The Cutter in a Dewey call number comes from the Cutter-Sanborn table, and
a language model does not apply
that table accurately: for Julián Rodríguez Fernández it may return R63
or R444 when the value in the table is R696. With a wrong Cutter the
book is shelved in the wrong place.
Collect calculates the Cutter and the ISAD-G level code after extracting the record, discarding whatever the model put in those two fields. The model still proposes the title, the author and the subjects.
Dewey Cutter
In Dewey the Cutter completes the call number (082$b; in UDC, 080$b).
Collect looks up the surname in the Cutter-Sanborn Three-Figure
Revised table, about 12,000 public entries, by longest prefix. If it
finds it, that code goes in the record and in holdings (852).
If the surname is not in the table, Collect falls back to the code for the initial and that field needs review, until someone adds the surname to the table.
LC Cutter
In Library of Congress classification the Cutter (050$b) comes from an
algorithm published by LC which, unlike the Cutter-Sanborn table, yields a
result for any word. Collect calculates that value and replaces whatever
the model proposed.
The full algorithm lengthens the Cutter letter by letter, so Rodriguez
would give .R6375449, whereas in practice two or three digits are used.
Collect cuts to the configured length, two by default, and leaves .R63. If the
record has the data, it can add:
- The work mark: the first significant letter of the title, to tell works by the same author apart.
- In
CTclasses, if there is a personal name as subject, the biographee’s Cutter. A biography of Lincoln by Smith takes Lincoln’s Cutter.
The biographee rule applies only when both conditions are met (CT class
and a personal name as subject). If one is missing, the author is used.
Which classes count as biographical is an institutional decision.
ISAD-G description level
In ISAD-G each unit has a level (fonds, section, series, item) and that level is coded in the record. Whoever knows the fonds decides whether a set is a series or a section, and Collect writes the code that matches that decision instead of asking the model for it.
Limits
Each rule has a different reach. In LC the Cutter can always be calculated, while the work mark and the biographee’s Cutter need the record to carry a title and a subject, and in Dewey a surname missing from the table leaves only the initial. The ISAD-G code is written once the level has been decided, not before.
None of this removes the need to review the record. What no longer varies from one batch to the next is the Cutter for the same author and the level code.
To continue the conversation
If you catalog in Dewey or LC, or describe fonds in ISAD-G, and you want to see how these rules look with your surnames and your conventions, write to us at info@janium.com.