[SAC-FAST] Comments on Periods sample records

[Laura Akerman]

I was asked to comment on the periods sample - I will review the others 
but comments will not be as organized. I'll bring these to the meeting.

Laura

Notes on review of Period records

1. Use of FAST dates could provide for more precise subject searching by 
date range. But...

Because LCSH has not provided for chronological subdivision for date 
ranges narrower than a century, converted chronological subdivisions 
result in FAST temporal facets with broader dates than the material 
covers, which could be misleading. E.g.

ocm00137013
245 10 $a Cambridge book of English verse 1900-39
650 0 $a English poetry $y 20th century.
648 7 $a 1900 - 1999 $2 fast
650 7 $a English poetry $2 fast

The title limits the time period to 1900-1939, but the FAST heading is 
for the whole century.
If FAST were being applied directly, the 650 should probably be coded 
655 Genre/form.

Another example:

ocm00492049
100 0 $a William, $c of Malmesbury, $d ca. 1090-1143.
245 10 $a William of Malmesbury's Chronicle of the kings of England. $b 
 From the
earliest period to the reign of King Stephen. $c With notes and 
illustrations. By J.A.
Giles.
651 0 $a Great Britain $x History $y Norman period, 1066-1154.
648 7 $a 1066 - 1154 $2 fast
655 7 $a History $2 fast

A person looking for the subject in the "1150's" using a subject date 
qualification based on the FAST temporal heading would get this work in 
the results. But the author died before the 1150's so no material about 
that era would be included.

Another example:

ocm00631917
100 1 $a Monsma, Stephen V., $d 1936-
245 10 $a American politics; $b a systems approach $c [by] Stephen V. 
Monsma.
260 $a Hinsdale, Ill., $b Dryden Press $c [1973]
651 0 $a United States $x Politics and government $y 1945-1989.
648 7 $a 1945 - 1989 $2 fast
650 7 $a Political science $2 fast
651 7 $a United States $2 fast

The date range of the LC chronological subdivision, 1945-1989, extends 
beyond the date of publication of the book.


2. Conversion of "period" chronological subdivisions to a set of dates 
loses some important words.

ocm00355648
245 13 $a An introduction to Tudor drama, $c by Frederick S. Boas ...
650 0 $a English drama $y Early modern and Elizabethan, 1500-1600 $x 
History and
criticism.
648 7 $a 1500 - 1600 $2 fast
650 7 $a English drama $2 fast
655 7 $a Criticism, interpretation, etc. $2 fast

Granted, there is not currently a heading for "Elizabethan drama" 
(though there are headings for Art, Elizabethan).

Relying on wikipedia:

The Tudor period usually refers to the period between 1485 and 1558, 
especially in relation to the history of England. This coincides with 
the rule of the Tudor dynasty in England. Usually the term is used more 
broadly to include Elizabeth's reign as well (1558 – 1603), although 
this is sometimes treated separately as the Elizabethan era.

So, the book may indeed cover Elizabethan drama, but that word is lost 
from the record.


3. Use of FAST dates could provide for more precise subject searching by 
date range. But...

Open date ranges are then problematic.

A. The date of publication would in almost all cases be a "closer" on 
the temporal range of the topic or genre, but this is not incorporated 
into the FAST temporal heading.

B. I still see using "Since [date]" instead of an open date as being a 
bit harder for machine processing to deal with. We don't have a data 
standard for expressing date ranges / open dates in these fields so it's 
sort of a moot point.

ocm00482554
100 1 $a Martin, Judith, $d 1938-
245 14 $a The name on the White House floor, and other anxieties of our 
times.
260 $a New York, $b Coward, McCann & Geoghegan $c [1972]
651 0 $a Washington (D.C.) $x Social life and customs $y 1951- $v Anecdotes.
648 7 $a Since 1951 $2 fast
650 7 $a Manners and customs $2 fast
651 7 $a Washington (D.C.) $2 fast
655 7 $a Anecdotes $2 fast


General comments:

Conversion of chronological subdivisions to FAST temporal headings 
really magnifies the severity of limitations in LC subject headings to 
express temporal aspects of a work. For example, expressing the temporal 
aspect of a work whose subject covers a single year as a century range 
really limits what users can get from the data. This is something a 
"next generation" LC subject headings (or whatever vocabulary may be 
called) should address.

Many chronological subdivisions appear from what we're given in the 
records to translate nicely to FAST, particularly if the subject of the 
work is a war or period for which LC has established a close 
chronological date range.

Many of the examples gave a heading for what was essentially a genre, 
e.g. "650 0 $a Jazz $y 1971-1980." applied to a sound recording. To 
provide useful searching, either a "period" approach (grouping all works 
that are of a particular span of years together by applying a date span 
"heading" to all of them), which is what results from FAST conversion of 
chronological headings, could be used, or the individual dates of 
creation would be the data we record in a temporal heading, and then 
users or applications could group them as desired. The latter approach 
provides more freedom which is what I think we want. Combining records 
that use different approaches will be challenging indeed.

Might it be useful to encode the "provenance," or the approach used to 
generate a FAST temporal heading, in a subfield... ?