indexing system?
.....................................................................................................................
87
Indexing .....................................................................................................................
.....................................................................................................................
.....................................................................................................................
4) Enumerate the steps involved in chain indexing.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................

11.3.4 PRECIS (Preserved Context Index System)

PRECIS is generally considered as one of the most ambitious late twentieth century
attempts to create an indexing system from scratch. A major break-through in the field
of subject indexing was achieved by Classification Research Group (CRG), London
and Derek Austin came out with a new method of subject indexing called PREserved
Context Index System (PRECIS) in the early 1970s for the British National
Bibliography (BNB). When the BNB began in 1950, it used chain indexing system for
about 20 years. However it was not, for various reasons, ideal for a computerised
system, and in 1971, when BNB had developed the MARC system in the United
Kingdom and was also engaged in using computers for the production of BNB itself,
chain indexing was replaced by PRECIS.
What is PRECIS?
PRECIS is a system of subject indexing in which the initial string of terms organised
according to the scheme of role operators, is computer manipulated in such a way that
each sought term in the string functions as the approach term while preserving the full
context of the document. Entries are restructured at every step in such a way that the
user can determine from the format of the entry which terms set the approach term into
its context and which terms are context dependent on the approach term.
Objectives of PRECIS
a) The computer, not the indexer, should produce all index entries. The indexer’s
responsibility is to prepare the input strings and to give necessary instructions to
the computers to generate index entries according to definite formats.
b) Each of the sought terms should find index entries and each entry should express
the complete thought content / full context of the document unlike the chain
procedure where only one entry is specific—i.e. fully co-extensive with the subject
of the document and others are cross references describing only one aspect of the
thought content of the document.
c) Each of the entry should be expressive.
d) The system should be based on a single set of logical rules to make it consistent.
e) The system should be based on the concept of open-ended vocabulary, which
means that terms can be admitted into the index at any time, as soon as they have
been encountered in the literature.

88
f) The system must have sufficient references between semantically related terms.
Features of PRECIS Indexing Techniques

 It is more amenable to automatic manipulation than indexing based on the notational

classifications.
 The permuted entries read naturally, which is achieved by the prescribed order of
the role operators;
 The terms are linked to a machine-held thesaurus thereby providing possible
‘see’ and ‘see also’ references;
 PRECIS can be adapted to other languages.
 The indexer determines the meaning of the terms codes the roles and identifies the
lead terms, whereas the computer takes care of the permutations.
 Its subject formulation is completely independent of classification, therefore
exclusively geared to no classification numbers assigned in the MARC record.
 Context is preserved: It presents the full subject statement at every point of index
entry, by gradual inversion of the concept string, thus overcoming the problem of
the disappearing chain.
Principles of PRECIS
Two principles are followed in PRECIS:
a) Principle of Context Dependency: The “context-dependency” principle may
be seen as a combination of context and dependency. When this principle is
followed in a PRECIS input string, each term is qualified and sets the next term
into its wider context. In other words, the meaning of each term in the string
depends upon the meaning of its preceding term and taken together, they all
represent the single context. Each term is hence dependent, directly or indirectly,
on all the terms which precede it.
b) Principle of One-to-One Relationship: When the terms are organised according
to the principle of context dependency, they form a one-to-one related sequence—
each of the terms in the string directly related to its next term.
Syntax and Semantics
The syntax of PRECIS is based on the role operators, codes and logical rules which
act as instruction to the computer. The semantics of PRECIS is handled by linking the
terms to a machine-held thesaurus thereby providing possible ‘see’ and ‘see also’
references.
Role Operator: Role operators consist of a set of alpha-numeric notations which
specifies the grammatical role or the function of the indexed term and regulates the
order of terms in the input string. Role operators and their associated rules also serve as
the computer instruction for determining the format, typography and punctuation
associated with each index entry. There are two kinds of role operators: Primary
operators and secondary operators. Primary operators control the sequence of terms
in input string and determine the format of index entries. Any of the secondary operators
is always to be preceded by the primary operator to which it relates.
Codes: Use of codes in the string brings expressiveness in the resulting index entries.
Three types of codes are there: Primary, Secondary and Typographic codes.
89
Indexing Input String: A set of terms arranged according to the role operators which act as
instructions to the computer for generating index entries.
Schema of Role Operators

Primary Operators
Environment of core 0 Location
concepts
Core concepts 1 Key System
Thing when action not present.
Thing towards which an action
is directed, e.g. object of
transitive action, performer of
intransitive action
2 Action; Effect of action
3 Performer of transitive action
(Agent, Instrument); Intake;
Factor
Extra-core concepts 4 Viewpoint-as-form
5 Selected Instance: study region,
study example, sample
population
6 Form of document; Target user
Secondary Operations
Co-ordinate concepts f ‘Bound’ co-ordinate concept
g Standard co-ordinate concept
Dependent elements p Part; Property
q Member of quasi-generic group
r Assembly
Special classes of action s Roll definer; Directional
property
t Author-attributed action
u Two-way interaction

Schema of Codes

Primary codes
Theme Interlinks $x 1st concept in coordinate
theme
$y 2nd/subsequent concept in co-
ordinate theme
$z Common concept
Term Codes $a Common noun
$c Proper name (class of-one)
$d Place name

90
 Secondary codes Indexing Techniques
Differences
Preceding differences (3 characters)—
1st and 2nd characters:
$0 Non-lead, space generating
$1 Non-lead, close-up
$2 Lead, space generating
$3 Lead, close-up
rd
3 character: Number in the range, 1 to 9 indicating level of difference
Date as a difference $d
Parenthetical difference
$n Non-lead parenthetical difference
$o Lead parenthetical difference
Connectives
$v Downward reading connective
$w Upward reading connective
Typographic codes
$e Non-filing part in italic preceded by comma
$f Filing part in italic preceded by comma
$g Filing part in roman , no preceding punctuation
$h Filing part in italic preceded by full point
$I Filing part in italic , no preceding punctuation

Entry Structure of PRECIS

The entry structure of PRECIS string consists of a two-dimensional display rather than
the one dimensional that we have been accustomed to; instead of putting everything on
one line, so that the only relationship which could be shown was that of following or
proceeding, PRECIS uses two-line-three part entry structure as follows:

Lead Qualifier

Display
Lead is occupied by the approach term, which is the filing word and is offered as the
user’s access point in the index.
Qualifier position is occupied by the term(s) that sets the lead into wider context (i.e.
general to specific). Together, the Lead and the Qualifiers correspond to the Heading.
Terms in the heading set down in a narrower-to-wider context order. When the first
term of the input string appears in the Lead position, the Qualifier position is usually
kept blank.
Display position is occupied by those additional set of qualifying terms of the PRECIS
string, which rely upon the heading for their context. When the last term of the input
string appears in the Lead position, the Display position becomes empty.
Steps in PRECIS
Let us take the following title of a document for demonstrating the different steps of
subject indexing according to PRECIS:
91
Indexing 0. Title: University libraries in West Bengal
1. Analysis: Involves analysis of the thought content of the document and formulating
the subject statement in natural language:
Measurement of the performance of university libraries in West Bengal
2. Preparation of Input String: Involves the identification of the status or role of
each component term denoting key concept in terms of the role operators of PRECIS
and assigning the appropriate operators to prepare an input string. The stages of the
preparation of input string are furnished below:
[For understanding the stages of the preparation of input string, students are required to
consult the schema of role operators and codes as furnished under the sub-section
11.3.4. of this Unit]
a) Identifying the concept signifying an action (if there be any).
In the present example, the action concept is denoted by the term ‘Measurement’
and this term should, therefore be prefixed by the role operator (2).
b) Identifying the kind of action represented by this term, i.e. whether transitive action
or intransitive action. ‘Measurement’ is clearly a transitive action since it is capable
of taking an object. The object of transitive action is considered as the key system
and is coded by the operator (1). In the present example, it is the ‘Performance’
which is being measured, so the input string should now appear in the form:
(1) performance
c) Identifying the concept, if any related as property and/or whole-to-part. In this
example, ‘performance’ is the property of ‘university libraries’ and ‘libraries’ is the
part of the ‘universities’. As a result, we will get the following input string:
(1) universities
(p) libraries
(p) performance
(2) measurement
d) The remaining term ‘West Bengal’ signifies the environment (i.e. geographical
location) in which the whole thing takes place. Accordingly, the operator (0) is to
be prefixed to the concept and the resulting input string will be:
(0) “West Bengal
(1) “universities
(p) “libraries
(p) “performance
(2) “measurement
Note: The primary operators have ordinal filing values and the terms in the above
input string are sequenced accordingly. Thus, component terms are organised into
the above input string according to the principle of context dependency. The
secondary operator (p), prefixed with ‘libraries’ and denoting ’part’ of the
92
‘universities’, is preceded by the primary operator (1) to which it is related. Similarly,
 the secondary operator (p), prefixed with ‘performance’ and also denoting Indexing Techniques
‘property’ of the ‘libraries’, is preceded by the secondary operator (p) meant for
‘libraries’ to which it is related. The terms, except proper name (i.e. West Bengal),
are written in lower case initials. Index entries will be generated in upper case
initials by the computer. Tick mark (“) is to be provided for the terms which should
appear as Lead (access points) in the index entries.
3) Generation of Index Entries: The first index entry will be generated by the
computer by pushing the first term of the input string into the Lead position and
keeping the remaining terms in the Display position. As soon as any term goes to
the Lead position, it is printed in bold type face.
West Bengal
Universities. Libraries. Performance. Measurement
The second index entries will be generated by pushing the second term of the input
string into the Lead position and thereby replacing the existing Lead term into the Qualifier
position, such as:
Universities. West Bengal
Libraries. Performance. Measurement
Similarly other index entries will be generated as
Libraries. Universities. West Bengal
Performance. Measurement
Performance. Libraries. Universities. West Bengal
Measurement
Measurement. Performance. Libraries. Universities. West Bengal
Note: It can now be seen in the above examples that Lead and Qualifier are
separated by a full stop and 2-letter space. The standard separator between two
terms in the entry is full stop and one space. The terms in the Display position are
written leaving 2-letter space from the left. For over-run of Display in the next line,
4-letter space and for over-run of heading in the next line 8-letter spaces are to be
left from the margin.
4) Generation of supporting reference entries: ‘see’ and ‘see also’ references
are generated from semantically related terms taken from a machine-held thesaurus.
5) Alphabetisation: All the entries, generated by the process, as stated above, are
arranged alphabetically by headings. Under the common heading, displays are
organised alphabetically.
Formats of PRECIS Index
Index entries in PRECIS are basically generated in three formats: standard format,
inverted format and predicate transformation.
a) Standard Format: Index entries in the standard format are generated when any
of the primary operators (0), (1), and (2) or its dependent elements appear in the
Lead. The process of generation of index entries in the standard format has already
been demonstrated under the Section 11.3.4.7 of this Unit.
93
Indexing b) Inverted Format: Index entries in the inverted format are generated whenever a
term coded by an operator in the range from (4) to (6) or its dependent elements
appear in the Lead. The rule relating to the generation of index entries with this
format is that—when any of the terms coded either (4), or (5) or (6) or any of their
dependent element operators appear in the Lead, the whole input string is read
from top to bottom and is written in the Display. However, if the term appearing in
the Lead is last term of the input string, then it will be dropped from the Display.
Example: A report on the feminist viewpoint on marriage
Input String:
(2) marriage
(4) feminist viewpoint
(6) reports
Index Entries:
Marriage
– Feminist viewpoint – Reports
Feminist viewpoint
– Marriage – Feminist viewpoint – Reports
Reports
– Marriage – Feminist viewpoint
c) Predicate Transformation: When an entry is generated under a term coded (3)
that immediately follows a term coded either by (2) or (s) or(t)—each of which
introduces an action of one kind or another—the predicate transformation takes
place. An input string of this kind is shown below:
Example: Planning of libraries by architect
Input String:
(1) libraries
(2) planning $v by $w of
(3) architects
In order to bring expressiveness in the resulting index entries, the connective codes $v
and $w (see ‘schema of codes’) are attached to the action concept and it results in a
compound phrase. The rule relating to the generation of index entries with this format is:
when the term coded (3) goes to the Lead, the computer checks the operator assigned
to the next preceding term. If that operator is (2) or (s) or (t), the term coded with any
of these operators and the term accompanied by the Code $w for upward reading
connective (if any) are printed in the Display position instead of Qualifier position.
Accordingly, the index entries for the aforesaid input string will be:
Libraries
Planning by architects

