The Zthes XML schema: changes between versions 0.5 and 1.0

1st May 2006
  • Added new top-level element <postings> (optional, repeatable), occurring before <relation>. This is a structured element, containing <sourceDb>, optional <fieldName> and <hitCount>. (The postings sub-record has in fact been in the abstract model and Z39.50 profile since version 0.3b of 26th July 1999, but the XML DTD was never updated to reflect this change.)
  • Added new top-level ``control field'' element <termUpdate>, optional, immediately after <termId>. This is used during batch imports of Zthes records to indicate whether each record shoud be added to the database or deleted from it.
  • Added new top-level elements, <termVocabulary>, <termCategory>, <termStatus>, <termApproval> and <termSortkey> between <termLanguage> and <termNote>. All are optional; <termCategory> is repeatable, the others are not.
  • The <termNote> element is now repeatable, and may carry a disambiguating label attribute. The allows multiple separate notes to carry different information, e.g. the source of the term or its scope. The <termNote> element also may now carry an optional vocab attribute: if specified, this is the URI of a separate Zthes database that prescribes a controlled vocabulary for the field.
  • The <relation> element may now carry an optional weight attribute, indicating the strength of the relationship between the main term and that described in the relation sub-record.
  • The <sourceDb> element must now contain a resolvable URI, but the scheme is left to the application to decide. Obvious choices would be a z39.50s URL for records made available via Z39.50, and the http base-URL of an SRU server for records made available via SRU; no doubt there are others. (This is not to be used to point to a static XML document containing the whole thesaurus in XML format: the idea is that it points to a service, not to data; the service resolves the term-ID.)
  • In addition to the <term>s, the top-level <zthes> wrapper element may also contain a single, optional <thes> element describing the thesaurus as a whole. If included, this must come before the first <term>. This record contains Dublin Core elements in the usual DC namespace as described at dublincore.org/documents/2003/04/02/dc-xml-guidelines, and an additional, repeatable, label-bearing <thesNote label="whatever"> element analogous to <termNote>.