Colleagues :
Hi -- I agree with John Hostages's analysis of this proposal and agree
that his examples better illustrate the value of the proposed changes.
Something else about the examples in the paper reminded me of what I
believe may be a change that needs to be made in the MARC documentation
though. I don't believe many people would argue that recording serial
publication patterns and holdings is often confusing and very
complicated. Nonetheless, the point of doing so is to allow our systems
to precisely and consistently record, exchange and display the accurate
serial holdings of individual institutions.
As John points out, this requires using not only the MARC 21 Holdings
format for recording and exchanging this data, but also ANSI/NISO
Z39.71-1999 to generate accurate, consistent (and understandable)
holdings displays.
In the first 853 example of Section 2.1.2 of this proposal, and in at
least five of the examples used in the 853-855 Captions and Pattern --
General Information section of the MARC 21 Holdings format, the 'day'
element is recorded as a third level of chronology.
When you stop to think about it though, when a serial requires a 'day'
designation in addition to a month (generally for dailies,
semi-weeklies, weekly, bi-weekly, and semimonthly publications), the day
element is so dependent upon the month element as to represent a single
level of chronology -- much like the serial 'series' designations that
are combined with the highest level of enumeration as discussed in this
proposal and in John's response.
(Interestingly, in the MARC 21 Holdings Format 853-855 section, the
display examples shown for publication patterns with this 'day' 3rd
level of chronology, display the resulting piece statement with only two
levels of chronology (i.e., 853-855 p. 3 has a publication pattern:
854 00 $8 1 $a v. $b. suppl. $i (year) $j (month) $k (day)
864 40 $8 1.1 $a 16 $b 1 $i 1977 $j 06 $k 01
However, the display reads: v. 16: suppl. 1 (1977: June 1).
Levels of enumeration and chronology are always separated by colons, so
if a 3rd level of chronology were actually intended here, the display
example should read:
v. 16: suppl. 1 (1997: June: 1)
Now nobody records or displays dates like that, so if indeed part of the
point of recording and displaying serial holdings is to construct them
in a manner understandable to OPAC users, why would librarians choose to
do so?
I believe the above publication pattern should be constructed something
along the lines of:
854 00 $8 1 $a v. $b. suppl. $i (year) $j (combined month/day)
864 40 $8 1.1 $a 16 $b 1 $i 1977 $j June 1
And then, in fact the resulting display example would in fact read as
the current text does.
v. 16: suppl. 1 (1977: June 1)
I found further substantiation for constructing patterns and displays in
this manner in section 5.5.5.2 Dates (p. 34) of ANSI/NISO Z39.71-1999,
which states:
"When chronology below the first level is recorded, use a colon to
separate the year from the month. In recording chronology data that
contain day notations, do *not* treat these as a separate hierarchical
level, and do *not* separate them with a colon." [Emphasis added].
All of this is not to suggest that there are not times when another
level of chronology is necessary. One example I can think of, and one
cited in ANSI/NISO Z39.71-1999, is the example of many newspapers that
used to publish both a morning and evening edition (I don't believe
there are too many that continue to do that these days, but it used to
be quite the norm.)
I hope this is helpful (and understandable!)
everett
MARBI Rep. to CC:DA
**************************
Everett Allgood
Serials Cataloger
New York University Libraries
[log in to unmask]
212 998 2488
|