94 Planning. Libraries
 By architects Indexing Techniques

Architects
Planning of libraries

11.3.5 POPSI
All pre-coordinate indexing models are entirely based on the method of facet analysis.
Ranganathan pointed out in a paper entitled ‘Subject heading and facet analysis’
(Journal of Documentation, 20 (3), 1964, p.109-119) that facet analysis does not
depend entirely on notational scheme of classification. The rules of chain procedure, he
said, can be so framed as to implement any kind of decision about the sought first
heading and the other successive headings in conformity with the principle of local
variation. Since then, continuous research on this new line of thinking was going on at
Documentation Research and Training Centre (DRTC), Bangalore and a number of
papers on Postulate-based Permuted Subject Indexing (POPSI) based on
Ranganathan’s General Theory of Library Classification came out. Dr. Ganesh
Bhattacharyya first explained the fundamentals of subject indexing languages with an
extensive theoretical background which ultimately led to the development of newer
version of POPSI forming the part of his General Theory of Subject Indexing Languages
(GT-SIL). Bhattacharyya developed the POPSI through logical interpretation of the
deep structure of subject indexing language (SIL). POPSI drew attention to the
helpfulness of adopting a suitable device for ensuring an optimally effective organising
classification through the alphabetisation of verbal subject – propositions. It prescribes
the use of apparatus words – such as prepositions, conjunctions, participles etc., as
and when necessary to communicate the exact meaning of subject propositions. These
words are put in parenthesis and they are ignored in alphabetisation. Since the POPSI
Index of all verbal entries, filing them in one alphabetical sequence in a unipartite index
is made easy.
Major Working Concepts of POPSI
1) Deep Structure of Subject Indexing Languages (DS-SIL)
DS-SIL is the logical abstraction of the surface structures of outstanding SILs like
Cutter, Dewey, Kaiser and Ranganathan. According to the General Theory of SIL, the
structure of a specific SIL has been assumed to be a surface structure of the deep
structure of SIL. The DS-SILs has been presented diagrammatically as follows:

95
Indexing It appears from the above diagram that any specific subject may belong to any one of
the following elementary categories (D, E, A, P) and modifier:
2) Elementary Categories and Modifiers
a) Discipline (=D) refers to an elementary category that includes the
conventional field of study, or any aggregate of such fields, or artificially created
fields analogous to those mentioned above; e.g. Physics, Biotechnology,
Ocean science, Library and Information Science, etc.
b) Entity (=E) refers to an elementary category that includes manifestations
having perceptual correlates, or only conceptual existence, as contrasted
with their properties, and actions performed by them or on them; e.g. Energy,
Light, Plants, Animals, Place, Time, Environment, etc.
c) Action (=A) refers to an elementary category that includes manifestations
denoting the concept of ‘doing’. An action may manifest itself as Self Action
or External Action. For examples: Function, Migration, etc. are Self Actions;
and Treatment, Selection, organisation, and Evaluation, etc. are External
Actions.
d) Property (=P) refers to an elementary category that includes manifestations
denoting the concept of ‘attribute’—qualitative or quantitative; e.g. Property,
Effect, Power, Capability, Efficiency, Utility, Form, etc.
e) M=Modifier refers to a qualifier used to modify any one the elementary
categories D, E, A and P. It decreases the extension and increases the intension
of the qualified manifestation without disturbing its conceptual wholeness. A
modifier can modify any one of the elementary categories, as well as two or
more elementary categories. Modifiers are of two types:
 Common Modifiers: They refer to Space (e.g. Libraries in India),
Time (e.g. Libraries in India 19th Century), Environment (e.g. Desert
Birds), and Form (e.g. Encyclopedia of Physics). Common modifiers
have the property of modifying a combination of two or more elementary
categories.
 Special Modifiers: A special modifier is used to modify only one of
the elementary categories. It may be of Discipline-based, or Entity-
based, or Property-based, or Action-based. Special modifiers can be
grouped into two types:
i) those that require a phrase or auxiliary words to be inserted between
the term and thus forming a complex phrase, e.g. Cataloguing using
computers; and
ii) those that do not require auxiliary words or phrase to be inserted in
between the terms, but automatically form an acceptable compound
term denoting Species/Type, e.g. ‘Chemical’ in ‘Chemical
Treatment’.
3) Organising Classification and Associative Classification
According to the General Theory of SIL, classification is a combination of both organising
classification and associative classification. In other words, an indexing system is a
combination of both organising classification and associative classification. The tasks
96
involved in creating an organising classification are the categorisation of concepts and Indexing Techniques
their organising in hierarchies. In organising classification compound subjects are based
on genus-species, whole-part, and other inter-facet relationships. Here, classification is
used to distinguish and rank each subject from all other subjects with reference to its
Coordinate—Superordinate—Subordinate—Collateral (COSSCO) relationships. The
result of organising classification is always a hierarchy. In associative classification, a
subject is distinguished from other subjects based on the reference of how it is associated
with other subjects without reference to its COSSCO relationships. The result of
associative classification is always a relative index.
4) Base and Core
In the context of constructing compound subject heading, when the purpose is to bring
together all or major portion of information relating to a particular manifestation or
manifestations of a particular elementary category, the manifestation/category is Base.
In case of a complex subject, any one of the subjects can be decided to be the Base
subject depending upon the purpose in hand. For example: for a document on ‘Eye
cancer’, ‘Eye’ is the Base subject in an Eye Hospital Library, and ‘Cancer’ is to be
considered as the Base subject for a Cancer Research Centre.
When the purpose is to bring together within a recognised Base, all or major portion of
information pertaining to one or more elementary categories, the category or categories
concerned is the Core of the concerned Base. Core lies within the Base, and which one
will be the Base or Core depends on the collection or purpose of the library. For
example: In DDC, ‘Medicine’ is the Base, and the ‘Human body’ and its ‘Organs’
constitute the Core of the Base.
Features of POPSI’
From the operational point of view, the salient features of POPSI may be grouped
under three components: Analysis, Synthesis, and Permutation.
The work of ‘Analysis’ and ‘Synthesis’ is primarily based on the postulates associated
with the deep structure of SILs for generating organising classification. The task of
analysis and synthesis is largely guided by the following POPSI-table. The work of
‘Permutation’ is based on cyclic permutation of each term-of-approach, either
individually or in association with other terms for generating associative classification
effect in alphabetical arrangement.
Rules of Syntax
The basic rules of syntax associated with POPSI are:
a) Discipline is followed by Entity, both modified and unmodified.
b) Property follows immediately the manifestation in relation to which it is a Property.
c) Action follows immediately the manifestation in relation to which it is an Action.
d) A Property can have its own Property.
e) An Action can have its own Action.
f) A Species/Type follows immediately the manifestation in relation to which it is a
Species/Type.
g) A Part follows immediately the manifestation in relation to which it is a Part.
97
Indexing h) A Modifier follows immediately the manifestation in relation to which it is a Modifier.
The following POPSI Table, like Role Operators in PRECIS, is used in sequencing the
component terms for formulating a subject heading

0 Form modifier
1 General Treatment
2 Phase relation
2.1 General
2.2 Bias
2.3 Comparison
2.4 Similarity
2.5 Difference
2.6 Application
2.7 Influence
Common modifiers
3 Time modifier
4 Environment modifier
5 Place modifier
6 Entity (E) .1 Action (A) , Part
7 Discipline (D) .2 Property (P) . Speciator/Type
Note: Notations —Special modifier
.1 and .2 are Note: A
preceded by the Species/Type/ Special
notation of the modifier follows
manifestation in immediately the
relation to manifestation in
which it is (A) relation to which it is
and (P). a Species/Type.

8 Core I
9 Base (B)
Note: Features relating to Core I and Base (B) are analogous to
6 Entity / 7 Discipline /.1 Action / .2 Property.

Semantic Relationship in POPSI

Semantic relationship in POPSI is controlled by the vocabulary control device called
‘Classaurus’. A classaurus is an elementary category-based (faceted) systematic scheme
of hierarchical classification in verbal plane incorporating all the necessary features of a
conventional information retrieval thesaurus. Like any faceted classification scheme, a
classurus consists of separate schedule for each of the Elementary Categories, namely,
Entity, Property and Action, with their Species/Parts and Special Modifiers for each.In
98
each schedule, it displays hierarchical relationship among terms (broader, narrower,
and collateral). Classaurus also contains separate schedules for Common Modifiers Indexing Techniques
like Form, Time, Place, and Environment. Like any thesaurus, each of terms in hierarchic
schedule is enriched by synonyms, quasi-synonyms, etc. Unlike a thesaurus, a classaurus
does not include associatively related terms (RTs) because of its category-based (faceted)
structure. The implication of the faceted structure is that a term in one Elementary
Category has a high chance of being non-hierarchically related with another term in
another category. The task of showing what is non-hierarchically related to what, and
how they are related is left to the care of indexing procedure based on the information
contained in the document itself. All the component terms associated with the content
of the document are brought together under each approach-term by the permutation
technique of POPSI. It is assumed that RTs should not be dictated beforehand by the
designer of the classaurus, rather it should be dictated by the content of the document
itself. Any term may be related to any other terms depending on the content of the
document and hence, RTs should not be determined beforehand.
Steps in POPSI
The main steps in applying POPSI are as follows:
1) Content Analysis: Involves identification of different component ideas associated
with the content of the document with reference to their elementary categories and
modifiers.
2) Formalisation: Involves preparing the formalized expression of subject statement
obtained on the basis of the results of the step 1 (Content Analysis) according to
the rules of syntax.
3) Standardisation: Deciding the standard term in the formalized expression of
subject statement especially for the term(s) having synonym(s), if any. This step
calls for the use of Classaurus.
4) Modulation: Involves augmenting the standardized subject proposition by
interpolating and extrapolating, as the case may be, the successive superordinates
of each manifestation by using standard terms with indication of their synonyms, if
any. [Note: A Classaurus is the tool to guide the operation in steps 3 and 4 with
assurance of consistency in practice].
5) Entry for Organising Classification: Involves the preparation of entries for
organising classification by inserting appropriate notations for Elementary
Categories, subdivisions and modifiers from POPSI Table. Modulated subject
proposition with appropriate notations from POPSI Table sorted alpha-numerically
will produce organising classification effect by juxtaposition of entries in alphabetical
sequence.
Approach-term Selection: Consists of deciding the approach-terms for generating
associative classification effect and of controlling synonyms. The selection of
approach-terms may vary for one library to another library depending upon the
requirement of users.
6) Preparation of Entries of Associative Classification: Involves the preparation
of the entries under each approach-term by cyclic permutation of sought terms for
generating associative classification effect in alphabetical arrangement.
7) Alphabetisation: Involves in arranging all the entries in alphabetical sequence.

