VIAF Council Meeting Minutes Hilton Lyon, Lyon, France
Transcrição
VIAF Council Meeting Minutes Hilton Lyon, Lyon, France
VIAF Council Meeting Minutes Hilton Lyon, Lyon, France Friday, August 15, 2014 Attendees VIAF Council Institutions: Edita Lichtenbergova (Czech Republic – National Library) Ossama Mahmoud, Rania Osman (Egypt – Bibliotheca Alexandrina) François Mistral (France – ABES) Vincent Boulet, Françoise Bourdon, Pauline Chougnet, Gildas Illien (France – Bibliothèque nationale de France) Brigitte Wiechmann, Lars Svenson (Germany – Deutsche Nationalbibliothek) Tadahiko Oshiba (Japan – National Diet Library) Stefanie Zutter (Luxembourg – National Library) Mikolaj Baliszewski (Poland – National Library) Leszek Śnieżko (Poland – NUKAT) Ricardo Santos Muñoz (Spain – Biblioteca Nacional de España) Miguel Moreira (Switzerland – RERO) David Farneth (United States – Getty Research Institute) Beacher Wiggins, Susan Morris (United States – Library of Congress) Observers: Pat Riva (Canada – Bibliothèque et Archives de Québec) Cécile Davrieux-de Becdelièvre (France – Bibliothèque municipale de Lyon) Jihae Jeon (Korea – National Library) OCLC representatives: Ted Fons, Janifer Gatenby, Thom Hickey, Boaz Nadav Manes, Glenn Patton, Richard Wallis, John DeSantis (OCLC Global Council) Welcome and Introductions Brigitte Wiechmann, 2014 VIAF Council Chair, welcomed attendees and thanked OCLC for making meeting arrangements. Election of New Chair-Elect The incoming chair for 2015 is Vincent Boulet (Bibliothèque nationale de France). He became the chair following the Lyon meeting. Beacher Wiggins (Library of Congress) agreed to be nominated as chairelect and was unanimously elected. Beacher will serve as chair-elect in 2015 and as chair in 2016. VIAF Update Thom Hickey, Chief Scientist, OCLC Research, presented an update on the current status of VIAF. He began by noting that the 2014 annual report was issued just prior to the Lyon meeting. His update first emphasized infrastructure changes in the past year. He described xR records which grew out of OCLC Research’s Multilingual Bibliographic Structure project which mines the bibliographic records in WC for information related to works that have been translated. The mining process works with the most complete records and is initially concentrating on works of fiction. At the time of the meeting, the results of the work were available only in test but it became available in the production VIAF file in the August 2014 update. Thom also described the related work with xRefine and the records that are generated with “best guesses” about works, series, etc. Thom reported on the VIAF relationship with ISNI. VIAF and ISNI representatives met earlier this year in Paris to discuss ways of sharing information between the two databases. ISNI processing now considers VIAF’s xA file. When the resulting “reviewed” ISNI records are shared with VIAF, the relationships expressed in them are treated as if they came from the xA file and both merged and split VIAF clusters can result. Thom also reported on relationships with Wikipedia/WikiData. Wikipedia pages show both merges and splits and OCLC is hoping to experiment with using that information to inform clusters. In other changes related to clustering, Thom reported on gradual relaxation of the criteria for merging clusters. The intent is to avoid small duplicate clusters that contain sparse records (those with only a name and an ID) and duplicate records by moving them into an existing larger cluster. The intent is also to avoid large duplicate clusters (caused by duplicate records for a single source or by differences in pseudonym treatment). Statistics for the resulting changes are included in Thom’s slides. Thom concluded his update with a diagram of interrelationships among various OCLC Research files and OCLC production files. Criteria for Joining VIAF Following discussions during several conference calls over the past year, OCLC staff distributed a revised version along with the agenda for the Lyon meeting. BnF colleagues provided comments a few days before the meeting that focused on 1) giving more emphasis to stating the scope of VIAF; 2) putting more emphasis on data quality as a criterion; and 3) explicitly broadening the range of potential contributors to include LAMs (Libraries, Archives and Museums). The discussion touched on such questions as whether there are fundamental differences in cataloging practices used by LAMs and whether the document should be more explicit about who makes the final decision on an application. OCLC staff agreed to consider these comments and questions and to produce another draft for council consideration. Expansion of VIAF It was agreed that this agenda topic was adequately covered by the 2014 Annual Report. Sending Undifferentiated Records to VIAF Thom then used the final slide from his earlier update presentation to introduce the topic of including undifferentiated name records in the VIAF file. Currently, processing of undifferentiated names is done only for LC and DNB. These records have caused some problems in the past but they are better handled now. They do cause some problems with clustering but the problems are not huge. Changes have been made to insure that an undifferentiated record doesn’t pull clusters together and records now are marked in displays. Brigitte