1. June 2003INIS Training Seminar1 INIS Training Seminar 2-6 June 2003 Subject Analysis Thesaurus and Indexing Alexander Nevyjel Subject Control Unit INIS Section, IAEA

2. June 2003INIS Training Seminar2 1. Introduction to Subject Analysis u Subject Analysis should be carried out whenever possible by subject specialists with a good knowledge of the subject matter and a familiarity with the subject analysis tools of the respective database (subject categories, thesaurus, subject analysis rules) u Steps of Subject Analysis  subject classification  abstracting  subject indexing

3. June 2003INIS Training Seminar3 2. Preparatory Analysis Input preparation requires close collaboration between descriptive cataloguer and subject specialist in: u selecting literature relevant to the database scope and checking whether the information content makes its entry into the database worthwhile u deciding whether subdivisions should be made at the bibliographic level and each part treated as a separate piece of literature u selecting the appropriate literary indicator of the record

4. June 2003INIS Training Seminar4 2.1 Selection of relevant Literature u All the literature in the scope of the database should be included u Within the scope the database includes all relevant published material Rule: Select only those pieces of literature to which you can assign at least one appropriate subject category u In addition, it should be determined whether the information value of the piece of literature makes its entry into the database worthwhile. Non-scientific-technical publications, such as newspapers, newsletters, etc. should be excluded

5. June 2003INIS Training Seminar5 2.2 Bibliographic Subdivision Publications may consist of subunits such as chapters of a book or an annual report, or individual conference papers in conference proceedings. For these publications the subject specialist must decide whether bibliographic subdivisions should be made and whether all or only some of the items should be analysed separately A separate record for the entire publication must be supplied as a „lead“ record in addition to the „analytic“ records

6. June 2003INIS Training Seminar6 3. Subject Classification u The main topic of the document determines the primary subject category u If there are other significant topics, one or more secondary subject categories can be assigned in addition

7. June 2003INIS Training Seminar7 4. Abstracting (1) u Each input item should contain an English abstract (exception: short communications) u Abstracts in other languages are optional u If an author abstract is available, it should be checked by the subject specialist, and edited, if necessary u Abstracts can be: s informative s indicative u Abstracts should be informative whenever possible

8. June 2003INIS Training Seminar8 4. Abstracting (2) u Submit an informative abstract whenever possible u Emphasize what is novel about the information in the original document u Do not repeat the title of the original document in the body of the abstract u Do not exceed 6000 characters in length, including spaces and symbols u At the end of the abstract state the number of references, figures, tables and other illustrative material u The abstract may be followed by parentheses containing the name or initials of the person who wrote the abstract

9. June 2003INIS Training Seminar9 5. Subject Indexing 5.1 Co-ordinate Indexing u Co-ordinate indexing is the concurrent assignment of a set of descriptors to indicate the subject content of a piece of documents u The totality of the descriptors representing the information contained in the piece of literature is its set of „co-ordinated“ descriptors u In co-ordinate indexing any combination of descriptors is theoretically possible

10. June 2003INIS Training Seminar10 5.2 The Thesaurus and its Structure (1) What is a Thesaurus ? „A thesaurus is a terminological control device used in translating from the natural language of documents, indexers or users into a more constrained system language. It is a controlled and dynamic vocabulary of semantically and generically related terms which covers a specific domain of knowledge“ This definition has been adopted by UNESCO „Guidelines for the establishment and development of monolingual thesauri“, UNESCO, SC/W/255, Paris, September 1973

11. June 2003INIS Training Seminar11 5.2 The Thesaurus and its Structure (2) u A thesaurus is a tool in subject indexing for information retrieval u Descriptors („controlled terms“) chosen from the thesaurus and assigned to a document should clearely indicate the information content of the piece of literature u The meaning of descriptors must be well defined and unambiguous u Synonyms of controlled terms may be included in the thesaurus as „forbidden terms“ with a cross-reference

12. June 2003INIS Training Seminar12 5.2 The Thesaurus and its Structure (3) u Preferential indicators identify the preferred choice in cases of semantic ambiguity s USE; USED FOR; SEE; SEEN FOR u Hierarchical indicators identify the semantic relationship existing between descriptors on different specificity levels in the same hierarchy of concepts s BT broader terms; NT narrower terms u Affinitive indicators identify the semantic relationship existing between descriptors which are usually on the same level but do not belong to the same hierarchy of concepts s RT related terms