99
Indexing Demonstration of the Procedure of POPSI-Basic
0) Title of the document:
Use of computers for indexing of educational films in university libraries in India
1) Content Analysis:
Library and Information Science = Discipline (D) [Implicit]
University libraries = Entity (E) [Explicit]
Educational films = Part of E [Explicit]
Indexing = A of Part of E [Explicit]
Use = Application phase relation (PR) [Explicit]
Computers = E-based Special modifier (Sm) [Explicit]
India = Common modifier (Cm) of place [Explicit]
2) Formalisation:
Library and Information Science (D), University libraries (E), Educational films
(Part of E), Indexing (A of Part of E), Use (Application PR), Computers (E-
based Sm), India (Cm)
3) Standardization:
It is assumed that all the terms in the formalized expression of the subject as shown
above are standard terms.
4) Modulation:
Library and Information Science (D), Libraries. Academic libraries. University
libraries (E), Information sources. Films. Educational films (Part of E), Indexing
(A of Part of E), Use (Application PR), Computers (E-based Sm), Asia, India
(Cm)
5) Entry for Organising Classification:
Library and Information Science 6 Libraries. Academic libraries. University libraries,
Information sources. Films. Educational films 6.1 Indexing 2.6 (using) Computers
5 (in) Asia, India
6) Approach-term Selection:
The following terms are selected as the approach-terms:
Libraries
Academic libraries
University libraries
Information sources
Films
Educational films
Indexing
Computers
India
100
7) Preparation of Entries of Associative Classification: Indexing Techniques

LIBRARIES
Library and Information Science 6 Libraries. Academic libraries. University libraries,
Information sources. Films. Educational films 6.1 Indexing 2.6 (using) Computers
5 (in) Asia, India
The above organising classification will have to be repeated under each of the
following approach terms:
Academic libraries
University libraries
Information sources
Films
Educational films
Indexing
Computers
India
8) Alphabetisation:
All the entries are arranged according to the alphabetical order, word-by-word.
POPSI-Specific
The steps in POPSI with illustrative examples as demonstrated above fall within the
purview of POPSI-Basic. According to Bhattacharyya, there is no single absolute version
of organising or associative classification. POPSI tries to find out what is logically basic,
and amenable to systematic manipulation to meet specific requirement. The POPSI-
Basic is a product of the application of the GT-SIL and it is readily amenable to the
systematic manipulation to generate purpose-oriented specific versions known as
POPSI-Specific. POPSI-Specific is always a derivation from the POPSI-Basic
according to special decisions and rules to meet specific requirements at the local level.
It may be noted here that this approach is totally different from that of earlier contributors
of different SILs.
If the purpose is to bring together all or a major portion of information pertaining to a
specific topic in a discipline manifesting any of the Elementary categories, the above
version of POPSI-Basic can be systematically manipulated to generate the required
version of POPSI-Specific. This involves the decision about the ‘Base’ and ‘Core’.
For example, our purpose is to bring together all information pertaining ‘Educational
films’ in one place and hence, ‘Educational films’ is to be considered as the Base.
‘University libraries’ and ‘Indexing’ are to be considered as the Modifier and Action to
Base respectively. In view of this, we can prepare the following organising classification
entry:
Information sources. Films. Educational films — Libraries. Academic libraries. University
libraries 9.1 Indexing 2.6 (using) Computers 5 (in) Asia, India.

101
Indexing Self Check Exercise
Note: i) Write your answers in the space given below.
ii) Check your answers with the answers given at the end of this Unit.
5) Discuss the Principle of Context Dependency.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
6) Discuss the entry structure and entry format as followed in PRECIS?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
7) How do you categorise different operational stages of POPSI?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
8) What are the major steps in formulating index entries according to POPSI?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................

11.4 POST-COORDINATE INDEXING

All indexing systems follow the process of concept coordination to describe the contents
of the documents more precisely. We have seen in pre-coordinate indexing, component
concepts are coordinated according to the order of significance or citation order by
following the syntactical rules of the given indexing language. But the rigidity of the
citation order appears to be unsatisfactory to meet the varieties of approaches of all the
users. The provision for multiple entries in pre-coordinate indexing by rotating or cycling
of the component terms covers only a fraction of the possible number of the total
permutations and for this, a large portion of probable approach points is left uncovered.
Consequently, the searcher has no choice but to follow the rigid citation order specified
by the given indexing language. The above noted problems stemming from the pre-
coordination of terms with the rigidity of citation order triggered the development of
alternative indexing techniques where the component ideas of a subject are kept
separately, uncoordinated by the indexer. Here, concepts/terms are coordinated at the
102 time of searching (i.e. at output stage) by the user. A greater degree of search manipulation
is available in post-coordinate indexing system since the search terms can be coordinated Indexing Techniques
almost in any combination or order to retrieve records of information about the documents
as required by the users. The indexing systems which are based on this basic principle
are called Post-coordinate Indexing or simply Coordinate Indexing Systems. Based
on this basic principle a number of post-coordinate indexing systems like Uniterm,
Optical Coincidence Card, etc. were developed. Among the different types of post-
coordinate indexing systems, Uniterm system developed by Mortimer Taube is considered
as the most popular post-coordinate indexing model.

11.4.1 Pre-Coordinate Indexing versus Post-Coordinate

Indexing
Subjects of documents are not simple. There are compound and complex subjects
dealing with multiple numbers of concepts. When there is more than one concept in the
document the order in which we cite the concepts and their relationship to one another
become important. Both pre- and post-coordinate indexing systems are, by nature,
coordinate indexing, but the coordination is done in two different stages. The following
table furnishes the points of differences between the two systems:

Pre-coordinate indexing system Post-coordinate indexing system

Coordination of component terms Component concepts (denoted by
is carried out at the time of the terms) of a subject are kept
indexing (i.e. at input stage) in separately uncoordinated by the
anticipation of the users’ approach. indexer, and the user does the
coordination of concepts at the time
of searching (i.e. at the output
stage).
The most important aspect of this Rigidity of the significance order is
indexing system is to determine the very much absent in this system.
order of significance by following
the syntactical rules of the given
indexing language.
It is non-manipulative. The It is manipulative. The searcher has
searcher has no choice but to try to wide options for free manipulation
predict the citation order specified of the classes at the time of
by the indexer. searching in order to achieve
whatever logical operations are
required.
In this indexing system, both the This indexing system does not
indexer and the searcher are require the indexer and searcher to
required to understand the understand the mechanism of the
mechanism of the system—the system. However, the operational
indexer for arriving at the most aspects n need to be understood by
preferred citation order and the them.
searcher for formulating an
appropriate search strategy—in
order to achieve the highest
possible degree of matching of
concepts.

103
Indexing 11.4.2 Term Entry System and Item Entry System
In Term Entry System, we prepare entries for a document under each of the appropriate
subject headings, and file these entries alphabetically. Here, terms are posted on the
item (i.e. Term on Item System). In this type of post-coordinate indexing, the number of
entries for a document is dependent on the number of terms associated with the thought
content of the document. Searching of two files (Term Profile and Document Profile) is
required in this system. Uniterm and Peek-a-boo are examples of these.
It is possible to take the opposite approach and make a single entry for each item, using
a physical form which permits access to the entry from all appropriate headings. A
system which works in this way is called an Item Entry System. Here, items are
posted on the term (i.e. Item on Term System). In this type of post-coordinate indexing,
single entry is made for each item. Item entry system involves the searching of one file
(i.e. Term Profile) only. Edge-notched Card is an example of item entry system.

11.4.3 Uniterm Indexing

Uniterm indexing system was devised by Mortimer Taube in 1953 to organise a collection
of documents at the Armed Services Technical Information Agency (ASTIA) of Atomic
Energy Commission, Washington. Uniterm is a post-coordinate indexing system based
on term entry principle. Here, component term (uniterm) is independent of all other
terms and serves as a unique autonomous access point to all relevant items in the
collection.
Indexing Process
The processes involved in Uniterm indexing are as follows:
1) Preparation of Document Profile: Here indexer is required to prepare a card
for each incoming document to:
 assign an accession number or ‘address’ to each incoming document and
record the same on the card,
 identify the different component terms (uniterms) associated with the thought
content of the document,
 select the uniterms and record them on the card,
 prepare an abstract for each document and record the abstract on the card,
 write the bibliographical details of the document on the card, and
 arrange all the cards as prepared for all incoming documents according to
the ascending order of the accession numbers of the documents.
Cards so prepared and arranged according to the ascending order of the accession
numbers of the documents are called Documents Profiles.
2) Preparation of Term Profile: Here indexer is required to
 prepare a card (term card) for each uniterm,
 divide each term card into 10 equal vertical columns (from 0 to 9),
 record the accession number of the document relevant to the uniterm
according to the system of terminal digit posting. Terminal digit of the accession
number determines the column of its posting, and

104  arrange all term cards according to the alphabetical order of the uniterms.
Term cards so prepared as stated above and arranged according to the alphabetical Indexing Techniques
order of the uniterms are called the Term Profile of the system.
Searching
The process of searching in Uniterm system involves the following operations:
 Searcher identifies the different component terms/uniterms associated with the
content of her/his queries.
 Searcher, after identifying the component terms/uniterms, pull out the pertinent
term cards from the alphabetical deck of the Term Profile.
 Term cards thus pulled out are matched to find the common accession number(s).
The number(s) common in all such Uniterm cards represent the sum total of the
component concept of the specific subject.
 With the help of the common accession number(s), relevant card(s) are pulled out
from the Document Profile where full bibliographical information of the required
document(s) is available.
The main advantage of the Uniterm indexing system is its simplicity and the ease with
which persons without much knowledge of subject indexing can handle it. The criticisms
against Uniterm system centre around: (a) search time: involves much searching time
because of the searching of two files—Term Profile and Document profile; and (b)
false drops: possibility of retrieving irrelevant documents due to false coordination of
uniterms. For example, searching with the uniterms ‘Teachers’, ‘Students’ and
‘Evaluation’ may retrieve documents on both the subjects, ‘Evaluation of students by
teachers and ‘Evaluation of teachers by students’, one of which might be irrelevant to a
particular user.
To overcome the problem of false drops, the following post-coordinate searching devices
have been used:
1) Use of Pre-coordinated Terms
It is the introduction of pre-coordination to some extent in post-coordinate system
in which two or more terms in a subject are bound in place of isolated single term/
uniterm to get rid of false coordination.
2) Links: Links are special symbols used to group all the related concepts in a
document separately, so that inappropriate combinations of terms are not retrieved.
Suppose we have a document (accession number: 243) dealing with two different
topics—Classification of non-book materials and indexing of films. In order
to avoid false coordination like Classification of films and indexing of non-
book materials, alphabetical symbols, which serve as interlocking device, are
attached to accession number to indicate different groups:
Classification 243A
Non-book Materials 243A
Indexing 243B
Films 243B
3) Roles: Roles are the indicator digits/symbols attached to the terms at the time of
indexing to indicate the role or status or use of the term in a particular context.
Here, the possible roles of different terms are identified beforehand and terms are 105
Indexing tagged with these role indicators at the time of indexing. For example, roles
developed by the Engineering Joint Council, known as EJC role operators may be
attached to ‘Television’ to distinguish its functions as the product and tool:
Role Document
2 [Product / Output] Manufacturing of television
3 [Agent / Tool] Use of television in education
4) Weighting: It is the device of allocating quantitative values to the index terms
according to their degree of relevance in the document. Different ways for indicating
weights have been suggested. A simple system uses numbers 1 to 3, where 3
indicates maximum weight (i.e. the index term is highly specific and covers an
entire major subject of the document), 2 and 1 indicate weights of index terms in
decreasing order of their values or relevance in the document.

