For discussion at the March '99 ZIG meeting.

Note: Semantics are suppled for some attribute values, and semantics will be added for more values as this attribute set definition matures, but It is not intended that additional semantics will be supplied for values for which the name itself conveys complete and unambiguous semantics. An example is "Date/time Record Added to Database".

1. Access Point
   • Date/time Record Added to Database
   • Date/time Record Last Modified
   • Date/time Record Last Reviewed
   • All Access Points
     When the origin uses 'All Access Points' it is asking the target to search for the term via all supported access points.
   • Anywhere in Record
     When the origin uses 'Anywhere in Record' it is asking the target to search for the term anywhere it may occur within the record, subject to the targets search capability and interpretation of what "anywhere within the record" means. The target might search commonly used access points, or it might search the entire content of the record.
   • Server Choice
     When the origin uses 'Server-choice' it is asking the target to select an access point (which may be defined in any attribute set, not necessarily the Utility set), and to use its best judgment in making that selection.
   • Cost
   • Score
   • Rank
   • Language
   • Local Control Number
     A string assigned by the server that uniquely identifies a record in the database. This could take the form of an integer or character string.
   • Wildcard
     May be used for a single wild card. Thus:
     "field-1 within wildcard within field-3" would match "field-1 within field-2 within field-3" and it would match "field-1 within field-4 within field-3" but would not match "field-1 within field-2 within field-4 within field-3".
   • Wildpath
     May be used for an unanchored search. Thus:
     "field-1 within wildpath" would match "field-1", or "field-1 within field-2", or "field-1 within field-2 within field-3", etc.

2. Semantic Qualifier
   • Null
     The Null value of the Semantic Qualifier attribute may be supplied within a query operand, when that operand also includes one or more Semantic Qualifier attribute values defined in attribute sets other than the Utility set.

     The server pairs each supplied value of the Semantic Qualifier attribute with the Access Point attribute to try to find a best match. (The server determines what match is best.)

     By including the value Null, the client indicates that the server may, if it chooses, ignore the other semantic qualifiers (thus allowing the server to match the Access Point attribute value with Null, in effect rendering the Access Point unqualified).

     Whenever one or more Semantic Qualifiers are included in an operand, the server must either select one or fail the search. Thus the Null value is defined in order to allow the server to choose not to select any of the real Semantic Qualifiers (i.e. those other than Null).

     The Utility set defines only the value Null. Real Semantic Qualifier values are to be defined in other attribute sets. The value Null should not be supplied unless one or more other Semantic Qualifier values are also supplied.

3. Language
   The Language attribute indicates the language of the supplied term. It is a character string based on RFC 1766 .

4. Content Authority
   
   • Null
     The Null value of the Content Authority attribute is used in a manner analogous to the Null value of the Semantic Qualifier attribute. It may be supplied within a query operand, when that operand also includes one or more Content Authority attribute values defined in attribute sets other than the Utility set.

     The server pairs each supplied value of the Content Authority attribute with the Access Point attribute to try to find a best match. (The server determines what match is best.)

     By including the value Null, the client indicates that the server may, if it chooses, ignore the other Content Authority values (thus allowing the server to match the Access Point attribute value with Null, in effect rendering the Access Point unqualified).

     Whenever one or more Content Authority values are included in an operand, the server must either select one or fail the search. Thus the Null value is defined in order to allow the server to choose not to select any of the real Content Authority values (i.e. those other than Null).

     The Utility set defines only the value Null. Real Content Authority values are to be defined in other attribute sets. The value Null should not be supplied unless one or more other Content Authority values are also supplied.

5. Expansion/Interpretation
   • Left Truncation
   • Right Truncation
   • Word-by-word Left Truncation
   • Word-by-word Right Truncation
   • Phonetic
   • Stem
   • Thesaural Expansion
   • Singular Matching
   • Plural Matching
   • Case Sensitive
   • Case Insensitive
   • Punctuation Sensitive
   • No Stopwords
     Client requests the server not to treat any word in the term as a stopword.
   • Search Words Stopped
     Meaningful only in a response query. Used by the server to indicate that it treated one or more words in the term as stopwords. (May be supplied in a submitted query but should be ignored by the server. The server should not infer any semantics based on the occurrence of this attribute in a submitted query, however, nor should the server treat its occurrence as an error, because the client may have simply resubmitted a reformulated query or query otherwise previously returned by the server, where the server included this attribute.)

6. Normalized Weight
   The weight assigned to the term, for purposes of assigning scores to records. An integer from 0 to 1000. May be attached to a term in a request query (submitted by the client in a search request) in which case the supplied value indicates the weight that the client requests be attached to the term. May be attached to a term in a response query (returned by the server in a search response) in which case the supplied value indicates the weight that the server attached to the term, or the weight that the server recommends for use in a re-submitted query. When supplied by the server the value may be the same as, or different from, the value in the submitted query. May be supplied by the server even if it was not supplied in the request query.

7. Hit Count
   May be attached to a term in a returned query in the Search response, and its value is the number of records in which the term occurs. Meaningful only in a returned query, although it may occur in a submitted query but should be ignored by the server. (The server should not infer any semantics based on the occurrence of this attribute, however, nor should the server treat its occurrence as an error, because the client may have simply resubmitted a reformulated query or query otherwise previously returned by the server, where the server included a Hit Count attribute.)

8. Comparison
   • Following two used with null term:
     • Always Matches
     • Never Matches
   • For numerical and scalar comparisons: Integer and IntUnits; Externals and other expressions that evaluate to scalar values, including date/time; and character strings (for lexical comparisons):
     • Equal
     • Less Than
     • Less Than Or Equal
     • Greater Than
     • Greater Than Or Equal
     • Not Equal
   • For range comparison (may be used with an external datatype, for example, term-1 or term-2)
     • Between
     • Contained In Bounding-polygon
   • For character strings
     • Contains
     • Relevance Feedback
     • Regular Expression
       as prescribed by IEEE 1003.2 Volume 1, Section 2.8 "Regular Expression Notation"
     • Masking
       The rules governing the usage of this attribute are as follows:

       The character '?' (question mark) is used to mask a variable number of characters. It may be followed by a positive integer, i.e. one or more consecutive decimal digits (where the first is positive) in which the positive integer represented by the string of digits (beginning with the digit immediately following the '?', up to and not including the first non-digit character), indicates a range of characters to mask, from zero up to and including the specified integer.

       When '?' is not immediately followed by a positive decimal digit, it indicates an arbitrary number of characters to mask (from zero to a system defined limit).

       The character '#' (pound or number sign) is used to mask a single character. Multiple consecutive occurrences of '#' may be used to indicate a precise number of characters to mask.

9. Format/structure
   • Character String
   • Scalar Value
   • Interval
   • URI
   • ZURL Extracted Identifier
     An identifier extracted from a z39.50r URL (same as bib-1 Urx)
   • Key
     Functionally, the bib-1 Key structure attribute. Description to be developed.

10. Occurrence
    Indicates the desired occurrence of the specified access point. Integer. For example, to indicate second author, the value of the Access Point attribute is 'author', and the value of the Occurrence attribute is 2.

11. Indirection
    • URI