catwork: practical experiences in automation for retrospective conversion, reclassification and...
TRANSCRIPT
CatWork: CatWork: Practical Experiences in AutomaPractical Experiences in Automation for Retrospective Conversiotion for Retrospective Conversion, Reclassification and Backlog n, Reclassification and Backlog
ReductionReduction
LO Tin KingLO Tin KingThe University Of Hong Kong The University Of Hong Kong
LibrariesLibraries2003.12.09 (Tue.)2003.12.09 (Tue.)
CatWorkCatWork
BackgroundBackground
• History of CJK records History of CJK records
• Retrospective ConversionRetrospective Conversion
• ReclassificationReclassification
• Backlog ReductionBacklog Reduction
History of CJK recordsHistory of CJK records
• App. 168,000 records using Pinyin with hyphenApp. 168,000 records using Pinyin with hyphenation have been converted into standard Pinyiation have been converted into standard Pinyinn
• App. 70,000 records from RLIN have been convApp. 70,000 records from RLIN have been converted from Wade-Gile into Pinyin except Tag 1erted from Wade-Gile into Pinyin except Tag 1XX,6XX,7XXXX,6XX,7XX
• 79,015 old records’ Single Pinyin Syllable wer79,015 old records’ Single Pinyin Syllable were generated by a program in 1996 but not reace generated by a program in 1996 but not reaching the current Pinyin standardhing the current Pinyin standard
• Fung Ping Shan Library ClassificationFung Ping Shan Library Classification
Retrospective ConversionRetrospective Conversion
• Automatic Conversion again?Automatic Conversion again?• High Quality Records need human High Quality Records need human
involvedinvolved• Change the major direction from Change the major direction from
automatic conversion to automatic automatic conversion to automatic matching and retrievingmatching and retrieving
• Automatic conversion become an Automatic conversion become an assistant onlyassistant only
ReclassificationReclassification
• From Fung Ping Shan Library ClassificatiFrom Fung Ping Shan Library Classification to LCCon to LCC
Backlog ReductionBacklog Reduction
• From brief records to complete From brief records to complete cataloged bibliographical recordscataloged bibliographical records
How can automation help?How can automation help?
• Use Resources from JULAC members (AcUse Resources from JULAC members (Academic Libraries in HK)ademic Libraries in HK)
• Reasons:Reasons:– Overlapped collection (especially CJK)Overlapped collection (especially CJK)– Similar cataloguing standard/requirementSimilar cataloguing standard/requirement– Use the same library system (Innopac)Use the same library system (Innopac)
What is CatWork?What is CatWork?
• A byproduct of the recon & re-class projectsA byproduct of the recon & re-class projects• A Web version of InnoFace (see HKIUG 2001)A Web version of InnoFace (see HKIUG 2001)• Not a machine of automatic cataloguingNot a machine of automatic cataloguing• A cost-effective tool for reducing redundant caA cost-effective tool for reducing redundant ca
taloguing work among librariestaloguing work among libraries• TwinsPACTwinsPAC
– A new design of a user-friendly and two-in-one inteA new design of a user-friendly and two-in-one interface for cataloguers to compare two bibliographic rface for cataloguers to compare two bibliographic records from two WebOPACsrecords from two WebOPACs
TwinsPACTwinsPAC
What is CatWork? (Cont,d)What is CatWork? (Cont,d)• An all-in-one program includes features:An all-in-one program includes features:
– XML, UTF-8 & WebOPAC basedXML, UTF-8 & WebOPAC based– Cataloguers input Bib record number only for matchingCataloguers input Bib record number only for matching– Extract the Bib record and backup the MARC record from the Extract the Bib record and backup the MARC record from the
source librarysource library– Convert the Bib record into XML formatConvert the Bib record into XML format– Match the Bib record to other libraries’ recordsMatch the Bib record to other libraries’ records– Download the MARC record in ISO-2709 exchange format froDownload the MARC record in ISO-2709 exchange format fro
m the target librarym the target library– Add Tag 907 with the source record’s bib no. to the target MAdd Tag 907 with the source record’s bib no. to the target M
ARC recordARC record– Overlay the MARC record of the source library with the target Overlay the MARC record of the source library with the target
MARC recordMARC record– Options: Instant Matching and Batch MatchingOptions: Instant Matching and Batch Matching
IMDOIMDO
BMDOBMDO
Practical Experiences with CatWoPractical Experiences with CatWorkrk• Step 1:Step 1:79,015 HKU CJK Bib records were extracted and 79,015 HKU CJK Bib records were extracted and
converted into XML format for matchingconverted into XML format for matching• Step 2:Step 2:28,275 CU Bib records were found28,275 CU Bib records were found• Step 3:Step 3:28,275 CU MARC records were downloaded28,275 CU MARC records were downloaded• Step 4:Step 4:28,275 HKU MARC records were overlaid by CU 28,275 HKU MARC records were overlaid by CU
MARC recordsMARC records• Time:Time: The program’s running time of the above steps was The program’s running time of the above steps was
within 48 hourswithin 48 hours• Hit rate:Hit rate: 36%36%• Accuracy: Accuracy: 99.9% (Cataloguer found ONE mismatched 99.9% (Cataloguer found ONE mismatched
record record within 1,000 records)within 1,000 records)• Saved Cost: If a staff takes 10 min. to do recon and re-class Saved Cost: If a staff takes 10 min. to do recon and re-class
for for one record and the salary is HK$45 per hour, one record and the salary is HK$45 per hour, then then these 28,275 records could save these 28,275 records could save HK$212,062.50HK$212,062.50
How to balance the Matching How to balance the Matching Rules?Rules?