11.5 KEYWORD INDEXING

Keyword indexing is based on the usage of natural language terminology for generating
the index entries. The term ‘Keyword’ refers to a significant or memorable word (also
called ‘catchword’) that serves as a key in denoting the subject taken mainly from the
title of the document (so, it is called title index) and sometimes from the abstract or
text of the document. Common words like articles (a, an, the) and conjunctions (and,
or, but) are not treated as keywords because it is inefficient to do so. This system is also
known as Natural or Free Indexing language. It is to be pointed out here that the
concept of keyword indexing is not new and it existed in the nineteenth century as a
‘catchword indexing’. With the introduction of computers in information retrieval in the
1950s, Hans Peter Luhn, an IBM engineer, presented a computer-produced index in
1958 that became known as KWIC (Key Word In Context) indexing.

11.5.1 Key Word in Context Indexing (KWIC)

The production of a KWIC index by H. P. Luhn is the earliest example of an automatic
index produced using computers to perform repetitive tasks associated with subject
indexing. It was a great step forward in the technique of automatic indexing. Utilizing
the capabilities of computers, the KWIC method speedily and with a minimum of
intellectual effort produces indexes derived solely from the titles of the documents to be
analysed. All significant or key words of titles/title like phrases are alphabetised
mechanically, and then printed out in turn following a format which emphasises the
selected word. The computer uses the ‘stop-word’ list in order to ignore all syntactical
words such as articles; prepositions etc., and select the remaining words in the title as
indexing words. The remaining words of the title are arranged to stand in the context of
their original appearance. The use of ‘stop-word list’ reduces the volume of the index.
The result of the machine manipulation is an index of key terms printed in alphabetical
order, together with the text immediately surrounding each term. Each significant word
as entry point appears in the margin or a designated middle position while the rest of the
title printed on either side. The alphabetical filing is done on the basis of the key word
printed in bold letters. Chemical Titles (of Chemical Abstract Service) and BASIC
(Biological Abstracts Subjects in Context) are faithful adaptation of Luhn’s KWIC
indexing.
Let us consider the following title ‘Chemical treatment of cancer in the hospitals of
Chennai’ to demonstrate the index entries generated according to KWIC indexing:
106
 Cancer in the hospitals of Chennai / Chemical treatment of 614 Indexing Techniques

Chemical treatment of cancer in the hospitals of Chennai 614

Chennai / Chemical treatment of cancer in the hospitals of 614
Hospitals of Chennai / Chemical treatment of cancer in the 614
Annotation
a) Title in the above KWIC index has been rotated in such a way that each keyword
serves as the approach term and comes in the beginning by rotation followed by
rest of the title;
b) Last word and first word of the tile are separated by using a symbol say, stroke
[ / ] (sometime an asterisk ‘*’ is used) in an entry. In some computer-produced
KWIC indexes, keywords are positioned at middle of the entry;
c) Keywords are printed in bold type face to bring prominency in the approach term;
d) Identification / location code 614 is given at the right end of each entry; and
e) Entries are arranged alphabetically by keywords.
Advantages and Disadvantages of KWIC
KWIC method offers the following advantages: (1) the detection of formulaic expressions
and repeated word patterns; (2) simplicity; (3) speed; (4) maximisation of computer
use; and (5) minimisation of the indexer’s role.
The most common type of complaint against the KWIC indexing method is the lack of
terminology control as it is entirely dependent upon titles/abstract/text of the document.
Apart from this, KWIC has basically two problems: (i) KWIC shows sentences which
contain distant dependency as different context, and (ii) KWIC also shows sentences
which have different word order as different context. The effects of computer’s inability
to resolve these problems led to the redundancy, scatter of references throughout the
index, haphazard groupings and retrieval losses because the user is forced to guess at
the terminology the author actually used. The disadvantages of KWIC-type index can
be summarized as follows:
— Large number of index entries under a given keyword, which provokes difficulties
in searching;
— Lack of significant words in titles (therefore the title and the abstract are often
used as a source for indexing to increase the depth and range of indexing);
— No cross references, which make it difficult to find synonyms, spelling variants
and inflections in the index;
— Relatively high computing time, due to superfluous non-significant index entries;
— No combination of keywords;
— Lack of consistency in the indexing terms, because different authors can use different
form of words to communicate the same idea, or give different meanings to the
same word or phrase;
— References are not grouped under a convenient heading;
— Redundancy of the index. 107
Indexing 11.5.2 Variations of Keyword Indexing
A number of varieties of keyword index appear in the literature and they differ only in
terms of their formats but indexing techniques and principle remain more or less same.
Some important versions of keyword indexing are discussed below:
Key Word Out of Context (KWOC)
In KWOC, each index word is extracted from its context and printed separately in the
left hand margin with the unmodified title in its normal order printed to the right. For
example:
Cancer Chemical treatment of cancer in the hospitals of Chennai 614
Chemical Chemical treatment of cancer in the hospitals of Chennai 614
Chennai Chemical treatment of cancer in the hospitals of Chennai 614
Hospitals Chemical treatment of cancer in the hospitals of Chennai 614
Sometimes, keywords are positioned and printed as heading and the title is printed in
the next line under the heading instead of the same line as shown above.
Key-Word Augmented-in-Context Index (KWAC)
It has been observed that the dependency of keyword indexing on titles sometimes fails
to represent the thought content of the document co-extensively. In order to solve this
problem, KWAC index came into being. In KWAC, the keywords of the title are
enriched with additional keywords taken either from the abstract or from the original
text of the document and are inserted into the title or added at the end to generate
further index entries. KWAC is nothing but the enrichment of KWIC or KWOC. Further
enriched KWIC or KWOC gives index entries wherein additional terms are inserted
into the title or added at the end. This involves intellectual effort in the selection of
additional terms. CBAC (Chemical Biological Activities) of BIOSIS uses KWAC index
where title is enriched by another title like phrase formulated by the indexer.

11.5.3 Double KWIC

It is another improved version of KWIC. Double KWIC index is constructed in the
following way:
a) The first significant word in a title is extracted as a main index term and replaced
by an asterisk (*) to indicate its position in the title.
b) The remaining words in the title are then rotated, so as to permit each significant
word to appear as the first word of a wrap-around subordinate entry under the
main index term. Steps 1 and 2 are repeated until all of the titles of a given
bibliographic listing are processed. The index entries so created are then sorted
alphabetically, both with regard to main terms (primary sort) and subordinate terms
(secondary sort). Main index terms are not restricted to single words, but may
consist of multi-word terms derived from contiguous sets of words in the titles, for
example:

108
ANNUAL REPORT Indexing Techniques
EDITORIAL:*
1966=………………………………………………………………………... F 1
ANNUAL REVIEW
BOOK REVIEW:* OF INFORMATION SCIENCE AND TECHNOLOGY= B3-2.
INFORMATION SCIENCE AND TECHNOLOGY= BOOK REVIEW:*OF B3-2
SCIENCE AND TECHNOLOGY= BOOK REVIEW:* OF INFORMATION B3-2
TECHNOLOGY= BOOK REVIEW:* OF INFORMATION SCIENCE AND B3-2

11.5.4 Other Versions

In addition to the above variations in keyword indexing, a number of varieties of keyword
index are available and they differ only in terms of their formats but indexing techniques
and principles are more or less the same. They are:
i) KWWC (Key-Word-With-Context) Index: In KWWC, only the part of the
title, instead of full title, relevant to the keyword is considered as entry term.
ii) KEYTALPHA (Key-Term Alphabetical) Index: The KEYTALPHA is just
modified form with key terms arranged alphabetically. It is permuted subject
index that lists only keywords assigned to each abstract. Keytalpha index is being
used in the ‘Oceanic Abstract’.
iii) WADEX (Word and Author Index): It is an improved version of KWIC index
where along with the key words, the names of authors are also treated as keywords
and thus indexed accordingly. Thus, it appears that WADEX satisfies both the
author and subject. WADEX is used in ‘Applied Mechanics Review’. AKWIC
(Author and keyword in context) index is another version of WADEX.
iv) KLIC (Key-Letter-In-Context) Index: This type of index only takes fragmented
word (i.e. key letters), instead of the full word, either at the beginning or at the end
of the entry. In this system, the key letters forming the part of the word are specified
and the computer retrieves any term containing that letters either at the beginning
or at the end of the word. KLIC indexes are almost unknown today, the Chemical
Society (London) published a KLIC index as a guide to truncation.
Self Check Exercise
Note: i) Write your answers in the space given below.
ii) Check your answers with the answers given at the end of this Unit.
9) What do ‘Pre-’ and ‘Post-’ signify in Pre-coordinate and Post-coordinate indexing
systems?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
10) What are the search devices used to avoid false coordination of terms in Post
coordinate indexing?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
..................................................................................................................... 109
Indexing 11) Mention the different varieties of keyword indexing.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................

11.6 COMPUTERISED INDEXING

You already came to know that information retrieval deals with the problems related
with the storage, access and searching of information sources by persons in need of
information. In this digital network era, information sources are growing at an exorbitant
rate, available in many forms and formats, and accessible through various channels.
Moreover, recent advancements in Information and Communication Technology (ICT)
help in integration of different information sources and process them on a larger scale.
The results of ICT applications in library activities are related with the development of
wider and efficient information services. Remote database search service (both
bibliographic and full-text databases) is possibly one of the most prominent products of
ICT-enabled library services.
But you will be surprised to know that investigations of computerised indexing date
back to the late 1950s. The earliest and most primitive form of computerised indexing
method relying on the power of computers was invented by Hans peter Luhn, an IBM
engineer, who in 1958 produced what became known as KWIC (Key Word In Context)
indexing. Luhn reported his system in 1960 which has already been discussed under
the sub-section 11.5.1. The system was based just on simple, mechanical manipulations
of terms derived from document titles. Related forms are the Permuterm Subject Index
and the KeyWord Plus known from ISI’s citation indexes (this last system is based on
assigning terms from cited titles). Computers are now increasingly used in aiding indexing,
relying on stored dictionaries of synonyms and homonyms, lists of chemical compounds,
plants and animals, etc. They are also used for automatically arranging entries in
alphabetical order or subordinating subheadings and cross references in exact sequence
under a heading, and performing many other functions that previously had to be done
manually and therefore were quite expensive and often subject to errors. After the
introduction of Unicode (a 2 Byte-oriented encoding standard that can represent all
characters of all scripts of the world) as text encoding standard (in late nineties),
computerised indexing systems are able to store, process and retrieve multilingual
documents available in different scripts.

11.6.1 Meaning and Features

Computerised indexing can be defined as the process whereby a computer is used to
process a natural language text that is already in machine-readable form so that indexing
terms are allocated to its content without direct human intervention.
The features of computerised indexing are:
1) Computerised indexing starts with words.
2) Word association prompts the linking of target words in a search statement.
3) Computers scan text and create ‘inverted file’ which associates words in the file
with position in the texts.
4) Matches words in a search statement against ‘inverted files’ to identify texts that
have words in common.
110
5) Computer algorithms are used to carry out the above operations. Indexing Techniques

6) Humans do the programming and set the parameters for indexing.