13. June 2003INIS Training Seminar13 5.2 The Thesaurus and its Structure (4) u The structured alphabetical list of the Thesaurus contains s accepted terms --> descriptors s forbidden terms--> non-descriptors u For each alphabetical entry in the Thesaurus a „wordblock“ is given containing all terms associated with that particular entry u The date of introduction of the descriptor is given with the wordblock

14. June 2003INIS Training Seminar14 5.2 The Thesaurus and its Structure (5) RelationshipSymbCross reference hierarchicalBTbroader term (level 1, 2, etc.) hierarchicalNTnarrower term (level 1, 2, etc.) affinitiveRTrelated term preferentialUFused for (reciprocally USE...) preferentialUF+used for plus (reciprocally USE... AND...) preferentialSFseen for (reciprocally SEE... OR...)

15. June 2003INIS Training Seminar15 5.2 The Thesaurus and its Structure (6) Terms in natural language sometimes have several meanings, but each descriptor should be univocal. For those descriptors where ambiguity could exist, the semantic definition is provided by the Thesaurus in the following ways: u The descriptor is placed in its correct semantic context by the association with its forbidden, broader, narrower and related terms u A scope note provides the exact meaning which is intended

16. June 2003INIS Training Seminar16 5.2 The Thesaurus and its Structure (7) Examples: SOLUTIONS (For mathematical solutions see ANALYTICAL SOLUTION or NUMERICAL SOLUTION) BT1 homogenous mixtures BT2 mixtures BT3 dispersions NT1 aqueous solutions NT1 hypertonic solutions NT1 isotonic solutions

17. June 2003INIS Training Seminar17 5.3 General Indexing Rules and Procedures Subject indexing means analysing the information content of a piece of literature and expressing the meaningfull information content in the language of the database using the controlled vocabulary of the Thesaurus u Understanding of the content --> subject specialist u Familiarity with Thesaurus and indexing rules u Select a set of descriptors that describes the subject content of the piece of literature

18. June 2003INIS Training Seminar18 5.3.1 The Document Interpretation Rule In retrieval the user is searching for documents dealing with a topic defined by one, two or three important ideas. In preparing the search query, the retriever represents each of the main ideas by one or more descriptors. The indexer identifies for each piece of literature the one, two or three main ideas of the document. Each of these ideas must be represented by means of one or more descriptors. Rule: Choose such information items for indexing as you would yourself expect to find in the document if you were the user searching for that information.

19. June 2003INIS Training Seminar19 5.3.2 The Specificity Rule (1) The „up-posting“ procedure: For every descriptor assigned to a piece of literature all of its associated broader terms are automatically assigned as well Rule: Always use the most specific appropriate descriptor Examples: if appropriate usenot the more general INELASIC SCATTERINGSCATTERING HEAVY ION ACCELERATORSACCELERATORS COUNTING CIRCUITSELECTRONIC CIRCUITS MANGANESE BASE ALLOYSMANGANESE ALLOYS

20. June 2003INIS Training Seminar20 5.3.2 The Specificity Rule (2) Rule: Do not assign a descriptor and one of ist broader terms to the same item Exceptions: u If the document treats the same hierarchy of concepts in detail but at different levels of specificy u If the document treats several specific concepts, all of them belong to the same class but for only some of which specific descriptors exist

21. June 2003INIS Training Seminar21 5.3.3 Procedures for Indexing (1) u Carefully read the title and abstract and scan the body of the piece of literature u scan the full text (i ntroduction, table of content, tables, graphs, figures, conclusion) to find information items missing from the abstract or requiring more precision u Identify the concept(s) about which the piece of literature contains useful information u Translate the concepts into descriptors u something that was not examined should not be indexed

22. June 2003INIS Training Seminar22 5.3.3 Procedures for Indexing (2) u If part of the document is within the database scope and part is not, index the latter portion generally u Avoid overindexing. Do not add a descriptor merely because it appears as a term in the title or abstract u Check each descriptor to make sure that s the descriptor represents as precisely as possible the concept s the definition matches the use s the selected descriptor is the most specific appropriate choice

23. June 2003INIS Training Seminar23 5.3.4 The Proposal Rule Rule: If no suitable descriptor exists in the Thesaurus for the retrieval of a usefull concept, propose a new one Proposals for new descriptors are carefully evaluated by the Subject Controll Unit of INIS. The indexer is required to provide with his proposal: u a brief definition of the proposed term or description of its intended meaning u suggestions for its associated terms, usually its broader and related terms

24. June 2003INIS Training Seminar24 The purpose of subject indexing is to enable useful retrieval