Major Matching RulesMajor Matching Rules
1.1. Search Title and Publication Year Search Title and Publication Year [http://catalog.lib.washington.edu:1087/search/t{215f6f}[http://catalog.lib.washington.edu:1087/search/t{215f6f}{213b65}/t{215f6f}{213b65}/,,,B/frameset&FF=t{215f6f}{213b65}/t{215f6f}{213b65}/,,,B/frameset&FF=t{215f6f}{213b65};Ya=1989;Yb=2001&,,]{213b65};Ya=1989;Yb=2001&,,]
2.2. Target Record has been catalogued Target Record has been catalogued [Includes LC call no. in local tags: 09X][Includes LC call no. in local tags: 09X]
3.3. Target Record’s Tag 260 contains Source Target Record’s Tag 260 contains Source Record’s Publication Place and Publisher Record’s Publication Place and Publisher [Use Inclusive matching instead of [Use Inclusive matching instead of Exactly Matching]Exactly Matching]
Matching Flows for Each Matching Flows for Each FieldField
Source hasTag 880
Source hasTag 880
YesYes
NoNo
YesYes
Source = Tag 880Source = Tag 880
Source = Main TagSource = Main Tag
Target = Tag 880Target = Tag 880
NoNo
Target = Main TagTarget = Main Tag
CompareCompare Target has Tag 880
Target has Tag 880TC SCTC SC SC TCSC TC
Current Status in using CatWorkCurrent Status in using CatWork
• CUHK:CUHK: Completely OKCompletely OK• HKUST:HKUST: Completely OKCompletely OK• POLYU:POLYU: UTF-8 port of WebOPAC has not been UTF-8 port of WebOPAC has not been
releasedreleased• CITYU:CITYU: Matching is OK but the MARC exported Matching is OK but the MARC exported
from its WebOPAC cannot be uploadedfrom its WebOPAC cannot be uploaded• LU:LU: UTF-8 port of WebOPAC has not been UTF-8 port of WebOPAC has not been
releasedreleased• HKIED:HKIED: UTF-8 port of WebOPAC has not been UTF-8 port of WebOPAC has not been
releasedreleased• UWASH:UWASH: Completely OKCompletely OK• HKU:HKU: Completely OKCompletely OK
Future or TrendFuture or Trend
•獨樂樂不如與眾同樂 獨樂樂不如與眾同樂 - - 歐陽修歐陽修(Singing together is happier than singing alon(Singing together is happier than singing alon
e)e)• Resources SharingResources Sharing• Co-operationCo-operation
獨樂樂不如與眾同樂獨樂樂不如與眾同樂 A JULAC member can be selected as the source Library
Next VersionNext Version
• Backup & Download: using XML format from InnopacBackup & Download: using XML format from Innopac’s xrecord function instead of ISO-2709 format’s xrecord function instead of ISO-2709 format– Simplified the program’s proceduresSimplified the program’s procedures– FasterFaster
• More LinkagesMore Linkages– E.g. ISBN or Author for referenceE.g. ISBN or Author for reference
• Add User-defined matching rules: Flexible, Simple and Add User-defined matching rules: Flexible, Simple and User-friendlyUser-friendly
• Add a Bib record Pre-Editor on the Web for cataloguerAdd a Bib record Pre-Editor on the Web for cataloguers to edit before uploading the MARCs to edit before uploading the MARC
Thank you for all cataloguers’ Thank you for all cataloguers’ efforteffort
If there is no cataloguer, If there is no cataloguer, the matching result will be “0”.the matching result will be “0”.
LO Tin KingLO Tin King2003.12.092003.12.09