7) Computation techniques used include word frequency and keyword analysis.
8) Computerised indexing cannot replace human or manual indexing, rather
complimentary.

11.6.2 Manual Indexing versus Computerised Indexing

Manual Indexing Computerised Indexing
 Human indexer analyses texts  Computer analysis of texts has
and selects terms for indexing. not achieved the reliability of
 Human indexer interprets and human analysis. Computer
encodes text, and makes analysis of texts is carried out
inferences and judgment in by following the human
selecting index terms instructions in the form of a
judiciously. computer programming.
 Semantic, syntactical as well as Computers can select words by
contextual considerations govern employing statistical technique.
the selection of indexing terms  No text interpretation is
 Disagreement among indexers possible as computer cannot
on the determination of the think and draw inferences like
subject of a document as the human indexer. It can select or
process of determination of match keywords which are
subjects may variy from one provided as input text.
indexer to another indexer.  Computer algorithms are drawn
 Selected index terms less in to select, or exclude a term by
numbers. following the rules of semantic,
 It is time consuming. syntactical and contextual
 It is expensive. connotations, like human
 It is very difficult to maintain indexer.
consistency in indexing.  Determination of the subject of
 Relevancy of results are ensured the document is a mechanical
 Generally maintains a balance in process, based on what terms
recall-precision. appear frequently and/or
prominently in the text—i.e.
more frequently a term occurs
in a document, the more likely it
is that the document is about
that term.
 Selected index terms are more
in number.
 It takes less time.
 Index entries can be produced at
lower cost.
 Consistency in indexing is
maintained.
 Relevancy of results are not
always ensured.
 Generally shows high recall and
low precision.

111
Indexing 11.6.3 Advantages and Disadvantages of Computerised
Indexing
Advantages of computerised indexing are as follow:
 It is as effective as human indexing;
 It is cost effective compared to expensive human indexing;
 Maintains consistency in indexing;
 Indexing time is reduced;
 Help searchers find information quickly;
 Can be applied to large volumes of texts where human indexing becomes impossible
(e.g. Indexing web pages);
 Retrieval effectiveness can be achieved.
Disadvantages are
 Not flexible;
 Not precise when looking at unique materials;
 Not able to adapt new terminology;
 Not able to do the conceptual analysis of the content of the document;
 Not a term occurs several times in a document will always be a significant term.

11.6.4 Components of Computerised Indexing System

A typical computerised indexing system has four major components:
1) Database
Database acts as heart of a typical computerised indexing system. Bibliographic
databases which deal with metadata of bibliographic entities (e.g. author, title, subject
etc.) include two parts. The first part is sequential file (a combination of fields —>
records —> database) and inverted file (indexes to sequential file). Full-text databases,
apart from the above two parts also include field-less information entity (source object
in different formats like Web page (HTML), PDF file, Doc file etc).
2) Search Process
Database determines what can be retrieved, whereas search mechanism determines
how information stored in databases can be retrieved. Efficiency of search process is
very important factor for a computerised indexing system. Search process of a typical
computerised indexing system provides two sets of retrieval techniques – basic retrieval
techniques (Boolean operators – AND, OR, NOT; Relational operators - >, <, =, >=,
<=; and Positional search operators – NEAR, ADJ, NEARx etc.) and advance
techniques (Weighted searching, Fuzzy searching etc.).
3) Language of Indexing
Search mechanism determines what retrieval techniques will be available to searchers,
whereas computerised indexing language determines the flexibility in (1) document
representation; and (2) query representation. Computerised indexing language may
112
be grouped as natural language and controlled vocabulary (classification, subject heading Indexing Techniques
and thesauri).
4) User Interface
It is a layer of interaction between users and computerised indexing system. Efficiency
of user interface depends on mode of interaction, display features, online help, provision
of feedback etc. It is considered as the human dimension of computerised indexing
system.

11.6.5 Categories of Computerised Indexing Systems

Since the time of H.P. Luhn (i.e. 1950s) different computerised indexing systems have
been evolved to address information demands of different user groups. These indexing
systems may be categorised as follows –
Category I: Online Database Indexing
These computerised indexing systems allow users to search databases located in remote
places. Here computer technologies are applied to process, store, retrieve records and
communication technologies help in accessing records from centralised databases.
Generally, these IR systems include bibliographical, numeric, full-text and multimedia
information bearing objects. The online IR systems played a major role in the
development of computerised indexing systems over the years.
Advantages: Sophisticated retrieval techniques, cross-database searching, use of
integrated vocabulary control devices and many more.
Limitations: Heavy initial investments, need of intermediary in searching etc.
Category II: Optical Disk based Database Indexing
These computerised indexing systems emerged from online IR systems. These IR systems
offer subset of data (bibliographical, numeric, full-text, and multimedia) through the
optical media (like CDROM, DVDROM etc.). These IR systems adopted almost all
the retrieval techniques of online IR systems but the end users, rather than the intermediary,
does most of the searching in optical disk based IR.
Advantages: Low running cost, end-user friendly, browsing and searching facilities,
sharing of databases on LAN etc.
Limitations: Delayed updating (updating frequency ranges from quarterly to bi-annually)
and restricted access (remote searching is not possible beyond LAN).
Category III: OPAC based Indexing
An OPAC is outer / external form of library catalogue. OPACs are presently considered
as IR systems with their own characteristic features. For example, an OPAC is essentially
local (serve library resources of one or more institutes) but can act as global information
entity by integrating other IR systems e.g. CDROM databases, e-journals access, Z
39.50 based searching of other library catalogues etc.
Advantages: Supports field-level search (i.e. search by author, title etc.), and subject
access points, provides sophisticated retrieval techniques, and facilitates integration of
different information resources into a single search interface.
Limitations: Lack of federated search options, absense of multi-lingual user interface
etc.
113
Indexing Category IV: Indexing Internet Resources
The first three computerised indexing systems are dealing with structured objects (fields
and field values) whereas the last one deals with unstructured information bearing objects
(textual objects without metadata). Internet based computerised indexing systems are
generally based on automatic harvesting devices such as robots, spiders, crawlers etc.
for finding and gathering of information resources available in publicly indexable Internet
(see section 11.7).
Advantages: Free indexing services like search engines, subject directories amd meta
search engines, quick access to huge information sources, supports end-user searching,
simple to retrieve documents.
Limitations: High recall and very low precision, cross-disciplinary semantic drift, relevancy
of results not ensured etc.

11.6.6 Comparison of Computerised Indexing Systems

In the previous section we discussed different categories of computerised indexing
systems by covering features of the major systems including the relative advantages and
disadvantages of those indexing systems. In this section we are going for a comparison
of major computerised indexing systems on the basis of a set of defined parameters.

Features Online Optical Disk based OPAC Internet

Databases Databases
Contents Mainly textual Bibliographic Bibliographic Text, Images
objects; metadata, Fulltext metadata and Multimedia
Bibliographic and Multimdia
metadata
Retrieval Searching Searching and Searching and Searching and
approach browsing browsing browsing
Indexing Metadata and Metadata and Metadata and Limited
Keywords in Keywords in full- Keywords in metadata,
abstract text TOC Keywords in
full-text
Retrieval Basic and Basic and Basic and Basic and
techniques Advanced Advanced Advanced Advanced (full)
(Limited) (Limited) (Full) including Fuzzy
Search Yes Yes No Yes
modification
Search Indexing based End-user search End-user End-user search
method (expert search) search
Output / Traditional Traditional Traditional Different
Display ranking ranking ranking and modern ranking
Limited facilities
modern
ranking
Use of Yes Yes Yes No
controlled
vocabulary
Quality Yes Yes Yes No
control
mechanisms
Multilingual No No Yes Yes
search

114
However, the trend of in the domain of IR system is convergence. CDROM based Indexing Techniques
systems are integrated with OPACs, OPACs are linking online databases, document
delivery services, and other resource discovery services. MARC 21 bibliographic format
includes field 856 for encoding URLs of Internet resources. The Web is becoming the
platform for convergence of different IR systems e.g. Web-OPACs are linking open
databases (information mashup) and acting as the gateways for local and global
information resources to support users.

11.6.7 Index File Organisation

You must have noticed that all four major computerised indexing systems depend on
database as core component. Database, on the other hand depends on two basic
parameters – method for handling sequential file and method for handling inverted file.
Sequential file management ensures organisation of records in a database whereas
inverted file management helps in retrieving records from database against queries.
Therefore, design and development of the computerised indexing system mainly depends
on:
1) Method used for organising records in the file (file organisation), and
2) Method used for searching a record in the file (search techniques).
Further, these two factors are related in a sense that more efficient the file organisation,
more efficient the searching.
In computerised indexing, each document is represented by a record that a computer
can read and manipulate. Typically, these records contain data identifying an item and a
set of index terms to provide subject access. A basic problem in computerised indexing
is storing of files consisting of records, each record having an identical format. A record
format consists of a list of fields, with each field processing a number of characters and
having a fixed data type. A record also consists of values for fields. The complexity of
organising a file for storage depends on the operations we intend to perform on the file.
A file can be categorised as a logical file or a physical file. A logical file is the one
perceived by an application program, it may be different from the one which is stored in
storage units. Logical files are thus abstracts groups of data. Actual data are stored
using physical files. A physical file is a set of data stored on a physical medium such as
disk, tape, etc. It contains a number of data subsets, called physical records, which
have an identical layout.
Some of the important logical file organisations are discussed below:
a) Sequential File: The most common organisation of records in a file is the sequential
file organisation. It is the document file, which contains document records in their
normal form—the form in which they are sequentially entered into the database.
Here, document records are stored one after another in the computer memory—
this is actually the virtual structure of the database file.
b) Inverted File: An inverted file is a computer file in tabular format, in which rows
represent documents and columns represent words. Intersections of rows and
columns are marked when certain documents contain certain words. At the point
of retrieval, the computer scans the entire inverted index for documents which
contain the words in the search query. In an inverted file organisation, two files are
always maintained: sequential file and inverted file. An inverted file contains all the
potential index terms arranged alphabetically, drawn automatically from the
115
Indexing document records according to indexing technique adopted for the purpose. Each
index term in the inverted file is associated with the record number(s) in which the
index term occurs and it links with the list of records that represents documents.
Thus, for each index term in the database the inverted file contains an entry along
with a reference list which specifies position(s) in the database where the term
appears. Each term may occur in a number of documents.
c) Indexed Sequential Files: An indexed sequential file is also an inverted file in
which every record in the source file/document file.
d) Chained Files: It’s a special type of inverted file organisation which supports
dynamic and multiple record linking to ensure simple updating process and quick
response against query.
e) Tree Structured Files: Here records in a file based on the keys can be organized
using a tree structure. If the record is too large, only the keys (in the directory) are
organized in the tree structure.
f) B-Tree: Bayer and McCreight suggested a method of storing keys on a disk by
developing what is called a page. Here a block of storage of fixed size used to
transfer information between main storage and direct access storage. This is called
B-Tree indexing system that allows easy retrieval, insertion, and deletion of records.

