Formatted contents notes


One of the first changes seen at the outset of format integration (in fact, a bit before FI if memory serves) was the option of subfielding author and title information in MARC field 505 (Formatted contents note). The toggle for recognition of this information is the second indicator, which is left blank in non-formatted 505s. A second indicator value of "0" indicates that an "Enhanced" contents note follows, with the following subfields indicating information which is thus marked for indexing:

For example:


        505 00  |gv. 1.|tThe history of Grunge /|rby Kurt Cobain --|gv. 2.
                |tHead-banging and heavy metal /|rby Ozzy Osborne --|gv.
                3.|tThe Celtic revival /|rby Ashley McIsaac.

allows for the inclusion of information from each |t in the title browse and/or keyword indexes, with |g and |r excluded. Naturally, all subfields would remain searchable within the general keyword index unless you excluded these by subfield. Non-formatted contents notes would be accessible *only* through a general keyword search.

The example itself points to one of the difficulties with this approach. Contents notes, like the 245 |a|n|p|c, are transcriptions of natural language information (as it appears on the piece), not controlled through regularized terminology such as uniform titles or authority-controlled name headings. Natural language contents notes contain initial articles, which some libraries are masking using |g and others are omitting, and authors which are transcribed in forename/surname order often following "by". For this reason it obviously does not make sense to place |r in the author browse index since the collocation function of the author index would be compromised in this way. On the other hand, they might be sensibly placed in the author keyword index if we didn't worry about patrons NOT understanding why they could find the author in a keyword search but not a browse.

I propose that for now Cataloguing undertake to use formatted contents notes in places where they make sense. The following make sense (to me, at least) to treat in this manner (you may think of others):

Conference proceedings where the 505 includes titles and authors of presented papers.

Festschriften and other compilations of academic articles

Multi-volume sets where individual volumes bear distinctive titles. This would end the need for transcribing title twice--once in the contents note and once in corresponding 7xx.

Literary works where the 505 includes titles and authors of included short stories, poems, novellettes, etc. In cases where there are only a few works in the collection, we should stick to current practice of establishing author/title uniform titles in 700|a|t. Under Unicorn version 8, 7xx|t will finally move into the title indexes.

Works which represent compendiums of various specific laws, standards, etc. likely to be requested by specific title. This one has the potential to be overwhelming in some cases...
I propose the following NOT receive formatted contents notes:

Scores and sound recordings where controlled access (uniform titles and authority-controlled name headings) is the preferred means of access.

Contents notes with generic and/or non-distinctive titles (eg. v. 1. Executive summary -- v. 2. Final report -- v. 3. Appendices)

Chapter headings and other cases where primary patron access would be by subject rather than by author or title.
Furthermore, I propose that we consider indexing |t in both the title browse and title keyword indexes in Unicorn. We may wish to consider the possibility of indexing authors in the author keyword index only, after consideration of the patron expectation problem noted above.


First version: 21 April 1996.
Last revision: 30 May 1996.
Back to: MARC field view of local policies | Cataloguer's Toolbox.