{ 213}


The purpose of the following brief manual is to provide some guidance to indexers wishing to produce original NEPHIS input strings similar to those used for the five-year cumulative indexes to the Canadian Journal of Information Science. Production of original NEPHIS input strings includes not only coding, but also choice of terms, connectives, and citation order. In contrast to this manual, the NEPHIS implementation title coding manual (Craven and Fjerestad 1981) provides guidance only on the addition of NEPHIS codes and backward-reading connectives to existing descriptions.

In addition to the basic rules of NEPHIS indexing, certain policy decisions were made for the production of the CJIS indexes. A number of possible NEPHIS indexing techniques were excluded by the decisions and are therefore not mentioned in this manual. Notable among the exclusions are use of the comma as a backward-reading connective and the breaking up of long phrases by means of additional brackets. Index producers may modify the rules given here to allow specific techniques which they find useful.

The manual is divided into two parts. The first part gives a brief recapitulation of the general features of NEPHIS. The second contains the rules specific to the CJIS type of index. Both parts assume knowledge of certain basic terms in string indexing, such as "input string", "index string", and "term", which are defined in the Glossary. { 214}


Input string coding

  1. Angular brackets ("<", ">") mark off nested phrases within the input string.
  2. A question mark ("?") marks the beginning of a forward- or backward-reading connective. The left angular bracket at the beginning of a nested phrase marks the end of a forward-reading connective; the right angular bracket at the end of a nested phrase marks the end of a backward-reading connective.
  3. An "at" sign ("@") at the beginning of the input string or after the left angular bracket at the beginning of a nested phrase suppresses generation of the corresponding index string. If the input string or a nested phrase is not marked by an "at" sign, it is an access phrase.

Sample input string

@Use? of <Numeric <Data>>? in <Humanities? & <Social Sciences? & >>
Resulting index strings
  1. Data. Numeric -. Use in Humanities & Social Sciences
  2. Humanities & Social Sciences. Use of Numeric Data
  3. Numeric Data. Use in Humanities & Social Sciences
  4. Social Sciences & Humanities. Use of Numeric Data

Index string generation

  1. An index string is generated for each access phrase. For example, for the access phrases "Numeric <Data>", "Data", "Humanities? & <Social Sciences? & >", and "Social Sciences? & " in the sample input string, the index strings generated are respectively numbers 3, 1, 2, and 4 above.
  2. An index string begins with an access phrase minus its coding characters and minus the backward-reading connectives of any nested phrases within it. Note how the index strings above begin with "Numeric Data" (3), "Data" (1), "Humanities & Social Sciences" (2), and "Social Sciences & " (4).
  3. A forward-reading connective is omitted from an index string when, in the input string, it immediately precedes the access phrase or a nested { 215} phrase containing the access phrase. Note the omission of " of " from index strings 1 and 3, of " in " from index strings 2 and 4, and of the " & " after "Humanities" from index string 4.
  4. If a forward-reading connective has not been specified, a dash is inserted where the forward-reading connective would otherwise be omitted. Note the dash in index string 1.
  5. A backward-reading connective is included in an index string when it terminates the access phrase or a nested phrase containing the access phrase. Note the inclusion of the " & " following "Social Sciences" in index string 4.
  6. If a backward-reading connective has not been specified, a period-plus-space (". ") is used instead.