11.6.8 Methods of Computerised Indexing

The first step in indexing, both manual and computerised indexing, is to decide on the
subject matter of the document. In manual indexing, the indexer would consider the
subject matter in terms of answer to a set of questions such as “Does the document
deal with a specific product, condition or phenomenon?” Computerised indexing follows
a set of processes of analysing frequencies of word patterns and comparing results to
other documents in order to assign to subject categories. This requires no understanding
of the material being indexed therefore leads to more uniform indexing but this is at the
expense of the true meaning being interpreted. A computer program will not understand
the meaning of statements and may therefore fail to assign some relevant terms or
assign incorrectly. Human indexers focus their attention on certain parts of the document
such as the title, abstract, summary and conclusions, as analysing the full text in depth is
costly and time consuming. A Computerised system takes away the time limit and
allows the entire document to be analysed, but also has the option to be directed to
particular parts of the document.
The second stage of indexing involves the translation of the subject analysis into a set of
index terms. This can involve extracting from the document or assigning from a controlled
vocabulary.
Statistical Method
Statistical method involves taking words directly from the document. It uses natural
language and lends itself well to automated techniques where word frequencies are
calculated and those with a frequency over a pre-determined threshold are used as
index terms. A stop-list containing common words like articles, conjunctions, prepositions
and pronouns are excluded as index terms using a ‘stop word’ file. Automated extraction
of terms may lead to loss of meaning of terms by indexing single words as opposed to
phrases. Although it is possible to extract commonly occurring phrases, it becomes
more difficult if key concepts are inconsistently worded in phrases. The following
statistical methods are adopted in measuring the word significance:
116
a) Term Frequency method: In this method, terms (other than common words) to Indexing Techniques
be indexed are those occurring either very frequently (indicating concepts dealt
with) in a text or very seldom (indicating a topic mentioned expressly only once or
twice in the title or first paragraph but then being referred to by ‘it’ or ‘this’ and the
like). This method is based on the frequency of occurrence and co-occurrence of
terms, using probabilistic model.
b) Relative Frequency Method: Terms that occur infrequently may be highly
significant for example a new drug may be mentioned infrequently but the novelty
of the subject makes any reference significant. One method for allowing rarer
terms to be included and common words to be excluded by automated techniques
would be a relative frequency approach where frequency of a word in a document
is compared to frequency in the database as a whole. Therefore a term that occurs
more often in a document than might be expected based on the rest of the database
could then be used as an index term, and terms that occur equally frequently
throughout will be excluded.
c) Term weighting method: Not all the words have the same significance level.
Words occurring with high frequency in a document are better discriminators than
words of low frequency. Hence, term frequency method (how often a term is use)
is sometimes coupled with term weighting in which different degrees of importance
is assigned to terms on the basis of what terms are used in a search request or on
the basis of where and how terms appear (e.g. in the title, in an abstract, or in the
first and/or last paragraph of a text). For this, term weighting schema is prepared.
This method is based on statistical principle.
Linguistic Method
A quite a different approach to computerised indexing is by syntactic and semantic
analysis of text. The syntactical analysis identifies the grammatical role and relation
among the words in the sentence. It is concerned with automatic recognition of significant
word order in a phrase or sentence and with inflections, prefixes, and suffixes that
indicate grammatical relationships. Semantic analysis approach seeks to analyse noun
phrase automatically with the aid of stored dictionaries and other linguistic aids. Syntactic
and semantic analyses are often used in conjunction. Natural Language Processing
(NLP) is based on these two methods. Linguistic ontologies are used in NLP to assist
in the analysis of natural language text.
Artificial Intelligence (AI) based Indexing System
The application of artificial intelligence in information retrieval research basically involves
processing of source text to identify the roles of words and phrases and the relationships
between them. The results of this processing are used to identify appropriate indexing
expressions. The following indexing system researches based on AI are evident:
a) Natural Language Processing (NLP) based Indexing System: The basic
idea is to process the text of documents to generate indexing terms. Methods
focusing on the lexical level, attempting to identify grammatical classes or parts of
speech of individual words together with machine readable dictionaries to index
documents. At the syntactic level, there have been attempts to examine the
interconnections between words. It uses both co-occurrences of pairs of terms
and threshold distances (number of words between two terms). The ideas of co-
occurrence and threshold distance may be seen as an attempt to address the issue
of coordination (relationships between subject descriptors). This approach is based
117
Indexing upon a network of expertise or knowledge, the elements of which are associated
with the documents represented in the database.
b) Expert System based Indexing System: Expert systems have been used in
information retrieval research in an attempt to replace what is usually referred to
as the expert intermediary. The expertise referred to in this case is that necessary
to construct a search appropriate to the user’s needs and the functioning of the
system. The components therefore include knowledge of how to interrogate the
system, how to identify the appropriate search terms and how to link these terms.
An expert system shell consists of an appropriate conceptual hierarchy and a
dictionary of the appropriate subject area. This is created by analysis of subject
information by experts in the subject area. Associated with each node in the
hierarchy is relationship information for navigating the system and document
references. Searching basically consists of browsing the network. The network
nodes are linked by different kinds of relationship.
Self Check Exercise
Note: i) Write your answers in the space given below.
ii) Check your answers with the answers given at the end of this Unit.
12) Why should we use computerised indexing system in libraries?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
13) What are the major components and categories of computerised indexing system?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
14) Distinguish between Optical Disk based Indexing and Online Database Indexing.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
15) What are the different methods used for organising records in the index file?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
16) Highlight the different methods used in computerised indexing.
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
118
 Indexing Techniques
11.7 INDEXING INTERNET RESOURCES
You already know in sub-section 11.6.5 that Internet is evolving as the largest information
system in the world and Internet based indexing system is now considered as a major
category under computerised indexing system. The global growth in popularity of the
World Wide Web (WWW) has been enabled in part by the availability of browser
based search tools which in turn have led to an increased demand for indexing techniques
and technologies. Information resources accessible through the Web are very much
different from the bibliographic records of a conventional system. The Web presents
information access problems orders of magnitude greater than any encountered before.
The major defect of the Internet as an information source, apart from its sheer size, is
the fact that it lacks any form of quality control. Current chaotic situation caused by the
“every man his own publisher” phenomenon. Publishers of scholarly books and journals
apply reviewing/refereeing procedures that are, at least to some extent, effective in
eliminating the most worthless of what is written. The published indexing and abstracting
services provide the next level of quality filtering, mostly by choosing the journals, report
series, or other publications that they cover on a regular basis. It is obvious that human
professional indexing of the entire Web is completely impractical. Selective professional
indexing, of course, is possible.
There have been several attempts to organise the resources on the WWW. Some of
them have tried to use traditional Library Classification Schemes such as the Library of
Congress Classification, the Dewey Decimal Classification and others. However there
is a need to assign proper subject headings to them and present them in a logical or
hierarchical sequence to cater to the need for browsing.
When we talk about indexing Internet resources, we really mean indexing Web resources
or simply Web indexing. It centres round on the following:
a) search engine indexing of the Web,
b) creation of metadata,
c) organisation of Web links by category, and
d) creation of a Website index that looks and functions like a back-of-book index.

11.7.1 Search Engine Indexing

Internet search engines are special sites on the Web that are designed to help people
find information stored on other sites. A search engine searches a database of information
on the web. It is a tool to help users to locate information available via Internet. There
are differences in the ways various search engines work, but they all perform three
basic tasks:
 They search the Internet — or select pieces of the Internet — based on important
words.
 They keep an index of the words they find, and where they find them.
 They allow users to look for words or combinations of words found in that index.
To find information on the hundreds of millions of Web pages that exist, a search engine
employs special software robots, called spiders (other names for these programs are
crawler, worm, wanderer, gatherer, and so on), which traverse the web, following links
between pages. It builds the list of the words found on Web sites. When a spider is
119
Indexing building its lists, the process is called Web crawling. In order to build and maintain a
useful list of words, a search engine’s spiders have to look at a lot of pages.
How does any spider start its travels over the Web? The usual starting points are lists of
heavily used servers and very popular pages. The spider will begin with a popular site,
indexing the words on its pages and following every link found within the site. In this
way, the spidering system quickly begins to travel, spreading out across the most widely
used portions of the Web.
When the spider looked at an HTML page, it took note of two things:
 The words within the page
 Where the words were found
Words occurring in the title, subtitles, meta tags and other positions of relative importance
were noted for special consideration during a subsequent user search. Meta tags allow
the owner of a page to specify key words and concepts under which the page will be
indexed. The meta tags can guide the search engine in choosing which of the several
possible meanings for these words is correct.
Once the spiders have completed the task of finding information on Web pages, the
search engine must store the information in a way that makes it useful. There are two
key components involved in making the gathered data accessible to users: (1) The
information stored with the data, and (2) The method by which the information is indexed.
In the simplest case, a search engine could just store the word and the URL where it
was found. To make for more useful results, most search engines store more than just
the word and URL.
Search Engine Categories
Search engines may be divided into the following main categories:
 General Search Engine: It covers a range of services and compiles their own
searchable databases on the web. Examples: Google, Alta Vista, etc.
 Regional Search Engine: It refers to country specific search engine for locating
varied resources region-wise. Examples: Euro Ferret (Europe), Excite uk (UK),
etc.
 Meta-search Engine: A meta search engine does not use crawler for compiling
their own searchable database. These search engines utilise databases maintained
by other individual search engines. When a query is put before this type of search
engine, it forwards that query to other search engines. Examples of such search
engines are: Dog pile, Ask Jeeves, Inference Find, MetaCrawler, Profusion,
Surfwax.
 Subject Specific Search Engine: It does not attempt to index the entire Web.
Instead, it focuses on searching for Websites or pages within a defined subject
area, geographical area or type of resource. Examples: Geo index (Geography /
Environmental science), Biochemistry Easy Search Tool (Biochemistry). Because
this specific search engine aims for depth of coverage within a single area, rather
than breadth of coverage across subjects, they are often able to index documents
that are not included even in the largest search engines databases. Some examples
of subject specific search engines are: , Bioweb (Biotechnology), Scirus (Sc. &
Tech.), Medical world search, Health A to Z (Medical Sc.), Math Search
120
 (Mathematical Sc.), Agri Surf (Agricultural Sc.), Law Crawler (Law), KHOJ Indexing Techniques
(India specific search engine), etc.
 Directory-based Search Engines: Due to explosion of information over Internet
it is felt that the results fetched by search engines are often mixed with unwanted
ones. Subject directories, unlike search engines, are created and maintained by
human editors, not electronic spiders or robots. Directory editors typically organise
directories hierarchically into browsable subject categories and sub-categories.
The resources they list are usually annotated. Directories tend to be smaller than
search engine databases, typically indexing only the home page or top level pages
of a site. They may include a search engine for searching their own directory. It
enables the searcher to move from menu to menu, making one selection after
another until he gets to the level where the chosen sites are enlisted. These directories
offer access to the information that has been classified into certain categories.
Examples of such search engines are: Yahoos, Google, Look smart, Magellan,
Open Directory, and Project.
Salient points of differences between Subject Directories and Search Engines are as
furnished below:

Subject Directory Search Engine

Coverage It is a categorised list of A search engine indexes all
websites with brief the information on the
description, based on entire web automatically. It
submission by Web site deals with specific piece of
owners, scrutinised and information, not categories.
edited by professional Hence, contents coverage
editors. Contents coverage in a search engine is higher
in a subject directory are than the search engine.
fewer as compared with a
search engine.
Browse Allows searchers to Does not allow searchers
browse resources in the to browse resources by
home page of a Web site predefined subject
by predefined subject categories. It takes the
categories. From there, searcher to the exact page
searcher can explore the on which the words and
site for more specific phrases he/she is looking
information. for appear.
Searching Searches titles and Searches full-text of web
annotations of categorised pages.
resources.

