330 likes | 555 Views
Better, Not Perfect : Cataloging and Data Manipulation Strategies for Improving OCLC Records for Licensed Digital Resources. James L. Soe Nyun, Head, Music & Media Metadata Kurt D. Hanselman, Metadata Specialist, Music & Media Metadata. The Problem.
E N D
Better, Not Perfect :Cataloging and Data Manipulation Strategies for Improving OCLC Records for Licensed Digital Resources James L. Soe Nyun, Head, Music & Media Metadata Kurt D. Hanselman, Metadata Specialist, Music & Media Metadata
The Problem New licensed electronic resources, lots of them A promise to catalog many of them at the title level Flat staffing allocation
Some Strategies Others Have Used Catalog just the resource, direct users to the home page for the resource and let them use the vendor interface to find content Rely on an external knowledge base that will have records for resources in a licensed collection Load non-OCLC vendor records into the local ILS Harvest or purchase OCLC catalog records, load into local ILS
Record Quality Issues Some records are great! Often (really often) they are not Loading these unedited records can wreak havoc on an OPAC Some work can be done using local database management tools and vendor cleanup Corrections and improvements done post-load generally stay local and are not contributed back to the cataloging community
Impediments to Collaborative Cataloging Sheer volume of titles Siloes: Library systems haven’t made it easy to share all our work Priorities The “perfect catalog record”
The Glitches Round-tripping records back into Connexion uncontrols any headings that might have been controlled The MarcEdit OCLC Bibfile Reader plugin has a small issue with 007s …but we’ve got workarounds for those issues, details to follow But also some weird happenings we haven’t completely figured out…
Click to add title A couple options… Utilize KBART file from accessed through Collection Manager Possible alternative: perform access Method search w/ 856s
KBART method Generate search keys to use in Connexion OCLC# + “not li:CUS” Run batch search using search keys Review failed searches in the search report Deleting any records already held locally, and adding records with erroneous holdings information Once you’ve got your local file sorted out, you’re ready to start editing in MarcEdit!
Click to add title First step to editing records in MarcEdit: task lists Some tasks are simple, others more complex
MarcEdit Task Lists Tidy up bad 007s Fix 008 fill character Add “composer” designator Fix Lang: code (for instrumental music) Fix Lang: code (for vocal music) Provider-neutral cleanup Remove bad 856s …and others
MarcEdit Task Lists Tidy bad 007s
MarcEdit Task Lists Fix 008 fill character
MarcEdit Task Lists Add “composer” designator
MarcEdit Task Lists Fix Lang: code (instrumental music) (=008 .{35}).{3}(.*=650 .{2}\$a)(Sonatas|Triosonatas|Symphonies|Overtures|Concertos|Concertigrossi|Dances|Suites|Variations|Rondos|Instrumentalmusic|Instrumental ensembles|Keyboard|Piano|Organ|Harpsichord|String|Violin|Viola|Cello|Double bass|Woodwind|Flute|Piccolo|Altoflute|Bassflute|Oboe|Englishhorn|Clarinet|Altoclarinet|Bassclarinet|Contrabassclarinet|Basset horn|Bassoon|Contrabassoon|Brass|Horn|Trumpet|Cornet|Flügelhorn|Euphonium|Trombone|Bass trombone|Tuba|Percussion|Timpani|Xylophone|Marimba|Vibraphone|Plucked instrument|Harp|Guitar|Lute|Wind|Band|Orchestra|Chamberorchestra|Trios|Quartets|Quintets|Sextets|Septets|Octets|Nonets)(.*)/m
MarcEdit Task Lists Fix Lang: code (vocal music)
MarcEdit Task Lists Provider-neutral cleanup
MarcEdit Task Lists Remove bad 856s
MarcEdit Task Lists Identify missing 856s
Click to add title Batch macro operations Übermacro Batch headings check macro Selectively control headings still uncontrolled Replace master records and attach our holdings
Übermacro Simplified batch version of a complex macro devised to ease the transition into RDA This simplified version will go through a save file and, starting with the first record: Control all controllable headings Map $4 MARC relator codes to $e relationship designators Save the record and move on to the next record until the file has been processed
Batch Control Report Macro Usually used in conjunction with Übermacro, but could be used alone Goes through a save file and does one or both of the following: If desired, if not paired with Übermacro: Controls all controllable headings Generates a tab-delimited report of any uncontrolled headings
General Example Before After
20th Century Example Before After
Post-Cataloging “Features” to Look Out For Possibility the other institutions will further improve our records—Good! Possibility that removed institution-specific or provider-specific information will be reinstated Possibility that an improved record will get merged with a far less wonderful record, wiping out many improvements
Adopting this Workflow: Things to Consider This process requires some technical expertise beyond a basic cataloger skill set: Excel, working with delimited text files, sorting, string manipulation, possible use of regular expression plugin MarcEdit, including regular expressions, OCLC Bibfile Reader plugin OCLC Connexion macros: modifying, maybe creating Troubleshooting inevitable breakdowns in the tech above, issues in moving data from one program to another, possible issues added by local networks
Adopting this Workflow: Further Things to Consider This process is slower (=more expensive) than simply loading unedited records, or manipulating records only for local use Some of the stages in the workflow depend on tools that are in continuous development (MarcEdit) or represent a stable platform that now may undergo changes to support user needs (Connexion) A workflow to enhance master records needs to take into account the local resources and balance those resources against the importance to the local collection of the materials being cataloged E.g., you may opt to enhance only for certain sets, or for only certain titles within the sets
Sharing our Work General-use versions of some of the macros and MarcEdit task lists discussed here will be available at UC San Diego’s Technical Processing Online Tools (TPOT) website: http://tpot.ucsd.edu/toolbox/tools-and-tips/index.html