  1. Except as otherwise provided, use only nouns as terms.
  2. Except as otherwise provided, do not use words adjectivally; e.g.,
    NOT   Canadian <Government <Information>>
    BUT   Information? of <Government? of <Canada>>
  3. Use a word adjectivally when required as part of a multiword term; e.g.,
    Computerized Conferencing
    Information Systems
  4. Use a word adjectivally as an adjectival term when no alternative using nonadjectival terms would maintain the good qualities of all the index strings; e.g.,
    NOT   Libraries? for <Public>
    BUT   Public <Libraries>
    NOT   Information? for <Viewing>
    BUT   Visual <Information>
  5. Use only terms that describe the subject matter of the indexed item, not other aspects such as its form, author, or title; e.g., { 216}
    NOT   @Review? of <Publications of the Government of Ontario 1867-1900? /by <@Olga <Bishop>>>
    BUT   Publications of the Government of Ontario 1867-1900? /by <@Olga <Bishop>>>
  6. Except as otherwise provided, use only terms which contribute substantially to the description; e.g.,
    NOT   @Eye-opening Qualities? of <@Personal Experiences? with <EIES>>
    BUT   @Personal Experiences? with <EIES>
  7. Use terms such as "Use", "Applicability", "Role", and "Effects" to express certain types of relationships if indicating these relationships by simple prepositions is impossible or would seriously degrade one or more of the index strings; e.g.,
    NOT   Reference Work? with <Computerized <Bibliographic Systems>>
    BUT   @Use? of <Computerized <Bibliographic Systems>>? in <Reference Work>
    NOT   @Concept? of <Anomalous States of Knowledge>? in <Retrieval? of <Data>>
    BUT   @Applicability? of <@Concept? of <Anomalous States of Knowledge>>? to <Retrieval? of <Data>>
  8. Treat a surname preceded by intials or forename as a separate term; e.g.,
    NOT   @Model? of <Communication>? of <CC Shannon>
    BUT   @Model? of <Communication>? of <@CC <Shannon>>
  9. Qualify the name of a department or service by terms identifying the superordinate body if ambiguity might otherwise arise; e.g.,
    NOT   Citizen's Inquiry Branch
    BUT   Citizen's Inquiry Branch? of <Government? of <Ontario>>
    BUT   Agriculture Canada
    NOT   Agriculture Canada? of <Government? of <Canada>>
  10. Qualify titles of works by the names of the personal authors if any; e.g., { 217}
    NOT   Publications of the Government of Ontario 1867-1900
    BUT   Publications of the Government of Ontario 1867-1900? / by <@Olga <Bishop>>
  11. Use an initial capital letter on every noun or adjective in a term; e.g.,
    NOT   Mathematical model
    BUT   Mathematical Model
  12. Begin a term with a capital letter, not with a lower-case letter, numeral, punctuation mark, or space; e.g.,
    NOT   @Derivation? of < Bradford Distribution>
    BUT   @Derivation? of <Bradford Distribution>
  13. Use only letters, numerals, spaces, and hyphens in a term; e.g.,
    NOT   National Union Catalog: Pre-1956 Imprints
    BUT   National Union Catalog Pre-1956 Imprints
  14. Make the last character of a nonadjectival term a letter or a numeral, not a hyphen or a space; e.g.,
    NOT   @Concepts ? of <Information Science >
    BUT   @Concepts? of <Information Science>
  15. Make the last character of an adjectival term a space; e.g.,
    NOT   Public<Libraries>
    BUT   Public <Libraries>


  1. Except as otherwise provided, use only prepositions as connectives.
  2. Include a participle or an adjective before a preposition as a connective where use of nouns and prepositions alone would seriously degrade the good qualities of an index string; e.g.,
    NOT   @Research? in <Typography>? among <@Needs? for <Teletext>>
    BUT   @Research? in <Typography> needed? for <Teletext>
    { 218}
  3. When two or more parts of a description are coordinate, connect each subsequent part to the first with the forward- and backward-reading connective " & "; e.g.,
    NOT   @Relationships? between <Economic Systems and <Media>>
    BUT   @Relationships? between <Economic Systems? & <Media? & >>
  4. Use the connective " / by " to link a personal author name with the name of a specific work; e.g.,
    NOT   Knowledge Park? by <@Stephen <Franklin>>
    BUT   Knowledge Park? / by <@Stephen <Franklin>>
  5. Begin and end a connective, or sequence of connectives, with a space; e.g.,
    NOT   Computerized <MIS>?in<Libraries>
    BUT   Computerized <MIS>? in <Libraries>