11.7.2 Subject / Information Gateway

With the advent Internet many libraries are looking forward to go online with Internet.
Often they are finding the information available over Internet is enormous. For the
same they have devised subject-based portals which are known as Subject Gateways.
Major idea of Subject gateways came with the inefficiency of search engines as they fail
to give pin-pointed information. They employ subject experts and information
professionals to select, classify and catalogue Internet resources to aid search and
retrieval for their users.
A subject/information gateway is a web site that provides searchable and browsable 121
Indexing access to online resources focused around a specific subject. Subject gateway resource
descriptions are usually created manually rather than being generated via an automated
process. There are two kinds of gateways: library gateways and portals. Library gateways
are collections of databases and informational sites, arranged by subject that have been
assembled, reviewed and recommended by specialists, usually librarians. These gateway
collections support research and reference needs by identifying and pointing to
recommended, academically-oriented pages on the Web. Some notable examples of
library gateways are Librarians’ Index to the Internet, Internet Public Library, Academic
Information, WWW Virtual Library, Digital Librarian, Infomine, etc.

11.7.3 Semantic Web

Search engines usually employ statistical methods like frequency of occurrence of words,
co-occurrence of words, etc. which retrieve a number of irrelevant hits against a search
on the Web. Though some search engines like Google and Yahoo use human edited
entries; still they come up with a large number of wrong hits. Tim Berners-Lee introduced
the concept of Semantic Web to extend the web with semantic information. The idea
behind Semantic Web is to develop such technologies that make the information more
meaningful for the machines to process, which in turn makes search and retrieval of
information more effective for searchers. The conception of Semantic Web is
characterized by developing tools and technologies like languages, standards and
protocols so that the Web becomes meaningful.
Most of the technologies involved in the development of the Semantic Web are still in
their infancy. Some of them already in use are the URIs (for identifying documents
uniquely and globally), XML (for structuring the data semantically), RDF (to base the
structure of the documents on a common model base), Ontologies (to define the objects/
entities and the interrelations between these objects/entities), etc.

11.7.4 Taxonomies
Taxonomies can be considered another tool useful for organising web-based information.
For example, taxonomies provide an excellent means for organising subject-specific
information into an easily navigable format. The utility of taxonomies for displaying web
information can also be found when examining Yahoo’s site – which uses a taxonomy/
subject hierarchy to classify its indexed information.
Self Check Exercise
Note: i) Write your answers in the space given below.
ii) Check your answers with the answers given at the end of this Unit.
17) What do you mean by indexing Internet resources?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
18) What are the different types search engines used in indexing Internet resources?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
122
19) Discuss the methods of search engine indexing. Indexing Techniques

.....................................................................................................................
.....................................................................................................................
.....................................................................................................................
20) What is Semantic Web?
.....................................................................................................................
.....................................................................................................................
.....................................................................................................................

11.8 SUMMARY
In this Unit we have dealt with different techniques of subject indexing. It begins with a
brief discussion on derivative indexing and assignment indexing and is followed by the
discussion on different types of pre- and post-coordinate indexing systems. We cannot
understand the pre-coordinate indexing properly without being aware of the contributions
of C.A. Cutter and J. Kaiser. For this, principles and processes of subject indexing
techniques as enunciated by Cutter and Kaisers are discussed. Major pre-coordinate
indexing systems like Chain Indexing, PRECIS, and POPSI are discussed with reference
to their principles, syntactical and semantic aspects, entry structure and system of
references. Objective conditions that led to the development of post-coordinate indexing
and its differences with pre-coordinate indexing are explained. Term entry system and
item entry system forming parts of the post-coordinate indexing system are also discussed
with an emphasis on the operational stages of Uniterm indexing system. Different varieties
of keyword indexing are explained. Computerised indexing techniques are explained in
terms of its meaning, features, differences with manual indexing, advantages and
disadvantages, components, categories, index file organisation and different methods
associated with the generation of index entries with the aid of computers. Indexing
internet resources with particular reference to search engine indexing and other associated
concepts are discussed briefly at the end of this Unit. All indexing techniques are
demonstrated with illustrative examples.

11.9 ANSWERS TO SELF CHECK EXERCISES

1) Derived indexing is a method of indexing in which a human indexer or computer
extracts from the title and/or text of a document one or more words or phrases to
represent subject(s) of the work, for use as headings under which entries are
made. It is also known as extractive indexing.
Assigned indexing is also known as ‘Concept indexing’, because what we are
trying to do is to identify concept(s) associated with the content of each document.
It is a method of indexing in which a human indexer selects one or more subject
headings or descriptors from a list of controlled vocabulary, instead of the title
and/or text of a document, to represent the subject(s) of a work.
2) J. Kaiser in his “Systematic Indexing”, published in 1911, solved the problems
indexing compound subjects through classificatory approach. He pointed out that
compound subjects might be analysed by determining the relative significance of
the different component terms of compound subject in terms of two fundamental
categories: (1) Concrete and (2) Process. According to Kaiser, Concrete refers
to Things, place and abstract terms, not signifying any action or process; e.g. gold, 123
Indexing India, Physics, etc. Process refers to (a) Mode of treatment of the subject by the
author; e.g. Cataloguing of films; (b) An action or process described in the
document; e.g. Cultivation of rice; and (c) An adjective related to the concrete
as component of the subject; e.g. Strength of metal. Kaiser laid a rule that a
‘Process’ should follow ‘Concrete’.
3) The concept of ‘chain’ as the foundation of chain indexing is a structural manifestation
of a subject, which refers to the parts constituting a subject and their mutual
interrelationship. It is a modulated sequence of sub-classes or isolate ideas.
Since the chain expressed the modulated sequence more effectively in a notational
classification of subjects, this method takes the class number of the document
concerned as the base for deriving subject headings.
4) The basic steps in chain indexing are (1) Construction of the class number of the
subject of the document; (2) Representation of the Class Number in the form of a
Chain; (3) Determination of links like Sought Links (SL), Unsought Links (USL),
False Links (FL), and Missing Links (ML); (4) Preparation of Specific Subject
Heading; (5) Preparation of Subject Reference Headings; (6) Preparation of
Subject Reference Entries; (7) Preparation of Cross References, if any; and (8)
Alphabetisation.
5) Principle of context dependency in PRECIS is seen as a combination of context
and dependency. When this principle is followed in a PRECIS input string, each
term is qualified and sets the next term into its wider context. In other words, the
meaning of each term in the string depends upon the meaning of its preceding term
and taken together, they all represent the single context. Each term is hence
dependent, directly or indirectly, on all the terms which precede it.
6) Two-line-three part entry structure is followed in PRECIS. The first line is occupied
by two parts—Lead and Qualifier, which together constitute the Heading. Lead is
occupied by the approach term and Qualifier position is occupied by the term(s)
that sets the Lead into wider context. The second line is occupied by the third
part, i.e. Display position which consists of those additional set of qualifying terms,
which rely upon the heading for their context.
Index entries in PRECIS are basically generated in three formats: (a) Standard
format— entries are generated when any of the primary operators (0), (1), and
(2) or its dependent elements appear in the Lead; (b) Predicate transformation—
entries are generated under a term coded (3) that immediately follows a term
coded either by (2) or (s) or (t); and (c) Inverted format— entries are generated
whenever a term coded by an operator in the range from (4) to (6) or its dependent
elements appear in the Lead.
7) Different operational stages of POPSI may be categorised under three components:
Analysis, Synthesis, and Permutation. The work of ‘Analysis’ and ‘Synthesis’ is
primarily based on the postulates associated with the deep structure of SILs for
generating organising classification. The task of analysis and synthesis is largely
guided by the following POPSI-table. The work of ‘Permutation’ is based on
cyclic permutation of each term-of-approach, either individually or in association
with other terms for generating associative classification effect in alphabetical
arrangement.
8) The major steps in formulating index entries according to POPSI are (1) Content
Analysis, (2) Formalisation, (3) Standardisation, (4) Modulation, (5) Preparation
of Entry for Organising Classification, (6) Approach-term Selection, (7) Preparation
124 of Entries of Associative Classification, and (8) Alphabetisation.
9) In pre-coordinate indexing system, coordination of component terms is carried Indexing Techniques
out before the users come to search the index file, i.e. at the time of indexing by the
indexer (i.e. at input stage) in anticipation of the users’ approach. But in post-
coordinate indexing, component terms of a subject are kept separately
uncoordinated by the indexer, and the user does the coordination of concepts at
the time of searching (i.e. at the output stage).
10) Search devices used to avoid false coordination of terms in Post coordinate indexing
are: (a) Use of Pre-coordinated Terms, (b) Links, (c) Roles, and (d) Weighting.
11) Different varieties of keyword indexing include Key Word in Context (KWIC)
Indexing, Key Word Out of Context (KWOC) Indexing, Key-Word Augmented-
in-Context (KWAC) Indexing, Double KWIC Indexing, KWWC (Key-Word-
With-Context) Indexing, KEYTALPHA (Key-Term Alphabetical) Indexing, KLIC
(Key-Letter-In-Context) Indexing, and WADEX (Word and Author Index).
12) Library is a collection of databases. For example, Accession register is a database,
Card catalogue is a database, Circulation records constitute a database, Member
register is a database. We need to add, edit and maintain all these databases
regularly. The advantages of computerised indexing system are as follows:
Redundancy can be reduced; Inconsistency can be avoided; Data can be shared;
Standards can be enforced; Security restrictions can be applied; Integrity can be
maintained and conflicting requirement may be solved.
13) A typical computerised indexing system has four major components: 1) Database,
2) Search process, 3) Language of indexing, and 4) User interface.
Computerised indexing systems can be categorised into following four groups on
the basis of the varieties of information demands of different user groups: 1) Online
Database Indexing, 2) Optical Disk based Database Indexing, 3) OPAC based
Indexing, and 4) Indexing Internet Resources.
14) A comparative study of online database indexing and Optical Disk based indexing
may be done as follows:

Optical Disk based Indexing Online Database Indexing

1) Locally searchable. No 1) Remotely searchable.

requirement for Communication network is
communication channel. a must.
2) Cost based procurement. 2) Renewal based
3) End user searching. procurement.

4) Offline access. Generally 3) Intermediary searching.

updated on monthly 4) Online access. Updating is
basis. daily and sometimes hourly.
5) Contains multimedia 5) Bibliographic databases
database to reference those are large and
databases. frequently updated.
6) Suitable for 6) Suitable for large number of
homogeneous user simultaneous users.
environment.

