Notes on the prototype and testing environment

1. A diagram of the flow of transformations
2. An XML schema for MARC rules

The purpose of the XML scheme for MARC is to provide a formalism for the representation of MARC rules and human oriented information currently held in MARC manuals. The scheme will allow not only HTML, PDF, Windows HELP version of the MARC manual, all generated automatically, but also the production of stylesheets for record transformation with validation or decoding purposes.

See file FORMAT.html
3. Generation of valitation stylesheets

The RecordValidator.xsl stylesheet is automatically generated by the RecordValidatorGenerator.xslt stylesheet. This stylesheet uses the information contained in the sample subset of the UNIMARC manual files to build validation rules to create the validation stylesheets.

The RecordValidator.xsl has the following features:

  • It's automaticaly generated from an XML version of the UNIMARC manual.
  • Handles positional fields.
  • Checks for mandatory fields.
  • Creates variables to hold the content of the current field and the value of the leader. These variables are used for tests to decide which rules to apply.
  • Handles ranges of legal values in positional fields tests and indicator values.
  • Uses vocabularies of acceptable values in positional field testing.
  • Adds templates for positional fields validation:
    • validatePSubfield
    • checkPSubfieldValue
    • checkRangeCode
    • generatePSubfieldError
4. Generation of decoding stylesheets

The HTMLFormater.xsl stylesheet is automatically generated by the HTMLFormaterGenerator.xslt. This stylesheet uses the information contained in the sample subset of the MARC manual files to create the decoding stylesheets.

5. HTML formating of rules

The FORMATtoHTML.xsl stylesheet transforms the XML version of the sample subset of the MARC manual in an HTML document for referencial purposes.

6. How to use the sample set of stylesheets

Requires Java VM 1.3 or later.

At the DOS prompt or Linux/Unix terminal:

  1. Decompress the archive into a directory. This will be called the MARC_DIR.
  2. Change dir to MARC_DIR.
  3. Three sub-directories have been created:
    • bin : java libraries and two short scripts for testing: (Unix) and transform.bat (Windows)
    • doc: documentation, specially these notes and representation of the scheme in file FORMAT.html
    • src
      • src/schemas: Schema for the MARC manual
      • src/stylesheets
        • FORMATtoHTML.xsl: Generates an html version of the UNIMARC manual
        • HTMLFormaterGenerator.xslt: Generates a MARC to HTML stylesheet
        • RecordValidatorGenerator.xslt: Generates a stylesheet for record validation
      • src/xml: Sample records and the FORMATDescription.xml file that describes the localization of the sample UNIMARC manual files
      • src/sml/manual: Sample subset of the UNIMARC manual
  4. Examples (run from MARC_DIR):
    The and transform.bat are simple scripts that call the command line interpreter of the Saxon java XLST processor. They take three arguments: XML document, XLS stylesheet, output file name. They will apply the stylesheet to the XML document and save the result in the output file.
    Generate a HTML version of the MARC manual:
    ./bin/ src/manual/Unimarc0.xml src/stylesheets/FORMATtoHTML.xsl output/Unimarc0.html
    Generate a Stylesheet for decoding a record in HTML
    ./bin/ src/xml/FORMATDescription.xml src/stylesheets/HTMLFormaterGenerator.xslt output/HTMLFormater.xsl
    Now HTMLFormater.xsl is able to render a record in HTML:
    ./bin/ src/xml/Record.xml output/HTMLFormater.xsl Record.html
    Generate a validation stylesheet from the MARC manual:
    ./bin/ src/xml/FORMATDescription.xml src/stylesheets/RecordValidatorGenerator.xslt output/RecordValidator.xsl
    Now RecordValidator.xsl is capable of validating specific records
    ./bin/ src/xml/Record.xml output/RecordValidator.xsl output/RecordErrors.xml
7. Additional examples

This set of additional examples is intended to demonstrate the RecordValidator.xsl capabilities regarding error detection.In order to test these examples several errors are going to be inserted in the Record.xml file. Use the command: ./bin/ src/xml/Record.xml output/RecordValidator.xsl output/RecordErrors.xml

  1. Test mandatory control fields and unknown fields.
    Replace the 001 control field with:
    <controlfield tag="999">450981</controlfield>
    <error type="MandatoryControlfield" tag="001"/>
      <warning type="UnknownControlfieldTag"> 
      <controlfield xmlns="" tag="999">450981</controlfield> 
  2. Test subfield codes.
    Replace the 200 data field with:
    <datafield tag="200" ind1="1" ind2=" "> 
      <subfield code="a">UNIMARC</subfield> 
      <subfield code="y">Manual de operações</subfield> 
      <subfield code="f">Fernanda Maria Guedes de Campos, José Carlos Sottomayor</subfield>
    <error type="InvalidSubfieldCode" tagID="d0e49"> 
  3. Leader contents
    Replace the leader content with:
    <leader>00614kam  22002051  450 </leader>
    <error type="Leader" domain="RECORD-STATUS" start="6" length="1"> 
      <content>00614kam  22002051  450 </content> 
        <OPTION xmlns:xsi="" 
         name="Corrected Record" value="c"/> 
        <OPTION xmlns:xsi="" 
         name="Deleted Record" value="d"/> 
        <OPTION xmlns:xsi="" 
         name="New Record" value="n"/> 
        <OPTION xmlns:xsi="" 
         name="Previously Issued Higher Level Record" value="o"/> 
        <OPTION xmlns:xsi="" 
         name="Previously Issued as an Incomplete, Pre-publication Record" value="p"/>