Citation order

  1. Cite after a term all the nonadjectival terms that qualify it; i.e., that limit the class of things to which it could refer; e.g.,
    NOT   Legislation?. <Indexing>
    BUT   Indexing? of <Legislation>
    Except for coordinate parts, or where otherwise unavoidable, do not allow a nonqualifying term to intervene between qualifying terms and the term qualified.
  2. Cite adjectival terms before the terms that they qualify; e.g.,
    NOT   Networks?, <Bibliographic>
    BUT   Bibliographic <Networks>
  3. When qualifying a term directly by two or more terms, cite the qualifying terms according to the following precedence rules for the connectives that precede them:
    1. of
    2. at, for, from, in, on, to
    3. by
    e.g., { 219}
    NOT   Access? by <Public>? to <Information>
    BUT   Access? to <Information>? by <Public>
    NOT   Matrix? of <@Effects? on <Societies>? of <Media>>
    BUT   Matrix? of <@Effects? of <Media>? on <Societies>>
  4. When two or more parts of a description are coordinate, cite the parts in alphabetical order; e.g.,
    NOT   Communication? between <Human Beings? & <Computerized Systems? & >>
    BUT   Communication? between <Computerized Systems? & <Human Beings? & >>
  5. Cite the parts of a personal name in their ordinary-language order; e.g.,
    NOT   Knowledge Park? / by <Franklin, Stephen>
    BUT   Knowledge Park? / by <@Stephen <Franklin>>

Nesting ("<", ">")

  1. Except as otherwise provided, mark off a nested phrase beginning with each term in the input string after the beginning of the input string.
  2. Do not mark off a nested phrase at the beginning of the input string or at the beginning of another nested phrase; e.g.,
    NOT   <Data>? on <<Demand>? for <Books>>
    BUT   Data? on <Demand? for <Books>>
  3. Do not mark off a nested phrase which contains no access terms; e.g.,
    NOT   Citizens? of <@Canada>
    BUT   Citizens of Canada
  4. Mark off a nested phrase to contain a focal term (i.e., the first nonadjectival term of the phrase), terms which directly or indirectly qualify the focal term, and, if neccessary, coordinate terms; exclude all other terms; e.g.,
    NOT   @Effects? of <Literacy? on <Processing? of <Information? by <Human Beings>>>>
    BUT   @Effects? of <Literacy>? on <Processing? of <Information>? by <Human Beings>>
    { 220}
  5. Do not mark off a nested phrase which contains a focal term and one or more following qualifying terms but excludes adjectival terms qualifying the focal term; e.g.,
    NOT   Online <Databases? on <News>>
    BUT   Online <Databases>? on <News>
  6. Unless otherwise provided for, mark off a nested phrase so that the focal term is linked as strongly as possible to the focal term of the phrase within which the phrase is nested; e.g.,
    NOT   @Methods? of <Compression>? of <Machine-readable <Files>>
    BUT   @Methods? of <Compression? of <Machine-readable <Files>>>

Suppression of access ("@", and omission of "<" and ">")

  1. Within a single input string, do not define two access phrases beginning with identical terms; e.g.,
    NOT   Indexing <Languages>? & <Natural <Languages>? & >
    BUT   Indexing Languages? & <Natural <Languages>? & >
  2. Suppress access under a phrase beginning with a term which is unlikely to be looked up by searchers; e.g.,
    NOT   Use? of <STI>? in <Innovation? in <Industries>>
    BUT   @Use? of <STI>? in <Innovation? in <Industries>>

Forward- and backward-reading connectives ("?")

  1. Mark as forward-reading a simple preposition used by itself as a connective; e.g.,
    NOT   @Distribution of <Books> among <Branches of <Library Systems>>
    BUT   @Distribution? of <Books>? among <Branches? of <Library Systems>>
    { 221}
  2. If a participle or adjective plus a preposition is used as a connective, mark the preposition alone as forward-reading if the participle or adjective is required for the good qualities of all index strings; include an initial space in the forward-reading connective; e.g.,
    NOT   @Research? in <Typography>? needed for <Teletext>
    BUT   @Research? in <Typography> needed? for <Teletext>
  3. Mark " & " as forward- or backward-reading; e.g.,
    NOT   Information? on <Science & <Technology & >>
    BUT   Information? on <Science? & <Technology? & >>

<-- Appendix C: A Case Study of String Indexing Contents Appendix E: Some Organizations Involved in String Indexing -->