125
Indexing 15) Design and development of the computerised indexing system mainly depends on
the methods used for organising records in the file. Some of the important logical
file organisations are: (1) Sequential File, (2) Inverted File, (3) Indexed Sequential
Files, (4) Chained Files, (5) Tree Structured Files, and (6) B-Tree.
16) Different methods of computerised indexing include: (1) Statistical Method, which
may consist of (1.1) Term Frequency method, (1.2) Relative Frequency Method,
and (1.3) Term weighting method; (2) Linguistic Method, (3) Artificial Intelligence
(AI) based Indexing System, which may consist of (2.1) Natural Language
Processing (NLP) based Indexing System, and (2.2) Expert System based Indexing
System,
17) Indexing Internet resources or Web indexing means the following:
a) search engine indexing of the Web,
b) creation of metadata,
c) organisation of Web links by category, and
d) creation of a Website index that looks and functions like a back-of-book
index.
18) Different types search engines used in indexing Internet resources are: (1) General
Search Engine, (2) Regional Search Engine, (3) Meta-search Engine, (4) Subject
Specific Search Engine, and (5) Directory-based Search Engines.
19) In order to find information from the Internet, a search engine employs special
software robots, called spiders (also called crawler, worm, wanderer, gatherer,
etc.), which traverse the web, following links between pages. It builds the list of
the words found on Web sites, called Web crawling. Here, the spider begins with
a popular site, indexing the words on its pages and following every link found
within the site. In this way, the spidering system quickly begins to travel across the
Web and takes note of the (a) words within the page, and (b) location of the
words - title, subtitles, meta tags and other positions of the Web page. Meta tags
allow the owner of a page to specify key words and concepts under which the
page will be indexed. The meta tags also guide the search engine in choosing
which of the several possible meanings for these words is correct. After completion
of finding information on Web pages by the spiders, the search engine stores the
information in a way that makes accessible to users. In the simplest case, a search
engine just stores the word and the URL where it was found. To make for more
useful results, most search engines store more than just the word and URL.
20) The “semantic web” is an approach to extend the web with semantic information
to avoid wrong hits. The conception of Semantic Web is characterised by developing
tools and technologies like languages, standards and protocols so that the Web
becomes meaningful. Technologies involved in the development of the Semantic
Web are the Uniform Resource Identifier (URI) for identifying documents uniquely
and globally, XML (eXtensible Markup Language) for structuring the data
semantically, RDF (Resource Description Framework) to base the structure of
the documents on a common model base, Ontologies (to define the objects/entities
and the interrelations between these objects/entities), etc.

126
 Indexing Techniques
11.10 KEYWORDS
Action : An elementary category associated with POPSI
which refers to an idea denoting the concept of
‘doing’. An action may manifest itself as Self
Action or External Action.
Assigned Indexing : The process of indexing in which a human indexer
selects one or more subject headings or
descriptors from a list of controlled vocabulary
to represent the subject(s) of a work. Also
known as Assignment Indexing and Concept
Indexing.
Associative Classification : It refers to a classification in which a subject is
distinguished from all other subjects based on
the reference of how it is associated with other
subjects, without reference to its COSSCO
relationships. The result of associative
classification is always a relative index.
Back-of-the-book Index : An index which shows where exactly in the text
of a document a particular concept (denoted by
a term) is mentioned, referred to, defined or
discussed.
Base : It is a particular manifestation or manifestations
of a particular elementary category under which
all or major portion of related information are
brought together.
Boolean Operators : AND, OR, and NOT. Used to combine search
terms. AND finds only records that contain both
terms. OR finds records that contain either term.
NOT finds records that contain the first term but
not the second term.
Chain Indexing : The process of deriving subject index entries
based on the extracted vocabulary of a notational
scheme of classification. It retains all necessary
context but removes unnecessary context.
Classaurus : It is an elementary category-based (faceted)
systematic scheme of hierarchical classification
in verbal plane incorporating all the necessary
features of a conventional information retrieval
thesaurus. It is used as vocabulary control device
in POPSI.
Coextensive Subject Index : A subject index entry, in which a term, phrase,
or a set of terms define precisely the full thought
content of the document. Here extension and
intension of the ideas are equal to the thought
content of the document.
127
Indexing Common Modifier : It refers to the name of a place (space), Time,
Environment), and Form.
Computerised Indexing : A method of indexing in which an algorithm is
applied by a computer to the title and/or text of
a work to identify and extract words and phrases
representing subjects, for use as headings under
which entries are made in the index.
Concrete : An elementary category suggested by Kaiser to
refer to things, place and abstract terms, not
signifying any action or process.
Content Designation : The act of making a bibliographic record machine
readable by encoding its various elements
according to a specified scheme.
Core : It is a particular manifestation or manifestations
of one or more elementary category under which
all or major portion of related information are
brought together within a recognised Base.
COSSCO Relationship : It is a relationship in which COordinate—
Superordinate—Subordinate—COllateral
(COSSCO) relationships of a subject are shown.
Deep Structure of Subject : DS-SIL refers to the logical abstraction of the
Indexing Languages (DS-SIL) surface structures of outstanding SILs like Cutter,
Dewey, Kaiser and Ranganathan.
Derived Indexing : The process of indexing in which terms to be
used to represent the content of the document
are derived directly from the document itself. Also
known as Derivative Indexing.
Discipline : An elementary category associated with POPSI
that includes the conventional fields of study, or
any aggregate of such fields, or artificially created
fields.
Entity : An elementary category associated with POPSI
which includes manifestations having perceptual
correlates, or only conceptual existence, as
contrasted with their properties, and actions
performed by them or on them.
False Drops : Retrieval of unwanted documents because of the
false coordination of terms at the time of
searching.
Input String : A set of terms arranged according to the role
operators which act as instructions to the
computer for generating index entries.
Item Entry System : A type of post-coordinate indexing system in
which It takes the opposite approach to term
128 entry system and prepares a single entry for each
 document (item), using a defined physical form, Indexing Techniques
which permits access to the entry from all
appropriate headings. Here, items are posted on
the term.
Keyword : A term that is chosen, either from the actual text
or from the queries of the searcher, that is
considered to be a ‘key’ to finding certain
information.
Keyword Indexing : The process of using significant words from a
title or an abstract or sometimes from the text of
the document as index entries.
KWIC Indexing : Key Word In Context, format for showing index
entries within the context in which they occur.
KWOC Indexing : Key Word Out of Context, the use of significant
word from titles for subject index entries, each
followed by the whole title from which the word
was taken.
Meta Search Engine : a program that allows to search across many
search engines at once.
Modifier : It refers to a qualifier used to modify any one the
elementary categories D, E, A and P associated
with POPSI.
Nesting : Grouping terms within parentheses to specify the
order in which they will be combined. Terms in
the innermost parentheses will be combined and
searched first. Without parentheses, terms will
be combined in left-to-right order.
Ontology : A formal specification of a representational
vocabulary for a shared domain of discourse—
definitions of classes, relations, functions, and
other objects. Ontologies define data models in
terms of classes, subclasses and properties to
enhance the functioning of the Web.
Organising Classification : In organising classification compound subjects
are based on genus-species, whole-part, and
other inter-facet relationships. Organising
classification distinguish and rank each subject
from all other subjects with reference to its
COordinate—Superordinate—Subordinate—
COllateral (COSSCO) relationships.
Post-Coordinate Indexing : An indexing model in which terms associated with
the content of the document are kept separately
in the index file by the indexer and the searcher
coordinate coordinates the terms at the time of
searching or output stage. Also known as
‘coordinate indexing’.
129
Indexing PRECIS : PREserved Context Index System, a subject
indexing technique in which an open-ended
vocabulary can be organised according to a
scheme of role operators, usually for computer
manipulation.
Pre-Coordinate Indexing : An indexing model in which terms associated with
the content of the document are coordinated by
the indexer by following the syntactical rules of
given indexing language at the time of indexing
or input stage for use in the retrieval of information
collection on compound and /or complex
concepts.
Process : ‘An elementary category suggested by Kaiser
to refer to mode of treatment of the subject by
the author, an action or process described in the
document, and an adjective related to the
concrete as component of the subject.
Property : An elementary category associated with POPSI
which refers to the idea denoting ‘attribute’.
Role Operator : Role operators consist of a set of alpha-numeric
notations which specifies the grammatical role or
the function of the indexed term and regulates
the order of terms in the input string. Role
operators and their associated rules also serve
as the computer instruction for determining the
format, typography and punctuation associated
with each index entry.
Search Engine : A retrieval tool on the World Wide Web that, in
general, matches keywords input by a user to
words found at websites. The more sophisticated
search engines may allow other than keyword
searching.
Semantic Web : The “semantic web” is an approach to extend
the web with semantic information to avoid wrong
hits by developing tools and technologies like
languages, standards and protocols so that the
Web becomes meaningful.
Special Modifiers : A special modifier refers to a qualifier which is
used to qualify/modify only one of the elementary
categories associated with POPSI.
Stop-word List : The ‘stop-word’ list refers to a list of words,
which have no value for indexing / retrieval. These
may include insignificant words like articles (a,
an, the), prepositions, conjunctions, pronouns,
auxiliary verbs together with such general words
as ‘aspect’, ‘different’, ‘very’, etc. Each major
search system has defined its own ‘stop list’
Subject Analysis : The process of identifying the different component
130
 ideas associated with the thought content of the Indexing Techniques
document and establishing the interrelationships
among those component ideas.
Subject Gateways : Organized lists of web pages, divided into subject
areas by human indexers.
Subject Heading : A word or group of words representing the
subject of a document.
Subject Index : A tool that exhibits the analysed contents of the
collection of documents (either in the library or
database).
Subject Indexing : The process of representing the informational
content of the document by analysing its content
and translating the result of analysis into an
indexing language for creating a surrogate record
for it, especially subject access points, in an index.
Term Entry System : A type of post coordinate indexing system in
which index entries for a document are made
under each of the component terms associated
with the thought content of the document. Here,
terms are posted on the item.
Web Indexing : Web indexing means providing access points for
online information materials, which are available
through the use of World Wide Web browsing
Software.
World Wide Web (WWW) : a network of many thousands of servers linked
together by a common protocol.

11.11 REFERENCES AND FURTHER READING

Austin, Derek. PRECIS: A Manual of Concept Analysis and Subject Indexing. 2nd
ed. London: British Library Bibliographic Services Division, 1984. Print.
Chakraborty, A. R. and Bhubaneswar Chakraborty. Indexing: Principles, Processes
and Products. Calcutta: World Press, 1984. Print.
Chowdhury, G. G. Introduction to Modern Information Retrieval. 2nd ed. London:
Facet Publishing, 2004. Print.
Dykstra, Mary. PRECIS: A Primer. Rev. Reprint. Metuchen, NJ & London: Scarecrow
Press, 1987. Print.
Foskett, A. C. Subject Approach to Information. 5th Ed. London: The Library
Association, 1996. Print.
Ghosh, S. B. and J. N Satpathi., Eds. Subject Indexing Systems: Concepts, Methods
and Techniques. Calcutta. IASLIC, 1998. Print.
Guha, B. Documentation and Information: Services, Techniques and Systems. 2nd
rev ed. Calcutta: World Press, 1983. Print.
131
Indexing International Standards Organization. Documentation—Methods for Examining
Documents, Determining their Subjects, and Selecting Indexing Terms, 1985. Print.
Jones, Karen Sparck, Ed. Information Retrieval Experiment. London: Butterworth,
1981. Print.
Lancaster, F. W. “Do Indexing and Abstracting have a Future”? Anales de
Documentation, no 6, 2003, 137-144. Print.
Olson, Hope A. Subject Analysis in Online Catalogs. 2nd edition. Englewood: Libraries
Unlimited, 2001. Print.
Sarkhel, Juran Krishna. Information Analysis in Theory and Practice. Kolkata:
Classique Books, 2001. Print.
Sarkhel, Juran Krishna. “Subject Indexing, Vocabulary Control and Recent
Developments in Cataloguing”. Library Cataloguing Theory. (BLIS 04)New Delhi,
Indira Gandhi National Open University, 1999. 180p. Print.
Taylor, Arlene G. Introduction to Cataloguing and Classification. 10th edition.
New Delhi: Atlantic Publishers, 2007. Print.
Wellisch, H. H. Indexing from A to Z. 2nd. ed. New York: H. W. Wilson, 1995.
Print.
Wilson, T. D. An Introduction to Chain Indexing. Hamden, Conn.: Linnet Books,
1971.  (Programmed texts in library and information science).  Print.

132