Wiechmann described the history of how DNB undifferentiated records came to be. No other libraries seem to be having problems with undifferentiated names. Add to Records: Nationality – Profession This agenda topic arose from a suggestion made at the VIAF/ISNI meeting earlier in the year. It seems clear that the addition of information about nationality (i.e. the RDA element “Country Associated with the Person”) and profession could be useful in clustering. The discussion touched on the following points: would nationality involve the use of free text or a coding scheme such as ISO 3166; and are we talking about true nationality versus country of residence? Several institutions reported on what is in use in some of the files contained in VIAF. The DNB is using MARC 21 field 043 and ISO 3166 codes. The BnF and ABES use UNIMARC field 102 and ISO 3166 codes. The Library and Archives Canada is using standardized terms in MARC 21 field 670. It was agreed that Profession could also help, but use of it raises the significant problem of language. It was suggested that VIAF contributors be surveyed. A brief survey was posted to the VIAF-L list by Brigitte Wiechmann on August 25, 2014, with responses requested by the end of September. Quality Control in VIAF Mixed Identities: Vincent Boulet discussed how to resolve mixed clusters and posed the question of whether VIAF could send a notification to the institution whose record had caused an identity to become confused? He also asked whether institutions would be willing and would have staffing to look at problem cases. It was noted that one possible solution to resolve an incorrect cluster would be to include the VIAF id of the cluster to which the incoming authority record should link. Thom agreed to consider sending lists and asked whether reports should be done in clumps rather than as separate reports for each problem. There was also discussion of Excel spreadsheets vs. HTML pages, with “.csv” files as a compromise. It was also suggested that there should be a “best practices” document for contributors. The work will be endorsed by the VIAF Council for 2014-2015. Using Pre-alignment: Vincent also showed some examples and asked if VIAF could reuse “alignments” made by VIAF contributors (based on record numbers contained in the record)? Thom responded that VIAF processing should be doing this now if the numbers are present. VIAF-ISNI interoperability Janifer Gatenby gave an update on VIAF and ISNI synchronization. Included were examples of how the kinds of reports that ISNI receives from users of their database results in data being passed along to VIAF that assists in improving clusters. She also described how the ISNI Quality Team works to improve the ISNI file. She also gave additional information about the VIAF/ISNI Interoperability Task Force and the meeting held earlier this year. One attendee asked whether it is necessary to be an ISNI member to have full use and benefit of the database. Janifer confirmed that it is. Vincent Boulet then gave a complementary presentation which compares VIAF and ISNI and discusses how libraries can benefit from both ISNI and VIAF. He concluded by urging that VIAF contributors and other LAMs join both organizations. Bibliothèque municipale de Lyon’s authorities and VIAF The meeting attendees then welcomed Cécile Davrieux-de Becdelièvre of the Bibliothèque municipale de Lyon for a presentation. After an introduction to the BmL and its collections and catalogs, Cécile reported the results of research that she and her colleagues did to support the possibility of becoming a VIAF contributor. They compared three sample sets (one of regional personal, family and corporate names, a second of regional geographic names and a third “historical” set of names of former owners of materials in the BmL’s collection). Their comparison indicated that more than 60% of the names are not currently in VIAF and would therefore constitute a significant enrichment for the VIAF file. The BmL is not able to participate in VIAF through ABES or as part of the BnF file. Linked Data Lars Svensson and Ted Fons reported on the satellite meeting, “Linked Data in Libraries: Let's make it happen!” held at the BnF on August 14th. More information about the meeting, including the presentations, can be found at http://ifla2014-satdata.bnf.fr/. Their summary described the 3 main themes of the meeting: moving from records to entities; the need for reliable authority data everywhere; and, the emerging group of vendors and service providers offering Linked Data services. They also discussed the 2 tutorials that were part of the program: one for “beginner” beginners and a second one for managers. Both Lars and Ted mentioned that VIAF had been mentioned frequently during the day. Ted then concluded the VIAF Council with a presentation on OCLC’s Linked Data strategy. He emphasized the goals of improving access to libraries’ collections via the web and of improving the efficiency and effectiveness of library workflows. He also described OCLC staff’s efforts to model the data that is of interest to the web, to extract relevant data and to express that data in structures that are familiar to the web. Wrap-up Brigitte Wiechmann thanked attendees and was in turn thanked by them for her service as VIAF Council Chair. Notes prepared by Glenn Patton