[Yulcat-l] resending LC's comments w/ additions

[Joan Swanekamp]

Just from LC:


All, my apologies for resending this, I received some additional comments 
on LC's response. The substantive changes are marked by ** and occur in the 
last paragraph of the maintenance section and in the summary section:

LC response to the proposal to add additional records for e-serials to the 
CONSER file

Library of Congress Serial Record Division (SRD) discussed the OCLC 
proposal to add additional records for e-serials with OCLC in July and 
August 2005. The following LC response includes comments from other areas 
of LC as well as SRD staff. Comments were received from staff working with 
the ILS, those implementing LC's OpenURL resolver and ERMS, reference 
staff, digitization projects staff as well as SRD staff.

LC recognizes the importance of providing bibliographic records and 
holdings information for e-serials -- and of committing to a strategy that 
aggressively targets the inclusion of these records in the CONSER file. LC 
believes, however, that these e-serial records must continue to support the 
community's critical need for authoritative bibliographic records and 
accurate ISSN data. LC agrees with other CONSER comments that the timeline 
for adding these records to the CONSER file should be extended to complete 
OCLC's duplicate resolution process.  In addition, LC recommends that the 
PCC Policy and Standards Committees be consulted as part of the overall 
planning process. LC also recommends that CONSER and OCLC work with vendors 
to better understand how these additional records will impact "downstream" 
products using the CONSER file.

ISSN
E-serial web service linking and cross-database record maintenance rely 
heavily on accurate ISSN and eISSN matching.  LC staff have noticed 
incorrect or misleading ISSN in a random sampling of the auto-generated 
e-serial records. Cloning the print record involves flipping the print ISSN 
to subfield y of the 022 and subfield x of the 776 field in the record for 
the e-serial. For those serials not assigned an ISSN by NSDP or ISSN 
Canada, there is no guarantee that this is the correct ISSN for the print, 
since CONSER libraries do not all have the capability to check the ISSN 
Portal for correct ISSN when records are first authenticated.  Moreover, 
there is no systematic maintenance of non-U.S. ISSN in the CONSER 
database.   While the problem of ISSN maintenance for non-U.S. records is 
not new to the CONSER file, the additional records may amplify problem 
situations.

Currently mechanisms are not in place to provide correct ISSN information 
in CONSER records directly from the ISSN Network database. The "dirty" ISSN 
data problem in the CONSER file will continue until CONSER members, OCLC, 
and vendors of e-resource products use authoritative ISSN data.

The Medium-Neutral ISSN (MNI) being developed as part of the revision of 
the ISSN standard  will provide a new means to support aggregation of ISSN 
assigned to separate format versions of serials. Implementation of this 
important new identifier in the CONSER database will require increased 
cooperation and coordination with the ISSN International Centre in 
Paris.  Distribution of the MNI via CONSER records will provide the 
potential to enhance record displays for the utilities, ILSs and e-resource 
management products. LC recommends that CONSER begin now to plan for how 
the ISSN revision process will impact separate version records in the 
CONSER file.

Vendor use of the records
CONSER and OCLC need to ensure that down-stream system vendors are able to 
identify and understand the characteristics of auto-generated e-serials 
records marked 042$a oclcderive. LC is aware, for example, that some 
vendors do not use CONSER records coded msc and others use only CONSER 
print records for record set products.

Maintenance
LC staff expressed concern that maintenance responsibility is ambiguous for 
these auto-generated e-serial bibliographic records and for e-serials 
holdings data.  Will this large number of e-serial records added to the 
CONSER database be actively monitored for major changes in titles, issuing 
bodies, etc.? Will 776 links between print and e-serial records be 
maintained as titles change?  Will the traditional methods of maintenance 
(CONSER members catching changes and reports to OCLC Quality Assurance from 
other OCLC members) assure that new records will be created for e-serial 
title changes?

CONSER guidelines will also be needed for records created during the pilot 
to ensure that successive e-serial records, based on successive print 
records showing minor title changes, are handled appropriately.

While it appears that OCLC's e-serials holdings knowledgebase will be kept 
up to date as aggregations change during the pilot period, what happens 
after the pilot ends? Who will be permanently responsible for updating 
holdings data?  Do e-serial holdings reflect "master" holdings, **in the 
sense of an idealized run of all published holdings (as opposed to the 
holdings or a particular institution)**, or local institutional holdings?


