NNELS is constantly adding ebook and audiobook content to its collection. We purchase these files from third parties, such as distributors (like Blackstone Audio, Recorded Books, eBOUND & DeMarque) and publishers. Along with those files come all the metadata records, which need some "massaging" before they can be processed in the NNELS repository.

The metadata that comes with record sets will vary in thoroughness. We do our best to complete the necessary elements required before ingesting them - the records don't need to have every bit of metadata possible; just the fields that will be displayed in the end.

Some metadata will be added in Drupal after the files are ingested, for example Audience and Language.

When editing records in MarcEdit we want the files to be in .mrk format. When we receive files from publishers they are generally in either .xml or .mrc format. In opening screen of MarcEdit there is a button called "MARC Tools". Click on that, and on the next screen there is a dropdown menu under "Select Operation." MarcBreaker will convert .mrc to .mrk, and MARC21XML⇒MARC21 will convert .xml to .mrc. Using these 2 tools, we can convert the publisher files into .mrk files for editing. After editing records in MarcEdit we want to convert them to .xml files to ingest into the NNELS repository (see how to save to XML).

• Adhere to MARC format rules as best you can
• Include the field number, indicators and any subfields
• Ex: =245 04$aThe Codfish Dream :$bChronicles of a West Coast Fishing Guide /$cby David Giblin
• 2 spaces after the field number 245, indicators 04 (no added entry, 4 nonfiling characters - "The "), subfields $a, <space>:$b[Subtitle], <space>/$c[Author statement]
• I don't think these need to be perfect - editing records in Drupal doesn't seem to affect subfields and indicators, for example adding a subtitle in Drupal won't add a $b to the MARC record
• We are trying to find a compromise between thoroughness and expediency, for example the author should be in field 100, so if there is no subfield $c present in the 245 field that shouldn't affect how it looks in Drupal
• It largely comes down to what the records look like when they come from the publishers/distributors - we look to modify what we get so that it displays properly in the NNELS catalogue

• Records should come with LDR and other technical fields like 001, 003, 007, 008, 028, etc.
• We need to ensure these fields are present:
  • 100 (Author/Creator) - one for each author
  • 245 (Title)
  • 260 (Publisher)
  • 520 (Summary)
  • 650/related fields (Subject headings - more information below) - one for each subject heading
  • 655 (Genre - indicators will be /4 - we have a prescribed list here - only use these terms: https://wiki.libraries.coop/doku.php?id=public:nnels:publishing:projects:cataloguing:metadata-cleanup:genre) - one for each genre term

/* * 946 (If BISAC subject headings come with the records - more information below)*/

There is an automated process for adding subject headings in bulk to Record Sets. We use OCLC to add FAST and BISAC.

Here is a link to how that is done:

https://bclc.wiki.libraries.coop/doku.php?id=public:nnels:cataloguing:add-subject-headings