I'm fine with the BL proposals, but in the end we aren't going to solve the problem of undifferentiated name authorities by adding more attribute-based authorized access point qualifiers to RDA. In some cases the evidence for differentiation is so tied to a particular resource (e.g., stray bits of circumstantial information about an author of archival correspondence) or so difficult to express (e.g., a style difference between to identically named contemporary illustrators) that it offers no good basis for an authorized access point qualifier. 

We need a technique that will ALWAYS enable differentiation, such as the default inclusion of a system number in comparisons when a 1XX heading string is identical to another 1XX and the two identities are distinct in the cataloger's judgment. This may be something that gets done in authority systems, e.g., by a change to the NACO Heading Comparison rules rather than by changes to a rule set like RDA. So I support the BL proposals as a way of making entities more identifiable from authorized access points, but not as a solution to the problem of undifferentiated access.

Stephen

On Mon, Oct 8, 2012 at 5:35 AM, Moore, Richard <[log in to unmask]> wrote:
I came across n 2012041419, an RDA NAR which has been made into an undifferentiated name. I think it illustrates the difficulties inherent in creating undifferentiated NARs under RDA, and the important of being able to make all authorised personal name access points unique.

010   $a n 2012041419
040   $a DLC $b eng $c DLC $e rda $d DLC
1001  $a Zhang, Xiaomei
374   $a Translator
377   $a chi $a rus
4001  $a 张晓梅
667   $a Non-Latin script reference not evaluated.
670   $a [Translator of 1955 Nian Xizang ji xing]
670   $a 1955 Nian Xizang ji xing, 2010:$$bt.p. (张晓梅 = Zhang Xiaomei [in rom.] translator)
670   $a [Author of Heilongjiang Sheng guo you lin quan gai ge bao zhang ti xi yan jiu]
670   $a Heilongjiang Sheng guo you lin quan gai ge bao zhang ti xi yan jiu, 2010: $b t.p. (张晓梅 = Zhang Xiaomei) vita (B.A. from Dongbei nong ye da xue zhi wu bao hu xi in1989; master and doctor degrees in Guan li xue in 2001 and 2005)
675   $a HKCAN database, viewed Sep 19, 2012

374 and 377 appear to have been included with the original identity. When a second identity was added, no further 37X could be added, as it would immediately have been unclear to whom they referred.

In this case, should one of the authors subsequently be removed to a new, differentiated NAR, it is reasonably clear to whom the 37X belong (but only because none have been added for the second identity). In a record containing multiple identities, this would not be clear. Worse, if 046/37X had been added for more than one identity, confusion would ensue, both when examining the undifferentiated NAR, and in trying to move any identity to a new, differentiated NAR. The only safe option would be to remove any existing 046/37X, and not add any more.

If undifferentiated RDA NARs are to be created, it seems the choice is between confusion and inaccuracy if discrete data elements are included, and loss of data if they are not.

It seems better to aim for an differentiated NAR in every case. The BL has aimed to do this since our announcement earlier in the year that we would no longer create undifferentiated NARs. Since then we have pushed our notion of "Profession or occupation" as far as we consider reasonable, and find that it mostly works. However, in a small but annoying number of cases, when information is not available, RDA leaves us short of a qualifier. This is the reason for the BL RDA change proposals 6/JSC/BL/3 and 6/JSC/BL/4, which can be found at http://www.rda-jsc.org/workingnew.html.


Regards
Richard

_________________________
Richard Moore
Authority Control Team Manager
The British Library

Tel.: +44 (0)1937 546806
E-mail: [log in to unmask]




--
Stephen Hearn, Metadata Strategist
Technical Services, University Libraries
University of Minnesota
160 Wilson Library
309 19th Avenue South
Minneapolis, MN 55455
Ph: 612-625-2328
Fx: 612-625-3428