Duplicates
LC understands that the pilot provided a good testbed for tuning the 
matching algorithms used to auto-generate e-serials records.  It appears 
that duplicate consolidation of records coded "OCLCS" added in June is 
still ongoing. While many records have been consolidated, LC staff continue 
to find a number of OCLCS duplicates.

What is the timeframe for completing consolidation? What matching processes 
will be used for avoiding future duplicates as much as possible? LC 
believes that, for all users of the CONSER database and distributed CONSER 
products, it is important that the e-serials record creation process 
minimize the addition of duplicate records.

URLS
Maintaining accurate URL links for e-serial titles is a problem CONSER has 
struggled with.  Current practice recommends that CONSER records avoid the 
use of institution -specific links, some of the aggregator-specific links 
in these records appear institutional specific.


LC staff found that URLs in the auto-generated e-serial records vary in 
accuracy and usefulness -- some portion were incorrect, dead, or otherwise 
not working; some represented aggregator-specific links.  LC staff found a 
number of duplicate e-serial records where the only difference was 
the  aggregator-specific URL.  How do we establish and enforce guidelines 
that prevent record creation based on the occurrence of a new or different 
e-distribution channel for the same work?

OCLC has also indicated that, at this point, there are no plans for 
checking URLs. LC recommends that CONSER work with OCLC to ensure that URL 
links to non-publisher sites on these auto-generated e-serial records 
follow current CONSER guidelines.  In addition, LC suggests that CONSER and 
OCLC continue to explore options for checking CONSER URLs with e-serial 
vendors and publishers.

Titles in article databases
LC staff reviewing records which failed the ISSN matching process have 
discovered that many of these auto-generated e-serial records represent 
titles for journal articles found in article databases rather than full 
e-serials that have complete issue coverage. CONSER should continue to 
discuss the impact of adding records in this category of e-resource to the 
CONSER file. For example, currently, only e-serials that are published as 
separate electronic journals rather than as articles licensed to article 
databases are eligible for ISSN assignments.  Thus, many of the 
auto-generated e-serial records will not carry separate ISSN, a potentially 
confusing situation, especially in light of the revised ISSN standard where 
the policy of separate ISSN for e-resources is clarified and 
mandated.  Other problematic situations were cases where titles that are 
serials in print, such as A&I services and some directories become 
integrating resources in their electronic forms.  Cloning print records for 
these kinds of resources results in inaccurate and often duplicative serial 
records for titles that should be cataloged as integrating resources.

Overlap with digitization projects
LC staff working with digital reformatting projects asked for clarification 
on possible overlaps or conflicts with the cataloging practices established 
with the DLF/OCLC Registry of Digital Masters.

Staff associated with projects involving digitization of serials originally 
in print, e.g., the National Digital Newspaper Program, also wondered what 
types of material were included in the records to be added to the CONSER 
file.  Understanding the criteria for inclusion, and thus the likely 
overlap, would help those involved in such projects at LC and other 
institutions.  In particular, it would allow them to determine whether they 
need to adjust plans, processes, or workflows.

Summary
Recommendations on adding auto-generated records to the CONSER file:

Begin discussion with the PCC Policy Committee (PoCo) and the PCC Standards 
Committee **on the impact of adding these records to the CONSER file. **

CONSER members, PoCo, PCC standards Committee and OCLC should develop a 
plan for addressing concerns raised by members and a new timeline **for the 
pilot should be considered. Continued planning for an agreed strategy 
related to pilot titles should  addressed ** at the ALA mid-winter 2006 
CONSER at-large meeting.

CONSER members, PoCo, PCC standards Committee and OCLC should develop a 
plan to ascertain how the records will impact the products of vendors that 
subscribe to the CONSER file. Continue planning at ALA mid-winter 2006 
CONSER at-large meeting.

Recommendations on e-serial cataloging **for records to be included in the 
CONSER file.**

Call for dialogue at Midwinter CONSER at-large meeting to brainstorm the 
scope/goals/intent of e-serial cataloging in the CONSER file.